<%BANNER%>

Evaluation of Electrochemical Processes Occurring in the Cathodic Reaction of SOFCs

Permanent Link: http://ufdc.ufl.edu/UFE0021426/00001

Material Information

Title: Evaluation of Electrochemical Processes Occurring in the Cathodic Reaction of SOFCs
Physical Description: 1 online resource (140 p.)
Language: english
Creator: Smith, Jeremiah Robinson
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: cathodic, impedance, lscf, lsm, reduction
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The need for high efficiency and low emissions power sources has created significant interest in fuel cells. Solid oxide fuel cells are desirable for their fuel versatility. The cathodic reaction is known to be one of the major causes of power losses in SOFCs, but the exact manner in which the cathodic reaction occurs is not well understood. The cathodic reaction was investigated using primarily lanthanum strontium manganite (LSM) cathode / yttria-stabilized zirconia (YSZ) electrolyte symmetric cells, as LSM is one of the most studied solid oxide cathodes and the symmetry of the sample simplifies the study. An in-depth investigation of the cathodic properties of lanthanum strontium cobalt iron oxide (LSCF) was also performed. The areas of interest are identification of the individual processes occurring in the cathodic reaction and understanding how the reaction is influenced by experimental conditions such as temperature and pO2. Elementary steps of the cathodic reaction can be analyzed individually using AC electrochemical impedance spectroscopy (EIS). This characterization technique gives overall polarization impedance as a function of applied frequency. The output spectra were analyzed giving information about each of the significant steps of the cathodic reaction. The effect of microstructural and interfacial changes on the cathodic reaction was also investigated. These changes were produced by sintering at various temperatures and times. The microstructural changes were analyzed both qualitatively and quantitatively. Ultimately, a direct relationship was established experimentally between the cathode microstructure and electrochemical performance. This relationship was modeled based on theory involving reaction kinetics.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jeremiah Robinson Smith.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Wachsman, Eric D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021426:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021426/00001

Material Information

Title: Evaluation of Electrochemical Processes Occurring in the Cathodic Reaction of SOFCs
Physical Description: 1 online resource (140 p.)
Language: english
Creator: Smith, Jeremiah Robinson
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2007

Subjects

Subjects / Keywords: cathodic, impedance, lscf, lsm, reduction
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The need for high efficiency and low emissions power sources has created significant interest in fuel cells. Solid oxide fuel cells are desirable for their fuel versatility. The cathodic reaction is known to be one of the major causes of power losses in SOFCs, but the exact manner in which the cathodic reaction occurs is not well understood. The cathodic reaction was investigated using primarily lanthanum strontium manganite (LSM) cathode / yttria-stabilized zirconia (YSZ) electrolyte symmetric cells, as LSM is one of the most studied solid oxide cathodes and the symmetry of the sample simplifies the study. An in-depth investigation of the cathodic properties of lanthanum strontium cobalt iron oxide (LSCF) was also performed. The areas of interest are identification of the individual processes occurring in the cathodic reaction and understanding how the reaction is influenced by experimental conditions such as temperature and pO2. Elementary steps of the cathodic reaction can be analyzed individually using AC electrochemical impedance spectroscopy (EIS). This characterization technique gives overall polarization impedance as a function of applied frequency. The output spectra were analyzed giving information about each of the significant steps of the cathodic reaction. The effect of microstructural and interfacial changes on the cathodic reaction was also investigated. These changes were produced by sintering at various temperatures and times. The microstructural changes were analyzed both qualitatively and quantitatively. Ultimately, a direct relationship was established experimentally between the cathode microstructure and electrochemical performance. This relationship was modeled based on theory involving reaction kinetics.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jeremiah Robinson Smith.
Thesis: Thesis (Ph.D.)--University of Florida, 2007.
Local: Adviser: Wachsman, Eric D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2007
System ID: UFE0021426:00001


This item has the following downloads:


Full Text
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 E20101206_AAAAES INGEST_TIME 2010-12-06T23:37:14Z PACKAGE UFE0021426_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 17558 DFID F20101206_AACRKS ORIGIN DEPOSITOR PATH smith_j_Page_107.pro GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
7960c994549b2f2c94d28d0d68a2b528
SHA-1
100032b34a6e95d754aca69a2cd80eea3f5fe0df
13091 F20101206_AACQHQ smith_j_Page_007.QC.jpg
91e17eae752dda3a5b0533e92eb20a86
b1c5e29256f4d517761bf46956d243d87e1a3669
27989 F20101206_AACRLH smith_j_Page_010.QC.jpg
6a98949e74982779afeae89d309441b2
cc3c7a28382c2986ce84fb6a47d86a2a6e26f0a8
4749 F20101206_AACRKT smith_j_Page_066thm.jpg
4b65fb55c395b95b3abf3d471e2131ca
4cab0cb3c84363da20499c4c2fc0aaec3592800a
6466 F20101206_AACQIF smith_j_Page_048thm.jpg
10ac8673c352f11950b15ce716ef97ed
617329e8277cc64e0dec91cbd6b1882383a0b1bf
1604 F20101206_AACQHR smith_j_Page_067.txt
ce1bdc7352b2ea822340fc1772bff5b8
014934498f5f2f7f6c179fd1b1aee499490282c0
6834 F20101206_AACRLI smith_j_Page_070thm.jpg
6102793804a213e91e16d645caa4cf5c
ef1bd2121bd7b2d87de35282015477ca411b44c0
1051946 F20101206_AACRKU smith_j_Page_097.jp2
243b5bea2e414b4c201f85bf75a462e4
56ed7aa125b69fe284618282d1f020dd29fd9924
392551 F20101206_AACQIG smith_j_Page_128.jp2
db6812419a62184d03abee96bad56312
a2e54bdb9d3831a8df518e2cd0b6f1468ebe6234
53547 F20101206_AACQHS smith_j_Page_032.pro
cd0d5e0746fad98b3fbf2dcc62d63b97
f9ee529bf84b344fc86940bab025dffd8fde405f
37951 F20101206_AACRLJ smith_j_Page_023.jpg
83b989f3028c60cd6b4f9dcebdb5a20b
977ccc813714447dc9414204af1a5ce4758b2cb5
25271604 F20101206_AACRKV smith_j_Page_121.tif
b1eb3f512d348ae83f17ed2391bdb8c6
c7ad51e41ab08dd2bbca5fa14ad7226594c69476
F20101206_AACQIH smith_j_Page_038.tif
31a6986471716409e7dfbd0b5d475d0c
fbc82d0a0796d33c7c4e0a378c8a62c52a7360c9
26185 F20101206_AACQHT smith_j_Page_032.QC.jpg
073fcf78511dd1eba3f190a0d6bda54e
b8236663a4f9843d95b0105221617771c45b897a
2360 F20101206_AACRLK smith_j_Page_036.txt
89ad75ab5b77c7279a40ecc4e9bc5b67
fa77b0a1de4d1f1704eb04775dbe01445637dca1
54221 F20101206_AACRKW smith_j_Page_048.pro
06aa7a65d28f3345bd8beb861c0b1f25
a9ff47acd47c033811d5d4d1240dcf3cae7e4f26
36731 F20101206_AACQII smith_j_Page_129.pro
9c4f7e8fe21da652d1bd851ca8715b54
1bff6a7beb595abd32d5888f38b66167d7c6a94a
4839 F20101206_AACQHU smith_j_Page_063thm.jpg
8b2b803671a4656abdd6e5d14c3b1e80
762e2700127902b3c57895e96c624ae21d207836
1051982 F20101206_AACRMA smith_j_Page_013.jp2
63a9dfa49d094a62790dc335a618fad7
04720bcf54558546a05ed6744493e2eb9ae0a321
86432 F20101206_AACRLL smith_j_Page_096.jpg
8672578ba26de85dfc0df3991ab8b30d
bb2bfc3715e6663fcf4e9735a48ab3b7ffd58542
27443 F20101206_AACRKX smith_j_Page_110.QC.jpg
c13ec96c1fa539ee0d7fbfcc942fa077
0df06f39679bf02f194405934c965d5d282f4c9a
54978 F20101206_AACQIJ smith_j_Page_106.pro
18efbe93e11c7d48082d7630785a2d2a
da41c2c3516d06436586d9eefbb37e6fc01dda28
59814 F20101206_AACQHV smith_j_Page_013.pro
75c2fc7c2c16ca5ec2c4f1911819a61d
ec2f5a065867af8fa802e9e0c8a2e18587597954
1051977 F20101206_AACRMB smith_j_Page_042.jp2
5a17677838cba7616a90d330be6cb2e5
7db20f5edf4014ac935df718d0203108796ae961
8419 F20101206_AACRLM smith_j_Page_001.pro
c0911eac604bc7a11d765c2dc38dbb49
5187855661d06f5136bf6bbccce92be23737a56b
42655 F20101206_AACRKY smith_j_Page_104.pro
dfe78f97c142c2b5f315d673b79ae629
521b255ae61be1c81bfe7755a9ce88588ddaf187
44478 F20101206_AACQIK smith_j_Page_062.pro
ed75b973ad19b8f7dd86ab0727a40c71
1a85e73f5f209772800929429f8b3d76cad86cc7
326 F20101206_AACQHW smith_j_Page_082.txt
8e9eb6790a3c6b4af70adf5ac336fd4c
a0f2bc79fdb43c4ea592eb101d0b007a9c80fb44
1051967 F20101206_AACRMC smith_j_Page_061.jp2
f05f1ed3553d22e6053b9d5ef3a1cb74
d83e075aa3db357c6977347688998c2ae214416f
48422 F20101206_AACRLN smith_j_Page_098.pro
79619b20540583a2a9af5de49b7d8c24
55b73f7ab01608a2e70c7663aad47c157e87223c
46967 F20101206_AACRKZ smith_j_Page_040.pro
fb4dcd6acc399c1aca213dc7fff0cd10
0e6f3542370cbb2a00614e76847157677e84c656
39455 F20101206_AACQJA smith_j_Page_081.pro
73ce95a969305e25087b1465fddfc647
7fac37c208e31f495a210fca0009770a800c12ce
26929 F20101206_AACQIL smith_j_Page_069.pro
bb2c54662c51c49c4fd22380d9de981c
0d3b6abc5bf693220ee3fbc1945ce3b553e2cdc0
219 F20101206_AACQHX smith_j_Page_003.txt
0067319108594dc9793e9dff05e7b19f
ef3969824d165b2e91ad78d87cbdcb9c7d11dd89
1051957 F20101206_AACRMD smith_j_Page_076.jp2
8e892dbae1fa58fb5f0003984f38f990
1298c4bbb20e334ab7c62932cc487465d1d88d60
35914 F20101206_AACRLO smith_j_Page_123.pro
073bdec07dbb132bc2638c10295fdc8c
8663274e94a6e0bd547d3d586a2c75818f0b196c
1959 F20101206_AACQJB smith_j_Page_100.txt
12ae4c5f901a7cacc23217b2eda9aa8a
cea3a48e4d2d7d873acd618fdfbe17ce7d9c8379
57124 F20101206_AACQIM smith_j_Page_050.pro
841a7ded7cc7498ea63e9b3dfb7f5bbb
a2f11012fd9b6ef69cd4c3949d3445d6593b7cbc
F20101206_AACQHY smith_j_Page_068.tif
ad493a99e738c2c9280b173a394fb30a
9f693c2149587903889a108a1db1f261807bd6e9
1051900 F20101206_AACRME smith_j_Page_084.jp2
a3d512f273c72e94600c2cb1999ea6ee
1e76c0bb73ffac95b8cc73a6f073f0c2d125ea14
1051986 F20101206_AACRLP smith_j_Page_126.jp2
402fb2b46d265194ee5d9aa06f89ccd2
22f0d8241c00a550156c7ffada6f72099970dc9e
1051924 F20101206_AACQJC smith_j_Page_025.jp2
1bb9709d534a5becdac6213b654e81a2
886bc55f80d2ed5b1185fac7ba21431df87bb9da
851080 F20101206_AACQIN smith_j_Page_069.jp2
4a666e04dc2800d3428d1c4251398448
598c2a1760a07aa09a468e3ff8feaf28f873f002
2404 F20101206_AACQHZ smith_j_Page_086.txt
a0501927e72e6c801a431fdf5c207d0a
659f81fd401664e997b2b926a3f5ccb001908b12
F20101206_AACRMF smith_j_Page_015.tif
342fb46aee192a370e69cf1102483803
542f3197d569df0f71a8aa97c394f85e58bfd178
11966 F20101206_AACRLQ smith_j_Page_125.QC.jpg
28a87d179b392264753959d00140c415
d2df0f8286afb97e05c69ab0f2190a0f271b9e77
4894 F20101206_AACQJD smith_j_Page_087thm.jpg
368130489272a474664806c25b7a1401
f9abcec1a54eabdd870d31ac7e118bf78fac6b08
F20101206_AACQIO smith_j_Page_101.jp2
0ec54bddcfc67826bf32caf177d0d2a2
d2b2f1f6916f16773b87abe98c051a9ee1263048
F20101206_AACRMG smith_j_Page_042.tif
9d25dd948df31fa2cf0df79204c752d0
66dfbb69b81682c025072789451fecd5cf07f0a9
161268 F20101206_AACRLR UFE0021426_00001.mets FULL
985b06c928b8b47d9aeeafd03cca9831
2adc04f320102420d5818171e96ee160ef0e3f8a
2165 F20101206_AACQJE smith_j_Page_137.txt
5b1135e533a66e537144ae3af51883d2
ad2fa0d5321c54198d239aaac11539f1f85d7936
121962 F20101206_AACQIP smith_j_Page_132.jp2
ecb581c36e5fee30935a9a49b291ef58
15317164333c7315947aecc2657d65438f9fdd35
F20101206_AACRMH smith_j_Page_044.tif
ebb057cc6e7db61de8b6f1a5e55e947b
fa8cda6ba0daf4398799585f7342af4ff5457918
3901 F20101206_AACQJF smith_j_Page_117thm.jpg
448cfbaed0d535f28cf4d3e3cc4e7c77
1a85687e02f4be97929ff51237da11430e2f4acb
28461 F20101206_AACQIQ smith_j_Page_078.pro
054175010b414de2d52dd9c2dde8decc
6025bca8b0f590d3c937e761be2d701c629b47fb
1940 F20101206_AACQIR smith_j_Page_093.txt
f333668dcad2c2bc6111d154938791c8
965bc21adca8c2f1db1d99f8e47d346b65ca13ee
F20101206_AACRMI smith_j_Page_069.tif
4d8b574ea39011e4430f8eaeb938d478
10cf29888a93ce041f3f4c1e0b90bf6a30c8bd7e
85354 F20101206_AACRLU smith_j_Page_035.jpg
4805ebe764a23c501ca4a062eb3b600e
67ed1dbe9d753581e7abfa71a8e3c6d8685fd30a
1053954 F20101206_AACQJG smith_j_Page_134.tif
bd8b88d27cc0e6b94f9661537b014ced
61e3d28c710d72cada7ebf88d48dc262315d9705
F20101206_AACQIS smith_j_Page_067.tif
bdc6a701850c199400004dd4abb318b7
cf581feb4dd42fb4ff652fb35d3a3e9159624777
F20101206_AACRMJ smith_j_Page_079.tif
df8d6f2933dc7f128544b14096e3895b
3f50b7ed3804d80ffea6fb33524cf2c0b1b64839
78752 F20101206_AACRLV smith_j_Page_057.jpg
cfbd1e361bbc0112378fdb94fc75ad79
6f1997eebda292feaaa446fbc312f48fd0cd4892
20240 F20101206_AACQJH smith_j_Page_005.QC.jpg
85b89f0c75fc5c5f978e25e2afc158b8
754db6f2bf340ce70efa81b44d71603f3c2a9099
33846 F20101206_AACQIT smith_j_Page_102.pro
59b269c8f75743ebaad12b8c59d605c2
d9d8f3a1d4accb09ed8fb941cb51eb19e777bd0a
F20101206_AACRMK smith_j_Page_100.tif
59d5d289525ac550689398b1abad0609
71dd2d266e0902e15a1aaf0c63ebe2890e70ba5f
89696 F20101206_AACRLW smith_j_Page_070.jpg
571e346be9015d8fa182b1eb69680388
b2feef3bba5077c0c3426ef1bca2f7880198fea9
2302 F20101206_AACQJI smith_j_Page_110.txt
9e339d09b1e8975501ed38021e5fb776
82df721ecbedff07f76987b7814d1cd1922aebff
61263 F20101206_AACQIU smith_j_Page_102.jpg
9c690e72ba504dffaeec51c82879063e
ef179ffd271764b3cfd0d9440197e9fd23f1dadd
24025 F20101206_AACRNA smith_j_Page_055.QC.jpg
85a903a55a62c80ec5eaf49a7f3ba86a
70c3edc7df1084df82f294c11d3932d3a4d84871
F20101206_AACRML smith_j_Page_114.tif
958325bbb99376aba0d55243ce18c6e9
289d2ab77963692df7079c13c0603b624fcb51fd
83354 F20101206_AACRLX smith_j_Page_085.jpg
6021590ef7326c30bc916c35b1a7995a
fdf6075a1d82548d89c99b1949d29526a233a166
14916 F20101206_AACQJJ smith_j_Page_115.pro
be8230feed6724ff9b29816e55226b84
414145b5b9f690e3f5224ffb98fa458a62a22e79
626790 F20101206_AACQIV smith_j_Page_039.jp2
34b2e047996f6721fc6b6d991f46e0c2
fae8a0ca8187ca791e6e0324edc3a0d03e8e4b8f
21771 F20101206_AACRNB smith_j_Page_067.QC.jpg
4a470df6d4f4b8f1af46a6bc0dfb488e
944f84347ce44a9df11050f7e7884e24cbb04c2e
F20101206_AACRMM smith_j_Page_116.tif
22d7e9b17188b21615e782c43a9166bd
3113ea8eb79a0f53fe797a3094c27be86df4964e
73552 F20101206_AACRLY smith_j_Page_091.jpg
5836033c0a105b7d048f03c089e5c918
189d8865c13e55ba358d8ad9f26ff30a56361601
6703 F20101206_AACQJK smith_j_Page_097thm.jpg
20eb103fe050cd8af6a712fcc7a5d8cf
d292abef5343e01217ac1a38cc5df9cad8ea4937
70284 F20101206_AACQIW smith_j_Page_133.jpg
c8457f18edc7d5259fcfaf7c4c754aed
d40f7cca9c81885557bfcde6cad8a8748b355d38
5647 F20101206_AACRNC smith_j_Page_069thm.jpg
235b40321ff414272d3fe85530366800
187cfc4800d3db7dba15fc0f0e0f601b8cc51b28
58236 F20101206_AACRMN smith_j_Page_017.pro
b0e4ef3f0861d56ce970d1abfe4b7b4e
3adf306059d6a8d95087c5d4d35906f4a224d360
71333 F20101206_AACRLZ smith_j_Page_126.jpg
99cc3a23ecbcc9c26b60ec8f6d46ff58
9d9069f154ab7c8157c23f6422dc09122a6d64ed
28035 F20101206_AACQJL smith_j_Page_074.QC.jpg
e7a136fc858d4a6562c800f83fb1191c
48a31b40cfa2b311f66f241479c15bcc7253c17a
60392 F20101206_AACQIX smith_j_Page_132.pro
8e56843c0af178952b85a8a67a6e916f
3a66d818b6f59cc75c20da10818f588bcc88187f
1918 F20101206_AACQKA smith_j_Page_041.txt
d1a6f6a5fcafae5d0c619be8da9793f8
547c842052af76307952024cc98c4b52241d60f8
208695 F20101206_AACRND UFE0021426_00001.xml
2f8785c249f52a0b62363ab6b5cbe4b9
b8d41b211844fc07b5294c59f25e971d4df8167a
57544 F20101206_AACRMO smith_j_Page_021.pro
666d9282591e088777e99d2743bb4e0b
3005f6e2223b3ab2dec6e124968521ac6bcfa79a
6013 F20101206_AACQJM smith_j_Page_140thm.jpg
a10e96d7ca87e53d7611de5e30f41604
14d274a699d0e5c1c314913fb09875c396c6a1fe
2306 F20101206_AACQIY smith_j_Page_026.txt
f36d4eb327181464a200a4290c99fcce
e65b26dd887b0af68cc30eaaa7b8f4d153dedf68
15043 F20101206_AACQKB smith_j_Page_087.QC.jpg
087a7c2378fcd51f12ff7b2c831a0275
d62197942f3a4618a3f2ee4b881a02a61e38de1f
23723 F20101206_AACRNE smith_j_Page_008.QC.jpg
7f8d3f773aca4faf58320a7e5b4f972b
d5fc64972edec51df2ab6820f416302b0950631c
9096 F20101206_AACRMP smith_j_Page_066.pro
4702d2811b11f0fd0dec0e0a0788335f
d85c7507efd463e6f0c2a9db0c96ea27051293ed
89770 F20101206_AACQJN smith_j_Page_013.jpg
9b22f10f40f158adca0d7660018d43d2
487c60ef5929d429d0dbfbb67f7978da450aeef1
6990 F20101206_AACQIZ smith_j_Page_029thm.jpg
b1b45a309c0e0dbf43220502af0d51a8
09ec25c5441a61ce6fbcc2206ff5c464c78b7ac9
3904 F20101206_AACQKC smith_j_Page_113thm.jpg
e0ba82926e580739bbfbcd3a9d1520ae
4c4c1ab3b1a49c6f31793df0dff6bf0634f401ad
4803 F20101206_AACRNF smith_j_Page_012.QC.jpg
b4badaf4d73dbc79e0950a243e29c45a
6aac617f71bc03126fbd7a1cdc38a0f8294aa641
61252 F20101206_AACRMQ smith_j_Page_086.pro
2eda02f1028483a797c2a716a63dc66d
df7a23a6b0d9a6c80dd196edd396a187105c9f69
769104 F20101206_AACQJO smith_j_Page_131.jp2
9fd426f40db7b2c418756eab6000c3e1
b92b238a942e337f365ca0802720f5efccc2c976
F20101206_AACQKD smith_j_Page_084.tif
4c5ee5edf93841e271bf3ec1d77faaff
18ae0532b171432133f23de96a809c1f90533596
6376 F20101206_AACRNG smith_j_Page_020thm.jpg
746718799038f14d0eae85c824f937cd
9e1df941910bb573e8dc781f12d3e5ef0d2e1df7
51414 F20101206_AACRMR smith_j_Page_130.pro
b7d28063ec584a399f5db63f02378829
16e2d63d51a3a897893eb7a794997be553c8d64e
24491 F20101206_AACQJP smith_j_Page_020.QC.jpg
69011deb8b633d39d310b5a4c6cba55f
4663bad375ac62e8fd634df26bb2b2ca658410c5
F20101206_AACQKE smith_j_Page_016.tif
cb1eedb4d54a0c6c7d34a924b94e0410
c8c07d1ff0e2e6955a1d625fe09a8dabd735ddc3
24682 F20101206_AACRNH smith_j_Page_026.QC.jpg
b7eec340263b0fad0782c5d3c5dafacf
443c7703d2deb3fff2362e9f81469b9f86e43079
2421 F20101206_AACRMS smith_j_Page_029.txt
52dac197912e7e8f237e39d5cafc3303
e20ed91e910d76839158ccb36e65b296c9df6161
F20101206_AACQJQ smith_j_Page_046.tif
36a6787788204e3c759a81a1ee314d22
abef5182b7352fa3952196d58acfd255e7bf93e6
1051962 F20101206_AACQKF smith_j_Page_024.jp2
ed358372c77900677b6654dc5ab387f2
a5a72b17f87fe848ed9f366f17f8939f6d8a8868
28982 F20101206_AACRNI smith_j_Page_029.QC.jpg
efb53bfccea009b7a2587a7a9ca5b995
96916cb7989dfb1f42a66142691c3889d01246af
1093 F20101206_AACRMT smith_j_Page_069.txt
0873b1dfba36111d721e19e94428d6ed
e5d0e588bd18711b84a16f34f9ae626042ec797e
39856 F20101206_AACQJR smith_j_Page_117.jpg
b4bf050f26b259514556d1c57ade48f0
5dea4b19295bc914f8214fb91c7a1b414f7cfa95
54433 F20101206_AACQKG smith_j_Page_131.jpg
09654c9c7140ee2c1276a1acb7b09a3c
e3fdfce0956170e093c30c33a42ff42124e2a5b5
1728 F20101206_AACRMU smith_j_Page_090.txt
9bbb66885cfc9c6272a78b39982c3eb5
e3d6c486082d216e3a8f50c3a21f782d0d814004
6787 F20101206_AACQJS smith_j_Page_059thm.jpg
7fbcf2a39a0d4eec542aacb9118f076a
ad73112437b2c7247995be6eb60f966f31789170
24998 F20101206_AACRNJ smith_j_Page_031.QC.jpg
68e4112637779b8953d4f136ca4edc7a
710040a18da351194f9da8acb090eed0403dc14e
1282 F20101206_AACRMV smith_j_Page_099.txt
a34764abc15858df90812b3219584bfa
b523937d9093f20cfe77b8b95878b760feb3cf23
48598 F20101206_AACQKH smith_j_Page_014.jpg
ed55d86ff3badda5ed563872ed5e7eeb
e1a9ddc7d34ec94ad256458b22c81dee79e1900c
62144 F20101206_AACQJT smith_j_Page_076.pro
4aba8c311b8f5b95f0b68c2c35faf99e
aca81f471b240e47df4bc19f46a659f28e220bd9
25506 F20101206_AACRNK smith_j_Page_053.QC.jpg
4155dc5e3ca3f8961ef2157e903f43a6
9a92a6a57a18d94a53d5acedf4d27459656cfd40
1017 F20101206_AACRMW smith_j_Page_113.txt
6afeb3b4a6655862d83a5008b82f4f09
374888048804d2728a3badfae0b18ce838c60487
5901 F20101206_AACQKI smith_j_Page_030thm.jpg
06c73198aad9cb2ce00c4920439550b0
2b1a173a6e2704c0692f6634abc5fcc2b0cb2e91
54798 F20101206_AACQJU smith_j_Page_031.pro
7bf8020fde78ba782443e5b777cbafa1
e1c88a1ea4ec46791f6c1eae027637085cd6b5f9
21429 F20101206_AACRNL smith_j_Page_093.QC.jpg
5710b9fb0001910428f6f5b9a501ab87
a4dc8ba65a109ab32ff24a9d8f49a7edc042cb3c
6571 F20101206_AACRMX smith_j_Page_079thm.jpg
b12733bd018f4eb6a0a313964becd095
119a2d11777852534e56242ed16e758ea7905876
46785 F20101206_AACQKJ smith_j_Page_111.jpg
c984ee3a18db01f4dc702667270d41bb
d77263e1ee29f3b6148f0b850ebfe33cd1ad3f9c
107730 F20101206_AACQJV smith_j_Page_137.jp2
97a77b096828a605e37a7ff8c9d37c00
d9de6662b3dc44a0ca2fe0d7c9e171c76baf82fd
5992 F20101206_AACRNM smith_j_Page_093thm.jpg
f32108436c33a3c4a2986cd1edfbc43c
59df206c11c952333323877b1a3bf0b347de48b5
18947 F20101206_AACRMY smith_j_Page_027.QC.jpg
9b0636da2ff7ed4ba14a31f99f8979a4
9d214c60d586d2cad96a19b46dd00124c1cf33cc
F20101206_AACQKK smith_j_Page_105.tif
24cf586a5de32fae075bc21a04816518
d6b8a61548d492389df60ff42207a51df75be98b
F20101206_AACQJW smith_j_Page_080.tif
21049a45c1f8ed41526c98dd17adf3fd
e0e77ebe3e333c6fb01cc6d94fea4b556d16896d
20277 F20101206_AACRMZ smith_j_Page_044.QC.jpg
7683c1da84c9518092e94fba9b9d0c18
ee557166882f1fa7fa4e33eac0b7df2b0d864ae4
21978 F20101206_AACQLA smith_j_Page_133.QC.jpg
001f02f7fbc1427cb2356e060a19eb4e
01ce17c18a51521b16b5f6123aa6c80ce6178095
2287 F20101206_AACQKL smith_j_Page_017.txt
8cda6b2e9f883ffe18a9ebace0499f1a
82a8ec652113d64d7936d1c629b741ee7d7c8294
F20101206_AACQJX smith_j_Page_096.tif
4601485c513e5adf55ca1909c1634ad4
a15018e5f3452e0a736efba1e577455922e53d93
25788 F20101206_AACQLB smith_j_Page_084.QC.jpg
cbff484673834eca6232c575354e13b9
511a34f7b7de784ee7a9a485394fd8d19e4bef33
25601 F20101206_AACQKM smith_j_Page_085.QC.jpg
ca2b6a0c3146c4a97742feb3fc4c18db
5829f9ed1dc1db5f3da2336fc7baa576cf0af138
F20101206_AACQJY smith_j_Page_001.tif
0a6c4b27abafa75f507da3a021b34f18
8fd3ce144a46889a7ecc9d0463826181427a7ce6
26908 F20101206_AACQLC smith_j_Page_025.QC.jpg
4e8ad234abdccb3dfb89a1bfbae63f5e
8fbe74a949039b9594149ca15ecd1323a5aab0a1
27446 F20101206_AACQKN smith_j_Page_124.QC.jpg
e491a0bbb34e084f919a3aacc38f0769
c6c831be440dee63cb0d0932ccc3721ccca64a7c
46697 F20101206_AACQJZ smith_j_Page_066.jpg
4250da0778232f6f2aae4c447f1bc006
28d93d8e0b9ce34aa71e21df1dfb029d8a303448
54824 F20101206_AACQLD smith_j_Page_114.pro
8c9e3bf9c5cf5109715c4d902caf6882
692c338a8789ce772bb9dfa51e73888ea94ff906
67473 F20101206_AACQKO smith_j_Page_104.jpg
a38f60d0f08caa51bb1802a82f215e70
3e648fcdf78cc70b8f52832b385589deb5f02356
27734 F20101206_AACQLE smith_j_Page_094.QC.jpg
1dac79c0c9561e5884a98b8b7d03bce2
34229b540ded5ed9c024162766fa3432ca500964
59881 F20101206_AACQKP smith_j_Page_024.pro
555b74342d7a3e60a4de3f30c75a1ad0
de71d85c145a0ee56bcb8aed8c1dc2648f117c9d
16963 F20101206_AACQLF smith_j_Page_064.pro
86fc40b04ba4847fcd0aef4946bf23bf
e3b5c5c675a9c7d2966310c26143bf83f085d226
1051978 F20101206_AACQKQ smith_j_Page_050.jp2
e5ba5eb4749ea03d25e1c2941199f44c
0001812b8cf411e70bf8c74938e3c8c5a0147c58
F20101206_AACQLG smith_j_Page_002.tif
b6b997192f2538eaa551d4a031693735
b96237049a3cd03847413e46061d43cbd8835c07
5194 F20101206_AACQKR smith_j_Page_095thm.jpg
9b0a83db7b3c6ffad89317bd7232405c
cbadedc0403e5a8dcfc6246ccdbe15ca00f37621
1950 F20101206_AACQLH smith_j_Page_006.txt
046d5e147994f0f31af4dd5a4a7ac00c
6ba30ef138090243149913357e7674ee22d34a94
2277 F20101206_AACQKS smith_j_Page_074.txt
3289d62dc58240ecde80d1c5781fce0d
9272fc0ec9a7aa906775d945e7b9a331b0f37b61
5548 F20101206_AACQKT smith_j_Page_104thm.jpg
3d099fcdab41f8ed948fc0d07a1dd43a
3155a931d117babaaaa5baf429a95858eda330b9
44952 F20101206_AACQLI smith_j_Page_080.pro
256a94b190410517fcd368a18e9746a7
6860721112913256765105f6e21702364656d541
83396 F20101206_AACQKU smith_j_Page_053.jpg
27e49967809e2a1b7d8bbac25695e819
482915f82585b40f18c3acc7493d4947c83a2471
F20101206_AACQLJ smith_j_Page_122.tif
2cd406d96ab4916eb2d7eb2c5e8a3f49
de821640543c95437ceb7baa8f736a83ea300b54
2122 F20101206_AACQKV smith_j_Page_020.txt
1c728fd26b3af9379edae44e68602bf0
73fbae5f96c308798b39845c119d7f91f2aff6c7
4923 F20101206_AACQLK smith_j_Page_005thm.jpg
ee7204a6ed9213ffc862dcc0738a1ee6
de8fce9dd80a74df9c605b7acb2cc1431f4c3fef
F20101206_AACQKW smith_j_Page_047.jp2
3fff1833423c33cec2e08a8e8f46b40b
7c794b092c66a0bf1a579d12b0f7ac8a1e997f73
1051831 F20101206_AACQLL smith_j_Page_078.jp2
7786a6b1c8addcf03f4b32a638fb3083
4755309482d767b116a49c937e70434730774a76
716185 F20101206_AACQKX smith_j_Page_117.jp2
013aa9b71a343338fbf8cfa68d8f7db4
39df97e4f46a9f62f5cff703aa8c52b8a02172d1
33827 F20101206_AACQMA smith_j_Page_131.pro
e85df34d70ea1c230c9d9976b19d6de7
df7da4c15e919331d0c6bc6e174e3ec1e15580c9
2157 F20101206_AACQLM smith_j_Page_042.txt
8bd2ed945cf518cff44bae656c72d972
69efdce5170d7ec174dbd6dbcce8408958c27528
23043 F20101206_AACQKY smith_j_Page_040.QC.jpg
69884880fbe647566d7317194705487e
c4a54f1f97bb9dc111a281af946490e34a719464
977 F20101206_AACQMB smith_j_Page_039.txt
40271fbfaaef44322561e6b4bab2232b
fe847355a5000dc76520349b91d57d2d35ba3af5
28135 F20101206_AACQLN smith_j_Page_118.QC.jpg
fb29a95e430d678c7bf81ae56789add1
e5333a9890dc43fbc6224324192e5a736a1a387d
42133 F20101206_AACQKZ smith_j_Page_006.pro
b26cb7db950a77f12e99bbe46f19c060
d8082cd29f82588ae0f31c0e11540d29dbdb0470
6752 F20101206_AACQMC smith_j_Page_075thm.jpg
94bc730a41d35e2cab963432464a72ba
1326e186dc7c34fcd9cc7602f94aaffa93d7e932
70263 F20101206_AACQLO smith_j_Page_090.jpg
2e4444f81dcd2a1cc487c61972acd741
d2ca3aaa3d318079d7683857b42ed4e05eced847
10628 F20101206_AACQMD smith_j_Page_128.QC.jpg
45a85405c0b86a0043e0499b4fbdb206
4725c6e6c5d070735c9a5b4e10227fe9ba646bb2
65730 F20101206_AACQLP smith_j_Page_129.jpg
93491c641191b029410f4e9da7ffb57a
758a1c168d3bc64d296b3e514b2706c1ffe1cc11
4273 F20101206_AACQME smith_j_Page_115thm.jpg
6694acf2c434b254368aae8971d6e462
0daf30f1c700c12979e2e59ce9a4dca612b149e9
F20101206_AACQLQ smith_j_Page_011.tif
74ee5bd8dc311600c53ef3f3c766cfa2
70a0d848b375b55cbd4a8b210750b75ccf9acba2
27420 F20101206_AACQMF smith_j_Page_079.QC.jpg
76704235a4730b4cce3a53186f3e8afd
574e91953da01f609aa5d625a6bcab917377cf35
2261 F20101206_AACQLR smith_j_Page_025.txt
64b5c798cd4cdbf27d4863f6bfc1cf82
f72c1e668ef40948c116fba31af8a1244035dd97
73368 F20101206_AACQMG smith_j_Page_056.jpg
e72e8d18a70bd7c3abe0606d80c9fdf0
17aad983bd0ccc65b7ad60a9809096a56584f85a
21707 F20101206_AACQLS smith_j_Page_047.QC.jpg
41aa1e4a07e5b127b840646a7bebbb58
6000eb3bf0a41deeed1041e2b66ab87dd8cc6f69
5401 F20101206_AACQMH smith_j_Page_002.jp2
b40736ffaeb92d8f15b42bdcdaded1ec
8de13697c6194e8758e5c026c9debed93f82d0c0
2275 F20101206_AACQLT smith_j_Page_060.txt
27d735ec05faf46c73ec245516932ece
9fc3d82f828a680a5bbcf51c1133c04001ad3724
43919 F20101206_AACQMI smith_j_Page_082.jpg
7711370a9c254e9c2582298bfd6b4a99
1e6527882bcb3aea8469c28ec50e4e1cd2ff5980
802111 F20101206_AACQLU smith_j_Page_082.jp2
b44d3137dda24f4bd80dadff08156f8f
193505de3a3d6379342f51fe6e5551c1969d4ada
12799 F20101206_AACQLV smith_j_Page_039.QC.jpg
e9c24e1140f4610f8e94bdb447d92044
b8364337b9dedf31a4b1df8849cb671c6aec0ca3
74346 F20101206_AACQMJ smith_j_Page_083.jpg
de208204f20ecc8f7f0f0f9a38b8fcfb
2f513bd64d5c55100df553768cd281cdf922708d
6055 F20101206_AACQLW smith_j_Page_026thm.jpg
e8a792a62e8be5c0a64f616b78263145
ac79c0260d0d372a6b6c2d8ba1cd11a77ca655bd
65535 F20101206_AACQMK smith_j_Page_134.jpg
f510dcdaa8b41ae1bc45930b3a36f15f
9ea3ee52e7f14b2ca1a471f5e0e1fa330cd0a745
1905 F20101206_AACQLX smith_j_Page_083.txt
f3d537c5ab50d7aca8ada87764059b94
39e0710b8e3d682ece960a5670befbfe9d59247a
6487 F20101206_AACQNA smith_j_Page_057thm.jpg
b66c1e7553686017c644d3800f61b829
9db37426f7fd4b51cef99baa06186dc3aed7b900
12244 F20101206_AACQML smith_j_Page_117.QC.jpg
f5704711dd4bee275753c12f992f2bce
1783489579b5690e14562bd4bcf0c9c66cf225cd
23430 F20101206_AACQLY smith_j_Page_081.QC.jpg
f25697ddde1b1103c4db0160eacd0fbd
5dc2cbd3de9e00701cb23477ca1e0b749910bbf4
72557 F20101206_AACQNB smith_j_Page_092.jpg
2276d078ba9d1325ac595f400c4a275b
5b90a0d1914f380055f22b7beddcff9a836ea17b
2344 F20101206_AACQMM smith_j_Page_059.txt
9de4bafed6ed2897517fdc6138c670a6
1c2ee81e015802296a7e8e652c1e2f6900e20041
F20101206_AACQLZ smith_j_Page_049.tif
ce2b314e227499a1d850313ef2e01b44
578b6db2458ea85efbfc963ef31c316478e06d7e
489989 F20101206_AACQNC smith_j_Page_023.jp2
aefc68a95ea288353de6724c45646674
1c590fb160326a3ac0470e533abf605965f0180c
F20101206_AACQMN smith_j_Page_126.tif
a3cb84bba54336bdfcc96bd157bbbd9b
b73cbbef108fd582e92f2d883326460cf763ed21
23092 F20101206_AACQND smith_j_Page_080.QC.jpg
6a56a7562285a5448d46639dbb380f2b
f2b771a431d6c6bd61db82db8dc5ceeec9d7ff1d
F20101206_AACQMO smith_j_Page_070.tif
45eed80dcafc748c52c3326d9f24b43e
346010ef24211fa6ab7951427a3bf51544315c7d
1051964 F20101206_AACQNE smith_j_Page_116.jp2
922bb8068562a2940c2bdb04d578c7c8
b10c1f0dbc83ddad8c60703700ed7e05b767c492
13740 F20101206_AACQMP smith_j_Page_082.QC.jpg
373e5498306d363165b2ad67c5885183
ce04a5d8f53ce2ecb7ab2005119d37c4af3876b5
14069 F20101206_AACQNF smith_j_Page_121.QC.jpg
caa2d88f66e23efc9993aa6b9ca9870a
df8d016f3bf25da2198b2246b67e2ae91daf4eac
82951 F20101206_AACQMQ smith_j_Page_036.jpg
06690ecb94d69b0b43f598a69d861b3c
e6c516b8971ea281082f7c39cb4aba3bab8ff8d2
6672 F20101206_AACQNG smith_j_Page_060thm.jpg
d69108d3635a9a69a55ae7b6e6ecc499
142070b6cde620fabd5f0bb41241981f79b864ad
949 F20101206_AACQMR smith_j_Page_125.txt
9cd0281b80ac822a532662eefb9af8b2
1c83420e624c115347e7eed4bc809fb337f14fb2
2446 F20101206_AACQNH smith_j_Page_048.txt
688bcbcb184313c2454da11dd026490d
fa05167f437a79c77457f3f0868c1b628fc52c76
1270 F20101206_AACQMS smith_j_Page_126.txt
48ec436df92d10576f4266b7643a31f5
d4d87cd3521d1b3a28d4721497cb8d6e76a94e7b
2264 F20101206_AACQNI smith_j_Page_031.txt
b1e0c8a87d63e857bf11e8cd4fed5a22
365fe405033f2bee8af5b0c76c52af24275b0eb2
F20101206_AACQMT smith_j_Page_062.jp2
f95ed88b61ab6a56343693cd63ecba8e
d6f642887d981c7b6c5aee121736f9484c52f11e
23227 F20101206_AACQNJ smith_j_Page_108.QC.jpg
0b6f131124b3f80fe513fff902ba5018
4506cc0add41568efb352c13e705f98b9a28ed82
6938 F20101206_AACQMU smith_j_Page_094thm.jpg
65a013b0e0efc526dc74310be455a768
94ed29b2b1d6e456fb7e7b825526931b998f3c0f
2289 F20101206_AACQMV smith_j_Page_079.txt
3e23672ac5bd7d3d2395fdf8b5a98fb2
81d3123557919ec6d74a7de943cc54582b5d9117
F20101206_AACQNK smith_j_Page_103.tif
e27fb280c599cc0ab14d3dd3dc4bb64d
4ef54c075336b0f738e1b70099d4aa9d3ba62424
1051968 F20101206_AACQMW smith_j_Page_051.jp2
2b312ee67906f98e3477e0fe1ce498f8
aae0431eb5c1d24e2b0601f8013974c238bda555
1051979 F20101206_AACQOA smith_j_Page_094.jp2
4daf438b9d836f6e88c76e9eb438ea38
2daafc0d6aea3a5a249cb8453583020ad5289030
F20101206_AACQNL smith_j_Page_097.tif
c5d917b1284d4c754273cd0dc6177bd1
e0b3ae70a001c8aa4f545c33fe8b9e3ad9cce13d
F20101206_AACQMX smith_j_Page_112.tif
11fff7aeae86fa7282b3c69def568c81
a255f03406d1fe864cb098e8699593a8c8d0632b
F20101206_AACQOB smith_j_Page_058.tif
db70c3bf3a8b108d061c8203cecae558
064e3caf7b13be0f68988ff0338827fc631cd5dd
F20101206_AACQNM smith_j_Page_129.tif
cd82e5728f21cfbc8e8f662a1b298ed9
ae9df2a95c2a93cc6117a48689a1e00d732d25bb
F20101206_AACQMY smith_j_Page_037.tif
93c1f8616ed8bd88f92e702677293604
035151450cb0acdca6d1caf9c8387ca032ed480b
F20101206_AACQOC smith_j_Page_128.tif
60961f7e8829eec436c39333c0f0bcbb
eb4f39abfbf4937f5040fa95511ccbf752a75bc5
4156 F20101206_AACQNN smith_j_Page_082thm.jpg
a5a81004d60f1f95947660c85f580178
0782f35448ed42381d8cbbd69b64be4d7ad84c7d
22115 F20101206_AACQMZ smith_j_Page_113.pro
9b029e6074cf5afb99dbb9b4d23cfa83
5cc62fab1028f3f76947b01d19fb7d365c740dfa
F20101206_AACQOD smith_j_Page_130.tif
833fe419f1e0733987d89e4494a38c95
5a9f269abdc52ec971e9e82729b2d3eac1486a39
87906 F20101206_AACQNO smith_j_Page_059.jpg
ffc85b2b0dd2c8c7af760255082592ba
bff036fe1ff4d144bfbf249cf112e84e10e54dd4
6052 F20101206_AACQOE smith_j_Page_043thm.jpg
cbcb520e3fc0771ab163dde64ec06a56
3f835e04283a78ea15d95babe60b25a229879edd
6238 F20101206_AACQNP smith_j_Page_015thm.jpg
b170a2f5379e2c5fd36e60340d226619
5c6f30301d3855081165888aca99f26332c5ff6e
14415 F20101206_AACQOF smith_j_Page_012.jpg
ad2252b3cc65267a489e1ab08dc0012e
3e3b1a6f3c9d5df1b51b2b707ea942bad33794d4
12214 F20101206_AACQNQ smith_j_Page_058.pro
afaa0bb30f45c4cbb5807ae9ac6cf521
e7380430df32fc3e9eb7f472500beb72e726849e
23507 F20101206_AACQOG smith_j_Page_034.QC.jpg
f766f4d4a4be6ef3b739b58e950b7e9e
91f2227cba10e6ffad2de3c551f46e7465ceb12c
52023 F20101206_AACQNR smith_j_Page_009.pro
59d7058955b2ea3039e2949c94d14922
e245d2fa383ae45174a08204bcf8672d09bb16e1
4999 F20101206_AACQOH smith_j_Page_064thm.jpg
aa6804fe2b5a8454a0af92a97480e179
1f0ad72a567f023efb3381b05d82867865da0dad
23725 F20101206_AACQNS smith_j_Page_125.pro
197c82032fd8574f1799d5e3631722a9
e6f9660977d67140649d9add3259a600c2a8a477
47319 F20101206_AACQOI smith_j_Page_055.pro
c105a963ca4723b71970797f80b07f24
001f9e31f9d6a4dd5f8755e28b03a249c15ce874
5021 F20101206_AACQNT smith_j_Page_006thm.jpg
67d09894fa775ea6f07080a950cae275
62d25df29e22ea7b146918ad0e1de4a87e5884c4
323 F20101206_AACQOJ smith_j_Page_138.txt
015a11b1a8c65bb48b6245a56b614c2a
92409303172110b2941d069895c77419f4bd6db3
39651 F20101206_AACQNU smith_j_Page_090.pro
e808ee1b3422935fbc0fef4796c35178
51db4d3cdb24c19ea8fa0470fd6285277bba1c4b
962165 F20101206_AACQOK smith_j_Page_033.jp2
32bc2e69b90bcd9c8a2b5633b98797cc
942b4e77cfc214de0c8bbe184e2aa9e5cd8af72a
32498 F20101206_AACQNV smith_j_Page_046.jpg
bc2c7fa773ce215bbcf8a1c601e5d631
680ebd3fe0d2830b7238fbfbaaf2b9b78136401b
2381 F20101206_AACQNW smith_j_Page_116.txt
a78ebcde6cb7722ff410e759830ccab8
d13ca165aa77e6e84d22d2c4eed6515ddb4c9002
6396 F20101206_AACQOL smith_j_Page_008thm.jpg
e0d5eee093c6883b8431d1e1519a9e7d
3c10913d1182e4a638a4eca43106e1b8ae4904dc
2235 F20101206_AACQNX smith_j_Page_101.txt
a324c20fa9bd845e38e61c9323b25251
7401c6dd0858784eea3712496febe6181a49c0ea
F20101206_AACQPA smith_j_Page_095.tif
1d7f8dec4d229a676d068a5230de9aef
479228604421c7cc8d136294ff35cfcdbe8f74f7
886009 F20101206_AACQOM smith_j_Page_127.jp2
a7ce090849fe7e3ca418350fb2c8224d
16fd1beccf33023e0a2c02c39686aa81e7ea6ae1
F20101206_AACQNY smith_j_Page_073.tif
dc8aca1d638be24230d68a00e8ec4a8c
485941de7f81cc8472b4203cc9ed135a4efc96d3
F20101206_AACQPB smith_j_Page_083.tif
d876b6a83709b50348d6eac467c129f6
778edbb65744e68ee11dae1607000395b4d54485
F20101206_AACQON smith_j_Page_081.tif
1c5d631dd4cbc0a6377a62c3d9af8500
31a398b1c6ccf79a2d12936a7f0b7348d169401f
1051985 F20101206_AACQNZ smith_j_Page_037.jp2
b922153ef54da83b25fec465e22998b3
d9e645e43be0188f80ee0aec756e5ee61f641d39
58944 F20101206_AACQPC smith_j_Page_088.pro
77ffeb25a4c6938e49465ea3aa4b928a
55fcc0c8082fb9a2386737f1758ac1e0e64c85c4
2115 F20101206_AACQOO smith_j_Page_133.txt
63da189b4c831823908c5f9f872d1ba8
7337b4c1ce47782686ef3054ab0ffda86d131416
119655 F20101206_AACQPD smith_j_Page_140.jp2
43f6d87b8a038a6f3afa43f36eb7915f
b5f677354ca30419cba897f8e971d23a95560d11
652253 F20101206_AACQOP smith_j_Page_089.jp2
dd70a650b67c0e01550e3c8350c065a7
fc0e43aff3b2ab284da47b564fe1f6496be6c856
1602 F20101206_AACQPE smith_j_Page_127.txt
92a81cb61187d0ae62ea17273d390776
30cc847e1c94878c8ae4538e877c7def01be5eca
5304 F20101206_AACQPF smith_j_Page_129thm.jpg
40b4100fd56f8bdc66d30bb6e04e03cb
f579857381a6850c9135881847a096b13bf76ca0
38771 F20101206_AACQOQ smith_j_Page_112.pro
8bf3f4cc5a6904ae1ae8ca62de630a63
d6919c7a4f0fcd106fbcbc0080d8f0b69bc09fab
56048 F20101206_AACQPG smith_j_Page_016.pro
13f77a6604c2303bf2460c209c6e0682
5c269ea519232e7ceb65d54baf704a13ff447eba
1999 F20101206_AACQOR smith_j_Page_030.txt
7f9b9629472cd4dc48236fa298c10500
d59e16c33f6748b090d7d2616a96bfb79637dcb9
F20101206_AACQPH smith_j_Page_036.tif
fc8d4fa3ec36276b89433be0124cf40d
d315cbe8dc8928ae95cebd42575c8dea2738aa9f
1051928 F20101206_AACQOS smith_j_Page_056.jp2
6b0efeacc693aa85bd5d06226987a48e
55ec72529f91c70aecc562bb4f25e70b30c54f27
19688 F20101206_AACQPI smith_j_Page_138.jp2
7969f0547433292dee0c1bae1da1810e
2ae3601fd7be98078082428839aa9bc096a51c8e
1014335 F20101206_AACQOT smith_j_Page_100.jp2
e0391229f674105d817478ec0214902c
c92e6d1a5939a9a3cb3aaa25d26ee75a6a937bd7
22681 F20101206_AACQPJ smith_j_Page_051.QC.jpg
8a66ba3af2f579e3ed03555f463f0894
c74334d04b9edc4c0bf1d5c09db923995f37f427
1051981 F20101206_AACQOU smith_j_Page_059.jp2
6ded466709e16546b077fec184c0c28a
267e739291fc7cb989facd59f2710d353cf85a16
2150 F20101206_AACQPK smith_j_Page_032.txt
2582310c4fa1a30f5b54398f52ecdc9d
ab63a5131e23ed6dee741bb47b70ca4b40ae9add
40532 F20101206_AACQOV smith_j_Page_109.pro
b301b3ecdda0be1da11dcdf26c40d76f
cd92f8a48592c884c2b7718e05f0246de9799dda
2204 F20101206_AACQPL smith_j_Page_122.txt
cc21418563112ae49693129090c59df3
5798764a42947bc51944f570fb4e0514af5cb3d5
40596 F20101206_AACQOW smith_j_Page_065.pro
f179704d28ef46ad6e366910cb206a58
caaaebf738b76d70c632fa8e8e1cca6638280f1b
F20101206_AACQQA smith_j_Page_020.jp2
4da0a3fbefcce26503b5d0869b7d591c
0a96b25fb38e8afaa0f594ef9f6125b0a570d9cd
2752 F20101206_AACQOX smith_j_Page_010.txt
83bc6626deeae8736efb5af4efe82354
314ac537eb8bc6abe90cc35f2a31088016249d56
501111 F20101206_AACQQB smith_j_Page_046.jp2
538c59e4197ece0aabdf9d59d6ed01ee
6308fca1b4f427333e43e40131d71ae0e1f23249
54120 F20101206_AACQPM smith_j_Page_137.pro
be8686444d60e58ff50cf25d0ae69390
ae9b54c5bc982293eb45c52ffa5486df94b21aa1
16415 F20101206_AACQOY smith_j_Page_123.QC.jpg
ee38f15042cc96b4b90ef6ce03ce1621
2f31f83fc80d68a7737695902c03695f54f2262a
86638 F20101206_AACQQC smith_j_Page_103.jpg
e55948232ccc9455e04c0409e4084c92
261af4c375e376ec4a85c261e760a30c38d5cf6a
82157 F20101206_AACQPN smith_j_Page_031.jpg
fa134b54a6ce273ec98bebd40402f4de
756d2e60172730867734167e29aea20f3a874eab
55156 F20101206_AACQOZ smith_j_Page_073.pro
9fd15703ecdc085145be6f52cba61bc8
a959637bb91ec946afa53594943ef1999da3a9ec
50385 F20101206_AACQQD smith_j_Page_030.pro
c337a257f53c027cef29700b49d4dae7
899ddda60213ab33b9f6a06dc2d40e92ef058091
6589 F20101206_AACQPO smith_j_Page_084thm.jpg
860a9f22c7b5fc64c001d5208570487f
6c20b65cf1547852338ee7fd3520ce946a7a5802
1631 F20101206_AACQQE smith_j_Page_065.txt
2013eb77b741374dd4a747a7457277ad
ad834f6fef62c1fa6a154d40093cbc114ddd7c0c
27864 F20101206_AACQPP smith_j_Page_021.QC.jpg
f92a8cd76accd0dc3aeef50b77141179
c493b7d7c19ad592f9bed0b571d41f3bc98f5f99
47547 F20101206_AACQQF smith_j_Page_054.pro
e5ce01440b71692f5c547e198b1d65b1
61e5852805042fd4e4189dc586267c5b11396aa1
F20101206_AACQPQ smith_j_Page_027.tif
64fdc9d8c7cc5f2708d7ae50db9e973d
ec7955f9ed05cc42875c10af3e3447836f264e70
8361695 F20101206_AACQQG smith_j.pdf
5723c0da0f6981602ef60469ea2a1412
d0b5346f1f66fa56a15b0a365e857e0b4e840e4e
40137 F20101206_AACQPR smith_j_Page_127.pro
c19c1e5112f3b8be76593f77964ae8bb
1a8dc01d1e735dd11be8dde8e1129ae895e0dbe8
26913 F20101206_AACQQH smith_j_Page_009.QC.jpg
8c2271b7882875fdd3598606445a8aad
55114ebe45a839aa31760d7498adbfb8d0b3efc3
28518 F20101206_AACQPS smith_j_Page_116.QC.jpg
d06f181f70be9fa22b9e2aee5b7491ce
ebe989408d04b201a8fd6b2d363a05bf857d0c07
10564 F20101206_AACQQI smith_j_Page_089.pro
bd14591227543e603d752b44e77c98fe
8412a8cb174337715121fcf99ef80b5976300858
2193 F20101206_AACQPT smith_j_Page_114.txt
0f7668022023f66132aaa1f3fac943f8
92aaead6284d1e71e811c320d14d47e2aebf4231
92571 F20101206_AACQQJ smith_j_Page_120.jpg
13d0970621216c4c9eeac1a7f381e92a
d8f956c2a85192499aa88563051d732a449a863e
24178 F20101206_AACQPU smith_j_Page_068.QC.jpg
fffbb7812764ccd5d171f6d757a07cb3
a135d67bddd1dad3e230ddb9aec4a473d96a30be
69284 F20101206_AACQQK smith_j_Page_093.jpg
6723b273121e714bcda0bfb3535fcce9
5c2ffdbbf24cbb17546ff1515f33d8ecccad5930
F20101206_AACQPV smith_j_Page_034.tif
eadaa178d3861e0a5791b0bb2e6ce81c
547e6e193ae5a3ce102ccdf155d33a09d3ab24c0
23097 F20101206_AACQQL smith_j_Page_091.QC.jpg
f6e39bf78734c428f4e05b50d1f2bef0
2f3ce27bebfce807e6f7626402cd18830936f89b
15724 F20101206_AACQPW smith_j_Page_063.QC.jpg
803d1cd30bc719b33eac7da95253e717
4b0bbb594868d5a0e5401b8cb8bfc3a57adfd5fa
1383 F20101206_AACQQM smith_j_Page_014.txt
89abb663a75882d457225f97d0705aac
2a0919edb1711461766184f2edf91f08b7220e98
89638 F20101206_AACQPX smith_j_Page_118.jpg
408591f77a2085feedbad6a4460f13cb
ce00c9b070a454723a18c361ec8053bfb5d75f81
F20101206_AACQRA smith_j_Page_094.tif
dc66dbfbeb4bd9f90098cbd423d1b530
7d06cb8e06afd5c01991368f94c6ed3fae566b55
53743 F20101206_AACQPY smith_j_Page_101.pro
00dbe1104386de3bb91e250620ee083f
884e59f54ae7654ee3aa10190f309e7a1e25df1f
2379 F20101206_AACQRB smith_j_Page_004thm.jpg
c97f295d295257eae24243be3ef8947f
2e766bf01bff24562743e6af4cf64f964091a3d1
745503 F20101206_AACQQN smith_j_Page_115.jp2
e610b4c936e3a6198c675cda60e0bf69
632760f4aefaa9db174dd7c336117c9c7e1bffce
80359 F20101206_AACQPZ smith_j_Page_101.jpg
d7c6a814495a79ac6f753358157c92a7
b89f2fb9d40390af521b80f5b13be9e1117b8696
9975 F20101206_AACQRC smith_j_Page_002.jpg
748f0db7ecb2a180eadfea1cd4a2b3b9
7127aa307c072f5cb768b0412e3c0fb1ee2e1ec5
F20101206_AACQQO smith_j_Page_070.jp2
1b03e340b5a5ea9b40ccdfd3d79485b5
1189d6a0881a684b1f67c3cca0ab7d2aa9a6a7f4
20656 F20101206_AACQRD smith_j_Page_033.QC.jpg
a611919d3fc1fa30040c6d078f9fb5d1
bdb3105f88799ffe1ca0bcb4002389c1ab6c267a
108967 F20101206_AACQQP smith_j_Page_011.jp2
c2022a89036c8f243e82eccd217afe92
3829baab25d151b0d03bdc5a15301762a1cb0a26
43391 F20101206_AACQRE smith_j_Page_034.pro
114e27102afe45807171c58973ea4f45
8dcd23787a21c02b3b1138c69868aad92f76c717
5962 F20101206_AACQQQ smith_j_Page_067thm.jpg
c64edcc61a23d62f501304dc0f779ba9
85a1569810e79c24b759bc77c9d26b60907f6ad4
6248 F20101206_AACQRF smith_j_Page_056thm.jpg
af3fb050b7c0f17414c9b78814d06d7c
5edf4cdace7136a2d2d17a23e5cf00b067a9ab01
13480 F20101206_AACQQR smith_j_Page_004.pro
07104a46a1528554e0295c9348557592
d0d745b4e9bf20a32fbd53b2b9c35a2fdf0294e5
27759 F20101206_AACQRG smith_j_Page_070.QC.jpg
14715539762ec33103e76237eea4542f
40fb63f0e66c1384ee75dd8cd4676e232e69c92f
90717 F20101206_AACQQS smith_j_Page_028.jpg
28cd57ad8d1655a3f20ec599bdb7a6b3
74eaf9bfa40fc173361c5e3ff711a0a2282679d1
56507 F20101206_AACQRH smith_j_Page_025.pro
30732f3682307147f7af4a6a55d378de
917b637f94032429531b924db1216cd374eb41ce
1051934 F20101206_AACQQT smith_j_Page_106.jp2
c8a91927ee88123a3a76c80f983c694b
31a4ad7857eb79d54f16dbdd48eaaaea76311a1f
8009 F20101206_AACQRI smith_j_Page_004.QC.jpg
7b3a48c1922dcb1a44040f2df09014a0
0dc36ae6a490c3a11c6884a2cb5cbe5f3abd2635
1051980 F20101206_AACQQU smith_j_Page_124.jp2
af9b71331f5864374476f9cdcc996723
b9ff2eb9a0dd8fb2e8a2627d531d789c2ede2335
56644 F20101206_AACQRJ smith_j_Page_099.jpg
d9a69ccdc039bd5dfbd5a0d47d53ed25
a577f304f35eb529a3c0b3bc85057e250d933538
6442 F20101206_AACQQV smith_j_Page_055thm.jpg
dd8c0688c96ca3b5c43f0b67237d8885
3ea657338c676444d2ad985b464dbba1dfaf28d1
F20101206_AACQRK smith_j_Page_033.tif
ff086817f5a02a638e35d5f5d094b9e9
2371e90fe303cac95ae172b6212d83fb474c8604
5808 F20101206_AACQQW smith_j_Page_100thm.jpg
c9073c465c6903428d9e8f4d297ddcd3
4c5d1f89982e8e75de53d35277604a19c5dcc51f
1051972 F20101206_AACQRL smith_j_Page_040.jp2
c6d01f301158429783641822a4e08cbe
b3c801225eafeb9d2e7cf2c7a08b89a29924fe02
5929 F20101206_AACQQX smith_j_Page_133thm.jpg
f0d0f99472092a3da91ec72835fc8165
13b3f169de9dcaf91133e747e6b9307dbda148d1
6721 F20101206_AACQSA smith_j_Page_021thm.jpg
119a9bcea7845f1d116de263859be04d
bdfd022af266f97e4319586cd42ddaad8545b216
91080 F20101206_AACQRM smith_j_Page_022.jpg
c4eec76586dd0f6e51fbc757e32c123e
d368a180bcfa9f04910d08457086c40485c47062
52222 F20101206_AACQQY smith_j_Page_015.pro
e7338f70286c56a52e479cdcec2163f5
2b7600dc371286e3244fa5926c772f668191a00d
1051971 F20101206_AACQSB smith_j_Page_120.jp2
162bf7b89538c3503a2d07e9524a44bc
d60df237917f91ffea98957fe368483c57159f87
44739 F20101206_AACQRN smith_j_Page_007.jpg
210662c1d1154837e539cb3bbcccc55c
3ddf0f47d034b29f52371200e78cb79d4a8ec2a9
F20101206_AACQQZ smith_j_Page_048.jp2
8221f01c0d9c3eb3316395f4a103d0de
1c3ba37d2387eddd52971a0227f0432b17734ce1
48400 F20101206_AACQSC smith_j_Page_091.pro
f0952c529a901d4ecfaccc3d2da2ab17
f55843eb72356b6c1dcc5d245a9a340732e74a90
5061 F20101206_AACQSD smith_j_Page_127thm.jpg
952dc6fd553115c0b3e79b039e191f89
87760465840136067d1a4f09d44ed6c9b9c4e55d
27867 F20101206_AACQRO smith_j_Page_105.QC.jpg
39f74bf8d468818ed372ed93c432731b
e6ca6b1a7789c221df46a7d3dad698309b4006e4
1051959 F20101206_AACQSE smith_j_Page_034.jp2
07518593bdc5e6221a3653d2dbce9f05
b3118f66e2c29be6f0398347fb681cbfa6f29c2d
6618 F20101206_AACQRP smith_j_Page_085thm.jpg
1bbad2f38cbc0772661c615bc24a751f
55b96ccbf05918f3fbd73eb9e52bdd8a57c71216
6569 F20101206_AACQSF smith_j_Page_035thm.jpg
0992be5f835c1bf8fb513dd3aaa3be13
274abadb699ccbc237a4ef77e44f35adc4f186e3
F20101206_AACQRQ smith_j_Page_065.jp2
f6e60e4b3a4e31a9e9595dc9506c9f40
81c68a4a8748adc0b83bfd2c0fbf6b182dd0db09
107205 F20101206_AACQSG smith_j_Page_133.jp2
a61a6b57baf33e1aaf505bfb86edbf9d
25bcadd635d90aea75c12c9164d93f529274ca6e
76736 F20101206_AACQRR smith_j_Page_055.jpg
b1c9870991e5fab0b207c73ecd94d73e
5c328ad5984ca5e0b29a7611643760a3660d7110
23782 F20101206_AACQSH smith_j_Page_058.jpg
ab2b98e9b92ade4329fe8b5f8511f1fd
7c769bde70dcc92f885e0341ffb6e0e2b7c5537c
F20101206_AACQRS smith_j_Page_088.jp2
7168eb57f816d94ea531bfb932511c89
cd8a98e43056d271e3d349ee6c1ab57119424221
5561 F20101206_AACQSI smith_j_Page_102thm.jpg
2c6592a3c0b853f9f460a1063a285837
288a75f6e35892b0d9b329dfc7c7bcd360be319a
6928 F20101206_AACQRT smith_j_Page_088thm.jpg
44eaab57a2ac3a2efe3f8e5b57d22db0
63694839c40bd849ee77ed525403cd785165374d
22766 F20101206_AACQSJ smith_j_Page_056.QC.jpg
026e3fa33fb9586d17a6282b38b29393
4d41c34bfbf8e84e6056d4ca6c4d401ec817d153
6468 F20101206_AACQRU smith_j_Page_108thm.jpg
8152ab8228fedc7e1235044895fa4a4f
46f88f4afd6e29f2ce3cb068ac2e6dc41a76a992
2414 F20101206_AACQSK smith_j_Page_120.txt
d9e662e1d393bb3ffab33cff1c2df3fa
b6602bd3f3f8d4a899fc9045ce06822d54a496a4
26963 F20101206_AACQRV smith_j_Page_106.QC.jpg
1e51a823db87b2e229bf90242e966917
8050e28e6ec8ec6059873276af51c2f1a6ca48c1
27742 F20101206_AACQSL smith_j_Page_060.QC.jpg
7d76cec1704bbe6f5a5ff861092266c9
4f4602f535bfcee60f2fbf60a36a3d9d1537a531
29823 F20101206_AACQRW smith_j_Page_051.pro
a2b9e643b6f80360f07f3cfbf04a37cd
393182bb28f308ce6cfd79d96450f21eab32c4fb
12051 F20101206_AACQTA smith_j_Page_089.QC.jpg
03f2ee2496f6f8e68d27bf62a855e0df
b0e7ba08df4ae38228cffec9d07e320155a81c5a
2466 F20101206_AACQSM smith_j_Page_005.txt
7de6ad52633ffe1735753573cefb41c0
fa69aa27189ca3c208f7a6ff2f892157ffc8d3d4
17644 F20101206_AACQRX smith_j_Page_131.QC.jpg
f5e81412d2665765878a943fce5e872f
a9dc12709325070d50d647377c10f6079800911f
6440 F20101206_AACQTB smith_j_Page_112thm.jpg
0f37a9a41c2d041539f786bd8761a289
2393f43af662ff6e322935b97d220d3adc98d133
1051939 F20101206_AACQSN smith_j_Page_036.jp2
99e33285a2f058e4c116167517dfa4e2
e4a43ccd82759f4e6293e9f817b5667a6b64d9aa
6226 F20101206_AACQRY smith_j_Page_034thm.jpg
4186e2e96d7eb48efe32ce771e3e179f
b986c2c0d454dbb0dc5ff7ddbd1500495c75a228
4051 F20101206_AACQTC smith_j_Page_039thm.jpg
302e38c3c45a5f010e69f8d2e3c0aeaf
7faea36822f13c9336d4679450ed960ead477459
89511 F20101206_AACQSO smith_j_Page_018.jpg
923cd9da84e17467acc7fb3dda919e6c
5b713a57a0dd9c779835f1a64d95e911d9a45278
F20101206_AACQRZ smith_j_Page_028.jp2
27516a6717135f9ad188fabece66aab5
61c3cfdeb0528020b5689c6321c92ef9be81ef63
69013 F20101206_AACQTD smith_j_Page_047.jpg
baa06435606caa52e10243ddf1c39623
b79ba4ce3d2ae2a9dae870891510339e23614e11
14858 F20101206_AACQTE smith_j_Page_117.pro
7fd273e6a5d8c96425c48d6f45493972
af86cc7234dc0d010354a9b1c1b28d5e6c4dd5a7
F20101206_AACQSP smith_j_Page_032.jp2
dd7a6f04670de06e4510b6f683117a69
a88f18dcedfb703b9f36318a3fede976d7789489
F20101206_AACQTF smith_j_Page_054.jp2
d549ca736362ea40886eec36c57e1b31
dbeda507076565a6187fb3a1e639f0ecd8497791
2142 F20101206_AACQSQ smith_j_Page_130.txt
ae791ebdc4d8f36e5e2dd63d526468cd
9fefcefd634f0160ea633904654302b686435565
6714 F20101206_AACQTG smith_j_Page_074thm.jpg
e44bbad7ec783828c6aefcc82f78816e
ca94429ca4be2d06f352cb127dfb6228bf255eef
36101 F20101206_AACQSR smith_j_Page_089.jpg
fad80a1bbee2054ede8c872931cf5456
0b5a6872791a97451b7a4eebf9399be102c9b2df
3767 F20101206_AACQTH smith_j_Page_023thm.jpg
5e16f082c1fb2113e6033a698d716ded
43572d4e9f73249137e9e8623b2f7997db717e87
41170 F20101206_AACQSS smith_j_Page_115.jpg
d24d39d1ed8b2356900c9f36a6421022
20d06ab70474ff6c13d6e9152607ebf773ed7e40
F20101206_AACQTI smith_j_Page_079.jp2
3f5b9bec6c49c043817d135783693d39
4d4e8b26136d518a0aa238b80e47dba7ac51ccc3
F20101206_AACQST smith_j_Page_061.tif
31dd4b50ef2afd54b91d15479a42a962
cfa8ff24c70f88095d01630a87cde1556fa8601d
3548 F20101206_AACQTJ smith_j_Page_003.pro
04000c894204d2c6a16c15e4759563d9
46532f7e331673781c9444e19154bd932760f963
6611 F20101206_AACQSU smith_j_Page_130thm.jpg
d0bd79c874b05dd44bd5cdd990680e0f
00df4fbb057a0caee05a2687d4095d99f0fc6a28
14207 F20101206_AACQTK smith_j_Page_111.QC.jpg
b46d019a3bb196da51f612c9ed5412c3
c8ff40790888ae4ef875787e92a557fff70596d3
6684 F20101206_AACQSV smith_j_Page_017thm.jpg
cad4b616420e85ce47038716b2fa4dd9
54339a47cd6a36fe9a460cfaca4f276889bbcd2e
72934 F20101206_AACQTL smith_j_Page_108.jpg
09d384cad455184452f7c13edab0dcf0
1068cb05e1fca601c1ddb1da5fba1273dea4b781
27706 F20101206_AACQSW smith_j_Page_057.pro
d4ce40d1cc14e8bb73916589a612267e
b0f1bf6d54b7ad24dd66509aa64b9f36a9e016cd
F20101206_AACQTM smith_j_Page_099.tif
4e5b39363ec96479e10ab18b7efe6591
37bfd8a235ae2c72334ed348a53fb53801b99e38
6625 F20101206_AACQSX smith_j_Page_053thm.jpg
e1780262cf4e7278a343f863e4b4cf20
7532b8ab3f5aa3a7d793164249e18d66811c5431
27293 F20101206_AACQUA smith_j_Page_017.QC.jpg
da02a137f7351135f1f6c89f46b63e86
3b9a00e56745fedb6ca1d5408486de764acee0b2
F20101206_AACQTN smith_j_Page_008.tif
57fa898f7032db40cf198544bccef546
ab68f5ddbdff0e9be0fe09633f44bc24afcfdd30
50909 F20101206_AACQSY smith_j_Page_011.pro
2c0bfa94ea458f3b3de65fd64b1fc093
257759840e372eccf14d636c98a4c4575710e634
490 F20101206_AACQUB smith_j_Page_058.txt
ae70026a3836d220bcf033532e84637a
7b9e5f1783b3345ff3765a33cf32375f282a3ea3
61752 F20101206_AACQSZ smith_j_Page_029.pro
9e775095151026c5791f28c965f48d7d
d146685fdfbe61ee10f322bb1f55193cafd6c810
37570 F20101206_AACQUC smith_j_Page_033.pro
d75fd6ee07bfe2b28ed8dafd359a00a3
de9a969fda2d6ebb37f52b71f320727bb0b53cb4
41190 F20101206_AACQTO smith_j_Page_039.jpg
d411de5c0b2c1e07a10edc658240aae8
cd923bdda82f67c844d575c5da79f2112bd3d5c7
F20101206_AACQUD smith_j_Page_018.tif
ec52181fe546d3994aae9ad2d64c44a4
e9aff297eefddcb8db6bba0487141fa8f36fc05b
84343 F20101206_AACQTP smith_j_Page_114.jpg
65635fa65b6bb58dffc4805c6c9f7bb2
cae05402a8a363fbb9a31ceb3b6043756c69a880
30025 F20101206_AACQUE smith_j_Page_049.pro
8a53f185a3aa226442d6ab7876aa8b60
b6b1ffffffa82c05dba9ce06a63c06db42b25806
4226 F20101206_AACQUF smith_j_Page_123thm.jpg
d852dd5821d6598f83df3b3a47457ed7
c38911d3e10b007e2f25b753a96b56e5e6f9e80e
56689 F20101206_AACQTQ smith_j_Page_059.pro
2c3cafc61a4bc457dd0e182a3e4f2899
c331c3774eba0326ffb69c218cfba61d56c3c38a
F20101206_AACRAA smith_j_Page_017.jp2
f82da98d0d6b47ec284781735f316db5
9aa3892d05020890f94c0059e58917056a18b3f5
59715 F20101206_AACQUG smith_j_Page_118.pro
0cca4d73a87a9632f878e2f36dcf3652
15b1350526a8808d7d1a7a2c1091c308a5ae7bf7
21768 F20101206_AACQTR smith_j_Page_136.QC.jpg
2263e22a97bb2ec72f6722d31c5c47a8
d2acda513f262e55c69940faf32283b0a4d92194
6505 F20101206_AACRAB smith_j_Page_062thm.jpg
5364f9919acfc1f0cec4e87ef23de7cb
4db23b217e507663f1dd9591693164d6ffbf8314
88402 F20101206_AACQUH smith_j_Page_019.jpg
164935a774390eaf206a99644def7cbc
902e7a2b224f8a55170c72dba6e296c21fa49416
26568 F20101206_AACQTS smith_j_Page_073.QC.jpg
bbc470be6ae2d411f12feba4eefae941
38f2ffd6f0524e5481b378c705a565794fe1cf0c
102809 F20101206_AACRAC smith_j_Page_009.jpg
04cd04925954bd0b0f7852cb54c05ea7
d49ea8db7e4334c6f79f9440ced46ef63f10aabe
F20101206_AACQUI smith_j_Page_074.tif
7e75d4aeddd23d83a349031f3d805a89
69ec8d3c35e09e232d94a3b5963940ffa9d98b18
82896 F20101206_AACQTT smith_j_Page_038.jpg
bdf5bc516cd608d4e3b529eee18e3bbb
12b2cbe9a5631caf8216668c0f71d79ef11e7853
F20101206_AACRAD smith_j_Page_095.jpg
8acb0ca1661ab18bf93f6205487b5d75
8c33ead23349edbe2a43f1b643a83c58e0cf9460
12676 F20101206_AACQUJ smith_j_Page_003.jpg
c7479a895f7dc319096eebbf6df6c8f6
d853dc8ed8e4cc0aa775ae355bb1ee2567d71fa1
43199 F20101206_AACQTU smith_j_Page_047.pro
e5ca5c50088be7dd4b6686482ad15de1
89ec04f55fbeb05604b699aea91e58bf4473781b
7005 F20101206_AACRAE smith_j_Page_018thm.jpg
58b9e66a321721d846b9f86a2f0d5a34
f44f53565e7afdbf16485ebb70b15d4769185189
2333 F20101206_AACQUK smith_j_Page_075.txt
b43a9c977019a3e0a1890f71f4781465
cd744110a088adc33d93c5b08048b84dcfc3681f
1048 F20101206_AACQTV smith_j_Page_072.txt
1da9eccd721126a9bc58774ea83bb80b
556ac711904622f6a59e3bc4ab5ad947d0c6290e
25954 F20101206_AACRAF smith_j_Page_016.QC.jpg
6ce5cefa38a342a8e4375d11bbf36ecf
9c3f0f85545a510b7e5e7c56e7814b68d595700f
2415 F20101206_AACQUL smith_j_Page_013.txt
2f2341b930ff32de6093adcad6914729
28c95201b9246b8986505d0783cc7e12498d3aa9
1921 F20101206_AACQTW smith_j_Page_138thm.jpg
2956ca4a6b61108379ea8d8b94b11de0
48abfa090a45f02a19df831627ede8a0812238d9
1051983 F20101206_AACRAG smith_j_Page_074.jp2
2735b7e668ac44632ccf15a9b0ae14d0
9f2107bdea8ababc33930a593d8daa75a1f21c0f
F20101206_AACQVA smith_j_Page_107.jp2
d9d5396c59904aecd7c1c66bc865bb33
b8f2619d7561c78739bbc6f352f693d991be7a75
26582 F20101206_AACQUM smith_j_Page_035.QC.jpg
a34791fafd7db6329f8bd4b62730bb7e
2f3dd0973682aa650459b3fbde91fb4b15feb868
F20101206_AACQTX smith_j_Page_092.tif
970bfd45383106f4216dd8ad11c582de
4502eca903d6d4ea8ea360f5f9c33c1f0d28e2d4
7264 F20101206_AACRAH smith_j_Page_071.pro
48110facd59abd3ef1e1654debbe16db
30ef87c711a4c22ca84bae3e958e9899a821b312
50523 F20101206_AACQVB smith_j_Page_123.jpg
7ccd8e07748678c719dde13185ee659a
b291e9c73d8421501f9b54118e23b6d6d0635c21
F20101206_AACQUN smith_j_Page_060.tif
fb66f46bcc6cc3b342ef5410faba6379
6c750ac90999caa09c73939b9630e7a7cd129a05
31126 F20101206_AACQTY smith_j_Page_126.pro
e377d4fc5768b30e02751207358edf78
4eea0781e063a67bfc901f410d52cd51c5c63d97
88365 F20101206_AACRAI smith_j_Page_094.jpg
70fd49c69a3b750ae551f36a28d639c2
dc09ed972b87ddbb2e2f07f785049ef6c1a234e3
523 F20101206_AACQVC smith_j_Page_071.txt
0c5d8745f19d9deb057df68aa0c9765f
931276fc9f5ce60cee3c9a151886a52eeed741a9
F20101206_AACQUO smith_j_Page_006.tif
60a386a9adb77d53912ab6943b1b80ee
0cf084bb38988a55b38acddee1ae044b4c14ff09
77747 F20101206_AACQTZ smith_j_Page_020.jpg
3bc8ea47ad8d6e131eb1bf262709d313
58974e6c82833509f352723fb49218b207e8265d
6465 F20101206_AACRAJ smith_j_Page_031thm.jpg
70dbd7ffd1ddf9f9a3c2d1d8101de241
3a709c4b1cb7c7b5ef253d391c3cc9ec6cd79ce2
1829 F20101206_AACQVD smith_j_Page_068.txt
4453063fb22a59f166d4b8e633a488da
ed659df9ab8f1cc2224e6fa7bd930518e54736f7
73813 F20101206_AACQUP smith_j_Page_005.jpg
3d37d8e2c3424283a1336ca4c9de30aa
a1c4eee20946d8e785b96e512ae2c2cd26ebcaf7
3408 F20101206_AACRAK smith_j_Page_072thm.jpg
b684a91b7c0bb8ea3ec276d3b32bb93f
e7c0ed3a2c1a4293863e3f6b037076a84351ca8a
994545 F20101206_AACQVE smith_j_Page_109.jp2
e68256a23bf196c85ad454ed8ee91620
54996183bd2aa0ab9e783b949334c1d3e54cd34e
86555 F20101206_AACQUQ smith_j_Page_025.jpg
3eda9a8dd0ae8bf92da61fce4133350f
f9879dab9a942cf58799b2785722555c4cdc52cc
6317 F20101206_AACQVF smith_j_Page_098thm.jpg
8a02b0a42aebc1df0d5f99836ec95ba7
0653e341a513538f74395d25e3846e1e5608efd8
21291 F20101206_AACRAL smith_j_Page_090.QC.jpg
b12d7bf8638d0b462111f22625eef565
fda7713e67908b8ecfcef39ec391b39bc65965a1
1861 F20101206_AACQVG smith_j_Page_108.txt
11a1d76cfca7359488a23676281e0752
70f23c6b980ef0c2559bcc7b0fdbcb31a51a861f
815939 F20101206_AACQUR smith_j_Page_099.jp2
eb1cd78520af2597127db2baaf2ef471
779f6fbf6d5b02dfd02084aae25e15e3af165ec2
72538 F20101206_AACRBA smith_j_Page_081.jpg
6ca182a0c7dbeaf37432b43c377b4667
5a2cdd61f22c7f0ee296c415d93c8dc73c17b80c
24712 F20101206_AACRAM smith_j_Page_001.jp2
cd06f3328987fd6a79ae4000359d2e52
d5f8d38a7166fdacc3dfead41d18f494d5f9bbe1
74283 F20101206_AACQVH smith_j_Page_080.jpg
d01db22db611a38b10b8f77f7747f6b7
81e66b0d1a2dfc08360b10ca935cc60e4eb0e0f8
F20101206_AACQUS smith_j_Page_053.tif
c04b81c968a2347447af9fc67bb83a9f
10a1c448ffcfbe91ee0818e86311fe0b0ad63c55
45680 F20101206_AACRBB smith_j_Page_093.pro
b8e566bb9b0d2d0654876392002403ed
3ca9dbb0e7236732d07b255d50bd97e38d904b21
7018 F20101206_AACRAN smith_j_Page_086thm.jpg
55e1446abf0f03e0697573def8c65ee4
f1721ef82efb26c747d3d5392a3e5fb3da339489
1051902 F20101206_AACQVI smith_j_Page_108.jp2
340f3f27838e5835c1e9ced44e3a46cc
d52d3abef8e94f6f82c1fa6da0763f9326b2429e
F20101206_AACQUT smith_j_Page_059.tif
eaa8fc0e19d318b10e5213c3cc33caf7
4d84615dbe2c963230241c3175005adc12d7eace
F20101206_AACRBC smith_j_Page_119.tif
c4b989b3d2da02684a82316f7fb759ed
0ec3feb8a1e65f0a33cec3bcd927a8a5e85546b2
278870 F20101206_AACRAO smith_j_Page_058.jp2
bb95d1e8d437087339b753a126245383
4deec33dd859b0db3220ece07aae5e5b88a5e110
46447 F20101206_AACQVJ smith_j_Page_113.jpg
1536cbd71229b4edc6bd75220d587e90
d7357bd59727a03156a6cd5e568ec47b3f06f8b7
23392 F20101206_AACQUU smith_j_Page_037.QC.jpg
7216298d6caf98e6605649d412dad388
8759b6888ea5190b7a44fe28689011276a027cf4
19943 F20101206_AACRBD smith_j_Page_127.QC.jpg
4b1906ad0c8680b5f7feb7a48d4640c9
34be148b51063f4b62254b07ac7e9b30cab56688
27374 F20101206_AACRAP smith_j_Page_018.QC.jpg
b1f039358f00123f0843132c5db41f05
9897bcbd2b5622abb2c7e77b48664eb1bd35c2b8
15540 F20101206_AACQVK smith_j_Page_119.QC.jpg
5caf5c553d9675fba15c62ef9434eb0d
4f4a97504c1f087bf80ddb58eb7d9f811e5b01ff
76406 F20101206_AACQUV smith_j_Page_140.jpg
93c158cf3f9aafec450e437dc2afc073
347fc15412251f63b48cba3eb1284c15592b2385
11589 F20101206_AACRBE smith_j_Page_087.pro
f16a3758210dd55324bdfdb3e29e1351
ffc8a248843def8b5741c1ecc480e9e733193b14
65514 F20101206_AACRAQ smith_j_Page_078.jpg
73b29afd8857ab99f9e6f751a8b550a2
5fe40998fda82f576a2ab5710391323b5abed71a
2391 F20101206_AACQVL smith_j_Page_022.txt
225f26149419fa01ae408fe7b4fbf041
82a4a64d07717f856fce4120a7182662b382f981
F20101206_AACQUW smith_j_Page_111.tif
9c139494a6bd8a0692062063b22d3ae5
5742669e15b4ccb2e425e74e5f12f577a869c3c1
F20101206_AACRBF smith_j_Page_055.tif
e7cb0204d776840ba5b5e262f4cad3de
1ac59d502abfb7156db880ec3b329476d4610462
23887 F20101206_AACRAR smith_j_Page_015.QC.jpg
bab5c737a7973a78415bd27e7ab7d68a
f055619c33fb54951f7a9d728ab6e46799a9e7b8
56998 F20101206_AACQVM smith_j_Page_036.pro
41c72d8ecb328acb5086ce890c6656e1
536564a7ea5e011b6ecc21e7dab15f9c97b0d0c5
6178 F20101206_AACQUX smith_j_Page_040thm.jpg
30ef5446a49d2f1042d5a000fd59d785
e7623b74192791cfbe3c2f019154b7c0de2f0685
75858 F20101206_AACRBG smith_j_Page_034.jpg
25f703073632d342b1e32c4941be1e7a
1a29ad482fd881ddf0e84e841458219863a7e9f0
59499 F20101206_AACQWA smith_j_Page_070.pro
a7f956af2ba0b3067bb777a65755b49e
4f7c4b3bb7a2091c219ff1e886cf9d752c521910
23160 F20101206_AACRAS smith_j_Page_098.QC.jpg
088db11fc6f543115ce58e7cd19b3167
871f443be8b21a45392591399c3d3f5d89d1ac42
F20101206_AACQVN smith_j_Page_125.tif
5a13c983fdd5cc5bdced8b23172f483f
ce5295d3b1c971a6c4fcb85ef67ba740b354fa93
1051927 F20101206_AACQUY smith_j_Page_083.jp2
ca774d9ca102977fae0dbcba070a588e
c84b76e22a715596f8ba2a2f0a0a09db21613b26
63035 F20101206_AACRBH smith_j_Page_008.pro
e72cdc27f2712c0fdf91092ebf5f264c
2786b830026b33526459e97adf4902119008ba02
54866 F20101206_AACQWB smith_j_Page_097.pro
d956d2f9fb18e9bdc4eeb691865b2ec6
4ec0a57ceec47ded30ba7f00e62b635caa6227ff
2395 F20101206_AACRAT smith_j_Page_105.txt
f451b6f97ce89343f8c47b6a65642139
976c6d1d9205342e8c78ecb8cb7dc4713691ae51
57887 F20101206_AACQVO smith_j_Page_060.pro
0162f8467d396b481dab4ec2ca491196
1d83d44a77e175a776724e9d54410fb0ddba2815
5812 F20101206_AACQUZ smith_j_Page_135thm.jpg
ea8d0937ee6b00935a1a16e10f5a77e7
ecfee96d5a9e331bbc3b4b98462234a71eb2a144
105802 F20101206_AACRBI smith_j_Page_136.jp2
d95bab796daa007a8cc7a1854e26df0b
558634938456f9c404a0a412d1a968e26c5fc662
12549 F20101206_AACQWC smith_j_Page_115.QC.jpg
12b8a38ef3747827f388996f448f0b37
0e6ffa9d1f86ba8a23c03d13209dea093526c5f5
2440 F20101206_AACRAU smith_j_Page_076.txt
81cfaf7db4a6cd098669ab98327ce029
803e4200acd9e6ec492f4619c66765a4483c49dc
19927 F20101206_AACQVP smith_j_Page_102.QC.jpg
edac7de66b44ef8a198ef53690928a24
c3e6e7aebf56e012042f97ef58e1f53d2413e988
2375 F20101206_AACRBJ smith_j_Page_094.txt
35d62fa220978b5d3ce6559f326eaf15
6216374712574789bc56371e9b6f01b8ecc60763
72079 F20101206_AACQWD smith_j_Page_011.jpg
d9480f26428338e69dd3d1cafc0c24cf
a2dded3c5afe274c0ae82cac9b9034ca6cfa08a3
25195 F20101206_AACRAV smith_j_Page_004.jpg
dac84902a7e1900e7fc816eafbe014a4
9c3a5fbb8dabefb13288eedb0cc280996e5185f9
7779 F20101206_AACQVQ smith_j_Page_138.pro
8405fda33e27e26cdb4f93e0b44dbfa4
b6a6bb8703b6599e772a249d8668efb759400ee5
1051940 F20101206_AACRBK smith_j_Page_045.jp2
3b4115aa5226b322745a6baac341fc9e
559589d2a1ae22b31b4737051272b2c35b10f58d
84097 F20101206_AACQWE smith_j_Page_122.jpg
d7cb486e21a6a4d5e4f89cf6a8c7d126
9f3807d1c2a25a00ef8325ebad4dc3b563c8f305
F20101206_AACRAW smith_j_Page_076.tif
5bd634473f97007350ac0a026dd56ed4
0e2a3d549ca5f8656bf29c94994316437c6b3716
1051973 F20101206_AACQVR smith_j_Page_098.jp2
725466f6158a9dbb460057f2c1d78a52
73b64448d1436220bd3e3a14f54904fb140afdcc
68247 F20101206_AACRBL smith_j_Page_006.jpg
33173cba84df5c7a326572002175fb9f
468300c20f35abed85f801a9730694a2aac2806f
79832 F20101206_AACQWF smith_j_Page_048.jpg
87bf03f6caf6970cb9631e9fad15588e
74173e3b41364fcb1705fa4ef81d2b3bc40562a6
2299 F20101206_AACRAX smith_j_Page_103.txt
1b9a98bfdfe6df24edc29d85b45ed63d
46ef7ccb6f455799e7155cab084b250cf9a33586
F20101206_AACRCA smith_j_Page_021.tif
5a480078a614a3c379f4db6f15bb380a
bd416412e41390d119fbbff1352d35f70dccc2a2
46228 F20101206_AACRBM smith_j_Page_092.pro
80d71cb1f04142417cd40d6da78dade2
4f12bbbfe1e67ba3c1507cc953adaf5a00560541
47269 F20101206_AACQWG smith_j_Page_134.pro
9b45f0f3cbf80ca633fa26ca67b60548
cc37d6a5c6290bc207df9f821b59249433750955
F20101206_AACRAY smith_j_Page_048.tif
7ae7c2127712a8fe308331604ba50c6e
6e58f63ef57270c83fe3c829560aada15fe13a52
F20101206_AACQVS smith_j_Page_047.tif
e4c8f4b0fb7210c4a150eb963fe6d317
6de4d5d74cea81509cb3a7d967d62ff5f99d4aaa
75505 F20101206_AACRCB smith_j_Page_041.jpg
b600cf1305e10f378bbdc1fa4465ca4f
1bdceec96bfa3b69b2edda60c2ef85e62b464f84
45898 F20101206_AACRBN smith_j_Page_052.pro
4351cc51c2d6a447a8e748157c83307d
9e2fd9ac880541d2ab75984c4fc43327a3ea9559
58216 F20101206_AACQWH smith_j_Page_105.pro
ca2eaa597a15fe557421f0ce040df2ab
327d370e2aae8e5d648ffad1f76e1905c3dcc2c6
F20101206_AACRAZ smith_j_Page_075.tif
a2ef55fccd215c38cfc549104d8e8530
b1152910b8367e02b170a85e5f1ad355bf1c2226
F20101206_AACQVT smith_j_Page_133.tif
00a993aa4207eab0ee0dc51a1fc0cdf3
5242d49b5af820dfa13615c52da1e9de976ff53b
51520 F20101206_AACRCC smith_j_Page_049.jpg
adb4339fca61eca6432b471a5f88d7ee
9f2c883e85da390e154c30ea2ad5c83608e55473
5620 F20101206_AACRBO smith_j_Page_126thm.jpg
f548114dded679a00ea68448605b586c
4d6f14c1a3174e9621a81c403c14ebcf486502ee
2043 F20101206_AACQWI smith_j_Page_091.txt
29168293760b93bcb1d4d778ad98bcc9
6c12812bfaf9652fddaa2fe31b09da2ff12422ee
26012 F20101206_AACQVU smith_j_Page_077.QC.jpg
a54da985e9428a7297e9c03c9f0f6943
1d645eda5f2e1ef8e15385f9f18c3d66280e421f
59836 F20101206_AACRCD smith_j_Page_069.jpg
96e71cc393e2fcda4cc9b7d68620e544
fb85c7ae781fc77833e7f2b1c7c8dcc5b297414e
67538 F20101206_AACRBP smith_j_Page_136.jpg
3cc226d42fd1efd757baba52e8fec3bc
eebfe37a06039707c5239ffe36b5d66c10581543
3623 F20101206_AACQWJ smith_j_Page_007thm.jpg
f0455dff2da105fcae1d6e3f55b5380d
8e67fed5873722ecc7cce4e249258dcb388137f9
52331 F20101206_AACQVV smith_j_Page_064.jpg
a3bfb53b34ba49b91c4ee7fb0405a534
2a8f9d825176a41fcc6da83200c1590fb4b6bfb9
F20101206_AACRCE smith_j_Page_054.tif
fc59fe7fd14e81466edf4365697432d2
abcf3bf42d5a1c61d04aa30d9f0f6acfd6ff55df
1051944 F20101206_AACRBQ smith_j_Page_021.jp2
b1a6f182199f6931a1807bbc674026c0
9c5720ee62a3be3d02c3d937894ca1114c983bf9
77992 F20101206_AACQWK smith_j_Page_062.jpg
4724d1638c36ed6197cd81f75eb6dab8
277fb098b46de4ddb474eaab5efaac6f78f2dc32
55013 F20101206_AACQVW smith_j_Page_122.pro
f7dd7c29b6720189195ca9638fef6f3e
c9fb92c7084eb2b8e53cc1c36b0479693a02999d
23627 F20101206_AACRCF smith_j_Page_030.QC.jpg
43cf3de2d3e3565f130e370a7b9aa929
7ee08d0ec8a3c0135747e95d8dec97bd3314501b
63734 F20101206_AACRBR smith_j_Page_033.jpg
ce167dffa60047cd0c0dad10f6922132
4e931053d09c6d84cad144b1928e94403d2b8c8e
F20101206_AACQWL smith_j_Page_007.jp2
4e59e0caa74feb65d4b922cd4942b131
0574f03831cd61d85d24c4775f7840f6e24f67dc
75346 F20101206_AACQVX smith_j_Page_065.jpg
3f3778c0224bbc42084e52db9d924bcf
bf64284fe811c76e7ccfc398ff5effaf86341159
50082 F20101206_AACRCG smith_j_Page_005.pro
682e38ac54c1c4efaa7711fe371942da
8fa31ccecc4b25310be869108f3bc39ad9e2f844
78857 F20101206_AACQXA smith_j_Page_123.jp2
1b299876675bca826e775d69bed4123c
25c4158edd69f23fda6df3db16767dba47c9d236
22952 F20101206_AACRBS smith_j_Page_065.QC.jpg
f79e5983f003160e9b747ecb26fd9e86
f97c587a4c87960bc504c48b4fb935dc4adfc6f9
806970 F20101206_AACQWM smith_j_Page_095.jp2
7cc4ac8c9857c51476636ae13a969fe9
a00ddabe98c15bd3a6b2363e27d3403277b0a653
14538 F20101206_AACQVY smith_j_Page_023.pro
973986fa6886531a53f91de1aa5209af
08a7674fb9e63abbfd387c3526ea4819fb77f483
89347 F20101206_AACRCH smith_j_Page_075.jpg
6defc3f5db15f3da1d1590b0a352725b
5d352d5e0e412fb63f54a59c1f4bbc0eeff1985c
16221 F20101206_AACQXB smith_j_Page_064.QC.jpg
e6155b3e95ab3b0cf3d20afde9e5e11b
14151087ad9b09e4ac2f5108bf7e371662188975
89501 F20101206_AACRBT smith_j_Page_024.jpg
054f5a0e8fc2902b9bb67d7133704b2a
db27987c4315f5c72ce1c0acb3b8d4d3438a7082
F20101206_AACQWN smith_j_Page_041.tif
fc60868de8e53ae8973b781b9a94a5af
5655be35e61f69df4604fb12fbca1b1d81dbd446
494 F20101206_AACQVZ smith_j_Page_089.txt
35256e8d5512df71275b7e26f0b70591
0b9077f6103ce89730eb16eba2f659fef2a48f6e
5828 F20101206_AACRCI smith_j_Page_052thm.jpg
59c86f4797fe40dc84e67313b7fb4563
cd99a904343f6e849c644328244507e8b6e71000
59393 F20101206_AACQXC smith_j_Page_124.pro
b08af489d7d09f2fbce3a9a4fa76811d
83bc1fb4d5eff064a4054a1845303867981c2239
852875 F20101206_AACRBU smith_j_Page_087.jp2
a18850b46d5cb05c99185101a98dff98
3504cce281c3b32aab0befbece0d001a2575c279
6301 F20101206_AACQWO smith_j_Page_041thm.jpg
984379b15653135321c8fbfcb6691291
b113cb016d9b7db407515e120f8b3640ab434e69
53994 F20101206_AACRCJ smith_j_Page_125.jp2
bd898d2a84d8010770a1923044f0159a
798b4b7327bd05240a8590c0148ba3980148b6b3
681 F20101206_AACQXD smith_j_Page_121.txt
e42d3095a9d01e6fb8fb64bd646735ba
705d875c4b95b1df4673cd23c0c1e5379c3d09a7
73455 F20101206_AACRBV smith_j_Page_098.jpg
63b294334e953290a31a56f7f016b586
d1f17cbfc4a198815f5f5be4db4f5c3e08156c59
6846 F20101206_AACQWP smith_j_Page_124thm.jpg
0811f0f91b3efb23b65ea91b4df37dc3
4dcc2ac00377ef762e21856d67ab9b18d69db0ca
22221 F20101206_AACRCK smith_j_Page_052.QC.jpg
353c6b200a27897e1a6bfc9649eaab8e
3cff63aa1816955b94846ea19bbd4727b8e0b9ef
57637 F20101206_AACQXE smith_j_Page_072.jp2
f65f32893d0c2cbf885ba96ec1bebbc7
e457846aa8de7e19896f466fe568327dbe7af430
17963 F20101206_AACRBW smith_j_Page_099.QC.jpg
9552ab6cb94c5f7f59d1a997f47d000c
3fe828f00d5464739c5cc1d8fc3b516e40a31927
28670 F20101206_AACQWQ smith_j_Page_022.QC.jpg
6119324c453799b92704a792da4e1acd
0759e8e43c444cc536cade2867433c17f100b0bc
5617 F20101206_AACRCL smith_j_Page_107thm.jpg
49af9549cb126a6073e33969a5753efe
170df421a62dfe2b54e80470ace15a2c25e2934f
814712 F20101206_AACQXF smith_j_Page_066.jp2
77540bd8a6687994b0d83f1f90cdd0e4
3fb2f28fd4b067533d350bdcf3b93d2d19247bcc
20723 F20101206_AACRBX smith_j_Page_126.QC.jpg
6f291f42ba2d6bcabb274dc73ae0e3b1
5475884d2a59944fd81de52e0000fd95a3278f60
F20101206_AACQWR smith_j_Page_043.tif
cfd19609bee494bcbffd44315b9fb82e
b8b09f507c81845d4c02b8c3307827c09ec2f942
1401 F20101206_AACRCM smith_j_Page_102.txt
3a4e49a55dfffea541b7408295063ac9
f5f4f385918c262db5ac6a3c9b73ef19a5b56f3e
17939 F20101206_AACQXG smith_j_Page_006.QC.jpg
8870eac97c332a6ee3689e7c8cbe1ab3
0540186bd70d541284adb90badd06ffc248a17d3
25937 F20101206_AACRBY smith_j_Page_097.QC.jpg
002d971c357f84e92624ca1ecbbf5937
84171bc29817809a5a3a56b3236ab67c3a95b00b
24090 F20101206_AACQWS smith_j_Page_041.QC.jpg
d2b4fdfdb5d9f192dc8a61cb3bcc46f8
f43d03a15d31ea1ad08f3711307b530586644042
28051 F20101206_AACRDA smith_j_Page_075.QC.jpg
5f6da77cd6a5f1ec17b8bcd531bb8133
26c7045b675395b31edb5bd22f675a729e3b5d9d
955363 F20101206_AACRCN smith_j_Page_129.jp2
ee16a4b3fa338059d653e226a6ca43ad
7dab6a7c297e768d026ad20f065e9281a2f09047
1952 F20101206_AACQXH smith_j_Page_037.txt
232e36e7ff944c72714c6b27cc98c8b2
6f231120941b9c6df1622ec929194705d32cc4b5
F20101206_AACRBZ smith_j_Page_005.tif
9b20ecc7ad0f8bff9c8a6f1f2866af9f
25eedace73551549c1dbe5d2874fd70eedcbbb7e
2067 F20101206_AACRDB smith_j_Page_054.txt
fb593b85006894d87d57f1b5e01015ce
bdfb1a83a029b317081af56c19e526e8ccd83756
6901 F20101206_AACRCO smith_j_Page_106thm.jpg
929523370daec65966a64a80e3e831e8
018d30dd611eef013d3d4903c8fa2d916968a05e
1051912 F20101206_AACQXI smith_j_Page_064.jp2
3b356f76cbb68d1792fa571b01055344
b660ef2999b7db67766391dda67bcb4c376ddd01
1359 F20101206_AACQWT smith_j_Page_051.txt
f30fbb1166547d145cb31adda7acab70
1a209260881d1ec06564417a6dd1ac75784863f7
2021 F20101206_AACRDC smith_j_Page_038.txt
82c7b2949b8fac5fe9ecaa183a0d2765
75d6dcc1a8f00de67a22fed956060cba57107316
77388 F20101206_AACRCP smith_j_Page_030.jpg
6046bff6a0dd459fad5dca86f2ff5546
7e6e3af9ed1c40947ce0b82b458ff435053df012
F20101206_AACQXJ smith_j_Page_096.jp2
c029a5e018edf52af2c0662034f0b14e
52e4f1a2582e506f399f64b50141a3e01de05160
57445 F20101206_AACQWU smith_j_Page_103.pro
0d19506ebb31d34e3d61ae6df2f8db54
0030f9ac839301bb1574b0ec39b5f82009b709a2
27710 F20101206_AACRDD smith_j_Page_024.QC.jpg
95bd93493fe0d22972c4b435014eaa3b
9e644aabccf810842e64ec05a13de2608392db7c
24506 F20101206_AACRCQ smith_j_Page_038.QC.jpg
0b5e072b0c4e3b51f8013c311ba65a39
7ef029fae0d88f499f5c0d6d5a51c0320f074213
2377 F20101206_AACQXK smith_j_Page_070.txt
5993449f8e0e2f1cdf17a1ae4cefda2f
5d9faae1f06777da0c6715e724ae3ef7337c6f6c
25021 F20101206_AACQWV smith_j_Page_130.QC.jpg
dfb3f7b445855eabef342af1cb25eb5a
4a723cda1c9a29a9156d306bf96fb2cefb1e92f6
1990 F20101206_AACRDE smith_j_Page_092.txt
fefb81458a6dc96d7587d5e5869cc6f0
239f3c79fd97e1ca67cd574ca5e5d1d0b7f9ba1b
74185 F20101206_AACRCR smith_j_Page_137.jpg
332a16ed1c58e67e2316c5f72ac30702
f6dad86ce0c7d9fe3a8dced7a4235522583fa241
F20101206_AACQXL smith_j_Page_081.jp2
bfd1748885afb3a583190d0979cd6b84
9df33c492cb8f2d88741d41de604362bda2bcbe1
5739 F20101206_AACQWW smith_j_Page_090thm.jpg
6dc5f837a370adfa140500f42c3997ba
a8c4a87bea3df96ba7b249a159bab85d9919ab35
91194 F20101206_AACRDF smith_j_Page_086.jpg
b621782c94f59ed953c704c8714fe4ae
6a39499278979dd0c90a0402cbb5c4830636859a
1035240 F20101206_AACQYA smith_j_Page_090.jp2
ba0813748f50a00a9077ffb3b0c2a594
1f553b999c2eb6a152f528f500e82949fd132449
54167 F20101206_AACRCS smith_j_Page_135.pro
a7827d51cca24211b9f6e97dbc41c323
bdcd87816b1e46f048b617ef29a592bfdfbf9989
68109 F20101206_AACQXM smith_j_Page_109.jpg
76747b9fc92dbcd64522be20ab927505
fda157583a93824b6f4d61495d4355ff96800fa7
28020 F20101206_AACQWX smith_j_Page_028.QC.jpg
61a7c3bcc452f5c8a54b9c2c0bd89b5f
8f83532c019fb01035fc334d2894b4ecd725daca
F20101206_AACRDG smith_j_Page_093.tif
4e845a42d785944ea875ef7bb36392a8
e3cf5042e0d29375030a8d5fd7d4eccd08e7fd01
61592 F20101206_AACQYB smith_j_Page_107.jpg
42964782cff65019df558ea2d531b7e5
f3f8cf2248977791a4c9e592edc57826efc361ba
2168 F20101206_AACRCT smith_j_Page_001thm.jpg
24426a46b3a20ca4b73636c1562004fa
258ad66227f2e331602bf90a3ab5f38c56e05208
6568 F20101206_AACQXN smith_j_Page_096thm.jpg
b0634164f72b78b0f77843ea66be402a
3828a3e6c172253d7ede00d16b68805e87055827
16160 F20101206_AACQWY smith_j_Page_049.QC.jpg
15c63baf950187f81389dba5da966a65
b2179a71a7ddefabd5f127a13db506e508c28400
24438 F20101206_AACRDH smith_j_Page_048.QC.jpg
2ddcaea875cb0a4dc4257adc4d500eea
230ff97ae33b2c588073cc46fffdeb3ebee77edb
6295 F20101206_AACQYC smith_j_Page_139thm.jpg
cb8d4ecb888f569ddeb3be0d2c8891d8
d30a5f078f6622d8b28ed86d76eeba3e3e0af311
F20101206_AACRCU smith_j_Page_019.tif
d448a08819f80dbf337f96e40275059e
c1fc03fd4c41a6b2908a6158a1d1f606afab6d65
822885 F20101206_AACQXO smith_j_Page_121.jp2
8db37cac200c372036b3cedbd1c5c1ac
195bbc65bede8dd9a5da1f7d3e8488ca2b8b5cdc
90863 F20101206_AACQWZ smith_j_Page_116.jpg
02606488fd50a1205b5d4b26535dc219
a5860fe5fb0fadf6ba5377f959ec04214b65df11
F20101206_AACRDI smith_j_Page_016.jp2
2e5e7409941057f554ad46eb3918f360
c45c28a9327a089644f9aa4041b6b6edcc42e7f1
881 F20101206_AACQYD smith_j_Page_002.pro
9ea24e5c100604d69014e7b407180407
6b3f6fff4d5aa1249fdf4f5397a8c51977df5dc5
939721 F20101206_AACRCV smith_j_Page_044.jp2
a5e79ec3dae45aae070c944410ec0561
0b3111f2972eb575b2a7f677359cee1744129161
2336 F20101206_AACQXP smith_j_Page_050.txt
fe64a3a39226fa9350c8f7f42beefaf3
50c3b467119f320f10fd86f2f09fa7732e95ae43
F20101206_AACRDJ smith_j_Page_023.tif
e2a6028a30e0b2e6820381eae866bb54
7d65e50789cf11c8f9cc050ba8ce89b3d026705c
1960 F20101206_AACQYE smith_j_Page_062.txt
c37be5b3c0c37213245db893c051b9e1
6a2c0a5eddfbad52761d3bf7d163125ec036dc21
15722 F20101206_AACRCW smith_j_Page_012.jp2
a02a5f310817d526a9d651f7a9b73abb
a6e268ee0f62c01d1a27cc9a710cc6b0de86bc83
367 F20101206_AACQXQ smith_j_Page_128.txt
47822f3470d5a80f368a7d9c4b8ae4ff
9253988106e988a89acfd5b33917ab54d254069b
26740 F20101206_AACRDK smith_j_Page_036.QC.jpg
678ef600c04e35e413abea52dbeba28e
252d9deac090f3de1d66733bd99fd8144c13aef5
F20101206_AACQYF smith_j_Page_078.tif
a0710be4842fd473d2725cc4e7ed8f61
776e755dc02c87358e911615acbfec48be14d4f3
53534 F20101206_AACRCX smith_j_Page_085.pro
adb5911102203e61d2c9fe2506f0e312
a256208a875d54a1419b80f1c562069aeb808af9
22278 F20101206_AACQXR smith_j_Page_083.QC.jpg
d46f808d1b3a39d4ccab9484d6fbb149
4785b36eef05ed2bd11d4566f12d00c119c51c57
1012307 F20101206_AACRDL smith_j_Page_093.jp2
f5a9142c9490eeafff9af68ed3f2715d
8075e784587f59d4d468dbd54a4077de5919ff2a
F20101206_AACQYG smith_j_Page_029.tif
28113254dd69da41ec333bd2d68e9b4f
39dd9410af18716b2db26ee0ca2d42f5ceb04a62
F20101206_AACRCY smith_j_Page_118.tif
ddbcf25c70f253b787223591aca8289e
042fc1f4a206e7278c07d04c9d37c0f93a915ff1
56713 F20101206_AACQXS smith_j_Page_096.pro
08eae9e627fa5f75a4c10bab27411560
880e0aae8d4715f82981d499b9373c87afa9ce2c
1431 F20101206_AACREA smith_j_Page_123.txt
507419641ba1bbe25e6de5fa96642642
6d700340dd6c9f1b1982b15f2bb714eea9f887de
3833 F20101206_AACRDM smith_j_Page_089thm.jpg
e1064036709c866dd5e03ad63218ea56
cb79e302dfa5bd6e71cbd8c54d73f5e127259319
101112 F20101206_AACQYH smith_j_Page_010.jpg
4d74e5777d6f5f13b0e82edca4bb674d
bb68d01d8ad44eeb221877926b78c42ea88cb3a5
4165 F20101206_AACRCZ smith_j_Page_003.QC.jpg
bbff878aba7ef6c31de9b74404e1aa39
b2490c9f6756f781767fd02256ea294ca9272475
77385 F20101206_AACQXT smith_j_Page_139.jpg
5ef3706314c03f6f60b7882b949d6bc3
2b9fbe071e75d1928cfe81176a6aeb500c950377
25007 F20101206_AACREB smith_j_Page_042.QC.jpg
0d831d0f62d66a3ff9257a4c6e7efb7c
d3f226b1a1a5cb92e2d16e629f19fce4c3f05270
F20101206_AACRDN smith_j_Page_039.tif
a23777a726f4ecb5599aed96626a2256
b28463053500b3bce6c76584720803ef1e1eb93a
27700 F20101206_AACQYI smith_j_Page_088.QC.jpg
e70f3eeee06db429d40195adee189be3
e3d7978eb463f6234ad527a5e8bfe55bcb0aab98
99381 F20101206_AACREC smith_j_Page_134.jp2
92fc88ffff10167b970444e683233736
ad301b1d5f4b94942fd66dc5eff775bceea03ad0
5571 F20101206_AACRDO smith_j_Page_109thm.jpg
f3d560b97e14ca84f34e1be2e3b8467b
86e1c120cddf8d0ef64e4599943a330df8a89111
87923 F20101206_AACQYJ smith_j_Page_074.jpg
40683e06d0042f4e55541d0838200937
ed0236fac0dca40d11b995c1458b7f1c1e1b311f
F20101206_AACQXU smith_j_Page_053.jp2
fa2d3631e5a6218dd5141818c68a2240
ab8a8e0d032883f1df72ea4551e6e306615f8885
F20101206_AACRED smith_j_Page_052.tif
0ab17bec00002ba46045d412e65e7ff4
7604ab30fd1938392913f66e772c54d24b08146b
27897 F20101206_AACRDP smith_j_Page_007.pro
ed1eca4c3e1e3df52905644c81f2f42c
d13586659bbb16931c48f9c4b8d6bd0c4a7f98b0
1051937 F20101206_AACQYK smith_j_Page_122.jp2
805485e4a346220e6c228035335aed9d
2e050b59fd0e425db561de408e29d4486c3936f5
1051969 F20101206_AACQXV smith_j_Page_015.jp2
8e6335ef9d221f121e09052943959b8d
564dd974204e395b836fd80c4eab7daded91f685
58118 F20101206_AACREE smith_j_Page_079.pro
387edd91981ffa44609bda4629805e2a
c85269fda5721d1e83beb6d4c33b0274407e4a57
7287 F20101206_AACRDQ smith_j_Page_046.pro
16fc93af25260bbc4108afd2fbe45513
83fa4184f17a395eca7d2defe01e404263d6eb9e
42220 F20101206_AACQXW smith_j_Page_108.pro
3fe3949175626f4815b92128939ce4df
f1d1700e11c27067b3c7287e7b30a4eb52788fc5
1051976 F20101206_AACREF smith_j_Page_031.jp2
1371c33c41aa4c12074a02f9b422a882
9dd1460937929cb1db000ffb6004b19612941134
F20101206_AACRDR smith_j_Page_135.tif
888f639c7607e9fb15806c19c1c5766b
f76335cda16746e3241df755e2b8202a7197974c
F20101206_AACQYL smith_j_Page_075.jp2
598ea688429e1339f99a1030d9f87c39
33b48e792fde89c3272404ccac90b93c1b375c94
52800 F20101206_AACQXX smith_j_Page_133.pro
c813164c45fb4f0f32aaf65222671c64
93193acfbc09c21093307e68e09c524f17134555
F20101206_AACREG smith_j_Page_103.jp2
12e13651e0520d4729f3fe7981b7b179
7583134324f200e74167e7a38e8e8fc9a3facc7e
6265 F20101206_AACQZA smith_j_Page_081thm.jpg
6915006cb70326eadaa7d73dce5f9fa2
513de7b1f2fd0a89b516de9e41a99f24c47b1dd8
3236 F20101206_AACRDS smith_j_Page_002.QC.jpg
1e55830f5fbaf3c2d009761a944bc621
cca1d918b8017c750fe4bc863f086d9cf91c452c
1051984 F20101206_AACQYM smith_j_Page_130.jp2
af92a804e8eac04629afcdaf243bd29a
44eedb1c88cfdd878072b65f6eeadde0fe18b775
77260 F20101206_AACQXY smith_j_Page_068.jpg
ec701e71525a242650ee8deb980ca441
4dabf32e0ed543d0cf6006a9a959839edc0e1f99
1051956 F20101206_AACREH smith_j_Page_030.jp2
0b3e43ffcbd1eb6243b731b0953fd164
4dea94093143fcb4a5e865897c101af17f2e0482
14493 F20101206_AACQZB smith_j_Page_071.QC.jpg
09eff51e04660613404b551eb0bbd291
bdc9774c5cd5a644278591064844f1a9305c3c61
25395 F20101206_AACRDT smith_j_Page_101.QC.jpg
8f72130d40c1eae2ce5d0462ecdac42e
f7d66c1786c3e9d9debc7f41e7fcad1004211310
2241 F20101206_AACQYN smith_j_Page_097.txt
582259ab642fc13323ea185b98802bef
038b89052f277935084adda9afb5e51c6f2bb364
6345 F20101206_AACQXZ smith_j_Page_038thm.jpg
9c6f109bea46ff3f5bdc3e2fb0daa60e
7e52c17db83450aa11de049bea1878a404412ee0
1511 F20101206_AACREI smith_j_Page_033.txt
f69137f85c64c6541b19e409798f8717
f2dedacf64494a0497ade78d75f91ef559801709
1875 F20101206_AACQZC smith_j_Page_027.txt
3ad9219360fcbb813cb61f917c573ab2
878e2e042c6c0895600d5d312b05e8f1e75f4bae
85296 F20101206_AACRDU smith_j_Page_073.jpg
46835a90ca97b19c649e40793d3217eb
b56842e4325df72b2f2bf3286f7b7163bd965a5d
84192 F20101206_AACQYO smith_j_Page_084.jpg
402f44a5b830f22bc3f50109c5c5bba0
8f0ff9849c3a1081354168eda3b53542e0b74ee6
F20101206_AACREJ smith_j_Page_085.tif
c8d6e1265b8e01403ba645fc2ad13735
e577e86295ccd79cd321dd714ad2dee0155223a3
1337 F20101206_AACQZD smith_j_Page_095.txt
f6e9f330456119d8f50b186a1c4ec248
f1cc651868ecbb159e014683deb45e24bd95edaf
F20101206_AACRDV smith_j_Page_035.jp2
177ef16dcdeecb64e86d80d0fb0e109e
00bce6e7ad2b2d1ac4970f13b110ed03e3a2d543
22099 F20101206_AACQYP smith_j_Page_092.QC.jpg
ec73a5fff1c75e19a4356151521a61be
726f9c46499067e181b3d1f02ca760ce2f57191d
89853 F20101206_AACREK smith_j_Page_027.jp2
3ba591cf6a0a5c6da5b9cb7090297485
916c60c771902b302d3c401f684f340753cd2875
26724 F20101206_AACQZE smith_j_Page_095.pro
3ab3c2bb6701bc425ab5a351eb33db4e
0f215ab8654a328ee33847657c6c8030523b3526
F20101206_AACRDW smith_j_Page_004.tif
9b8e110a324b8bdcc856dc85b227869c
2c8127c698f0127a0dbc88f0c3cd7c4ae0ab12a2
F20101206_AACQYQ smith_j_Page_082.tif
c27ce09e7c6267f4f53bd23655ada550
6dcde4a76a9ab917e70984db340f8562b7a81c54
F20101206_AACRFA smith_j_Page_071.jp2
2433e3516edc3e1a1ec3ed555f42bdf2
4bbfab8890022cc73dbc1349249298d8eb4ada7b
5591 F20101206_AACREL smith_j_Page_011thm.jpg
cc5c89d5271fdc0323fb4fbaf2575991
51f5c663fb82efd065cb776c42170f986c80cd98
2118 F20101206_AACQZF smith_j_Page_085.txt
6e46f6fce5e481e06deb7da2d4f171d6
821cad24d65fca89ec25f47e0367e9d09dd284e7
2153 F20101206_AACRDX smith_j_Page_077.txt
9a5348d76be6de65c6c17086dd7541fd
2c2c39effce4f336db1ba5cb94ed1ba0e6df9520
F20101206_AACQYR smith_j_Page_057.jp2
73b0b65268bd77204d25d8f1a84ded01
bae6cff29b03c11f037867e995cce7a74f095d7b
57779 F20101206_AACREM smith_j_Page_139.pro
875820089d20a7df2adc0ceacd5ed02b
31bae177a73aee815d279454ce90fa01719abb44
121815 F20101206_AACQZG smith_j_Page_139.jp2
1193bd1f0f0f44543323d21b6ecd66f6
165b48251e8270d815ff4162244db6615e2fe73c
2238 F20101206_AACRDY smith_j_Page_055.txt
67f67a18b1fcfecf41c4620bc894c46b
d273e892d021575cb589cac679607e566560811a
6880 F20101206_AACQYS smith_j_Page_103thm.jpg
ae410b55582d70a37cae63ee1ad0c86c
26387b4b6b4d8943f9951c0912f73b57c7211ee0
6447 F20101206_AACRFB smith_j_Page_114thm.jpg
ee7d02070b49b5be955ce9bb627744a6
c5081dc2cd71a16a4e53b84993ecd5f753027ad5
F20101206_AACREN smith_j_Page_082.pro
c3b9588b7d8670390551091bfbd03cb5
b48b9e6f0861261f3383b92289d5c11cfcce4836
68773 F20101206_AACQZH smith_j_Page_043.jpg
9f2f870c495c43d8180aa9675a281561
2ab0c1e6a034b1624dd8c262d39ee6ab3055e2bf
17340 F20101206_AACRDZ smith_j_Page_095.QC.jpg
510ff7e33e7fec99e3c76d28fbca04ef
93f8f49339268b70dc56c1c0997de5a89cb79a64
5619 F20101206_AACQYT smith_j_Page_078thm.jpg
b26dfd7c85e96fc3270d3239b65dd353
5a9321a40ace71b8ddbf10e4e5c12abce413b45c
41663 F20101206_AACRFC smith_j_Page_083.pro
7b300485253ebdbade99a5bc50c324a8
0c3756f19b9e52f613bbbae3ffdaeffb5c346ab7
1818 F20101206_AACREO smith_j_Page_012thm.jpg
08f0a61727e68ce695507d11d7496f96
689c398c406ab7a558906b145026024892eec805
25157 F20101206_AACQZI smith_j_Page_139.QC.jpg
20f35ab962261349b08fd64cb66375fd
a61590f0a64289edc1ec27653204f59f73b83b5a
60803 F20101206_AACQYU smith_j_Page_022.pro
48f1e340b6d4b884722bdedb03ca8172
d42bca76db46edf50dd4b4e2984739c175b27240
F20101206_AACQCA smith_j_Page_132.tif
b0a6ba50d616d1aa35fb9a20ce4d6d22
2619942e16405df28404dca19cd2b9d316b85e88
F20101206_AACRFD smith_j_Page_052.jp2
9c24eb521b529350a229a850b414185a
a9ca3fe8bed98cf9a7eec1a3831e906936aaf0ed
57349 F20101206_AACREP smith_j_Page_027.jpg
2fc628f57e2b1519a1b5a729a4e1e5d6
96ba92d3042ba3b6964d0ff3f988af34350e6ede
28399 F20101206_AACQZJ smith_j_Page_086.QC.jpg
1bdce77c5fa94822b715c37e960c8e6c
8cd3baf7c33ee85edd25788a5aa653be1a18a666
F20101206_AACQCB smith_j_Page_137.tif
bc7b233c4429ed15ef843e3e3cc745e8
755fa227ea6f2041dca61f8f756fbc77cedede1d
F20101206_AACRFE smith_j_Page_010.tif
274f91de31e68780cef6714cd69ed364
fe37e443f10c18aaefd5d0dd5817bb97703aa2ed
47819 F20101206_AACREQ smith_j_Page_071.jpg
f394ab8917cd9f91b2af0bae86ec6a4b
5a51bd6d8f4ea356ee919872b7b2f4248d4ea886
1216 F20101206_AACQZK smith_j_Page_119.txt
80a650218a996b74a803762171d71f0d
a7ae4b405e02a4ed9eef654ebe48aefb7a04ce40
23961 F20101206_AACQYV smith_j_Page_001.jpg
17e00c5d38c9f39af97a2425b42a0017
4d2d7ff14462263bca10085800c41959887b7449
19678 F20101206_AACQCC smith_j_Page_078.QC.jpg
32df3f08efaf149e2c597a69371108a0
7d7ee3da683f43f661b3712478629670c707c408
2154 F20101206_AACRFF smith_j_Page_135.txt
218c098999527f196089bbe670ddc23e
08a83d6cca271485fa6343b0e97c9038c4f5da76
60284 F20101206_AACRER smith_j_Page_028.pro
eb44964d7e9d6ec98c3e490d4dd1945d
ff67f2f294f195350c5b3a408966844a8e5a7def
F20101206_AACQZL smith_j_Page_051.tif
0db02be9be84ffca496928518761d92c
7efa063a0badd7fdd6fd5927cbeefa2f82027c3a
74846 F20101206_AACQYW smith_j_Page_037.jpg
7810112eef2ace49736833488a1d0990
854ecfc637867dc5b6ae542c3a42d536085f1550
1699 F20101206_AACQCD smith_j_Page_104.txt
6a68db491e0d053ce6dfd8bec862e23b
ef94de306d5a9755c2be91f3cb7a119284f91664
2160 F20101206_AACRFG smith_j_Page_098.txt
f7b141bd04df380ee1df4d4350e198f5
270f3566fd8ebce0e70230ad11a7f1f564bf22c7
1812 F20101206_AACQBQ smith_j_Page_109.txt
12a31c647cc05fa3ace544abc66b5478
5bde4832066dd83885bde9f91cdd797c73608997
85448 F20101206_AACRES smith_j_Page_050.jpg
59538204241183c28d624a4c4edb5b3a
7a97c012134b38e71813bbb85939a46a5cd0add4
46768 F20101206_AACQZM smith_j_Page_037.pro
1ca34c5d1898c2e325d63377f45004db
595647ac58e26da0086752f15467724ea8dcbeb3
1051885 F20101206_AACQYX smith_j_Page_038.jp2
b95799c3ce62cd6e64d0d5156b3635b1
ce0500b16ad45f944e04419495916d516b76a27a
27570 F20101206_AACQCE smith_j_Page_059.QC.jpg
f8499817fcc45b726c910f0dbfb8bedf
af0c5109367055becd5130c788a7f73d9cfaeca6
34672 F20101206_AACRFH smith_j_Page_014.pro
b3eaba109e6db637764a74d1c33c8e29
971375f0fcb525fb7ccf04e4d2dcb06e4fb4ac37
F20101206_AACQBR smith_j_Page_014.tif
b2c8c5876b27e5a85b8df99771577937
d4898df6498b7a09da657181e664639154384bab
6115 F20101206_AACRET smith_j_Page_091thm.jpg
a67824ce19b0801234023c727f02f09c
79deb069cb5d4a9fde96103074db799481a98b3c
6586 F20101206_AACQZN smith_j_Page_122thm.jpg
dc19b184b908c1715bffef8806ae96fe
c422b6f80422a6aefb8d3dae211c478c85906124
43609 F20101206_AACQYY smith_j_Page_087.jpg
a619bd887cebe0f229e7c09bf75a830c
12e7d8046afd46abdcde4ab6468710a2266dffd4
20588 F20101206_AACQCF smith_j_Page_129.QC.jpg
f5b5d7a133be5bac30b50f1bc80feaed
8b9af2e855471d0c6777846f28f2ad9075b170e3
8291 F20101206_AACQBS smith_j_Page_128.pro
2553b958e7f7cb3bf9a14aac0604ff1f
a2c489a1ed3d43725fe72bcf4bd465c17321d13d
1051974 F20101206_AACREU smith_j_Page_080.jp2
26601c1a5c47dee8edda05349a9caf33
f0dbeeebf84eb9bec5421ca4cd2d8025c49ac789
4670 F20101206_AACQZO smith_j_Page_111thm.jpg
f2bf12c411b05eda894e720a954ec00c
38da1533e1c18e07cdd8dc9eb6a97ef6c9862a09
F20101206_AACQYZ smith_j_Page_113.tif
62559bc9b7226656b4f47aad9df86368
fb4fa9b7b25a5584a01d312c954a20944d5b0169
F20101206_AACQCG smith_j_Page_032.tif
9234b279f890d6c63ed0f3085777d6fb
59b3d5c0130b844f6e26542989120d03bd437fec
88084 F20101206_AACRFI smith_j_Page_021.jpg
751fabd7ee39f9cf9941f538c8054802
a98c3d19fac238d946b0ca8a1fd87821e5c1bce8
5618 F20101206_AACQBT smith_j_Page_134thm.jpg
53b3d74858aeaec550a62ac7b9b5387f
01f7ebd08f2da43966915bbd8510fcf7f8442955
F20101206_AACREV smith_j_Page_028.tif
6541d656520886a0b78a1b38a9e124c9
66280a7020dc09a328aec1b15e538a646421b23e
3306 F20101206_AACQZP smith_j_Page_125thm.jpg
996250f0a90f55994bd2902c76cac5c5
b1f55e2ae23b343b31349ea446b9506f741f4701
6845 F20101206_AACQCH smith_j_Page_019thm.jpg
b51bed793509c4be01c348ccffc47731
00b4293af7ef023d837f110c7c30049f941cd9ae
F20101206_AACRFJ smith_j_Page_062.tif
dff516ff5c4e9567ba10e86832a9afa1
b6117af207c491b9f5e0b2d430a2d136c4979742
F20101206_AACQBU smith_j_Page_013.tif
974b402bdcf3b86d4072cf4945188797
01f8496b5cff78e514e3436eafeee87eb3c35ec3
58744 F20101206_AACREW smith_j_Page_026.pro
e5abca2abd8c7822c9e5cf4bd1f7a864
a2e4c2c45ebb2e9e90adca57b9a7fa093ec7961e
2308 F20101206_AACQZQ smith_j_Page_052.txt
c5aa79e42cb9f9688dc63b0295bc57fc
77adb2ec9cb5345be375aaee0d8cd2b1b92a9c2a
2311 F20101206_AACQCI smith_j_Page_096.txt
69f9b132cbe706642a5c8aef2e6443b3
586177477d0e30b9eb17662cadedf0acf2f5103a
F20101206_AACRFK smith_j_Page_094.pro
9226768d236979e3622f14a6c669fc6b
fa9a53321bae6c5f73eb47c1b2ef78b5f08cec1f
F20101206_AACRGA smith_j_Page_031.tif
14d929e8df2381312f9615f2c7b65bfc
67ddd5525cd439766a3d22ef4c69be5a95beb00c
6517 F20101206_AACQBV smith_j_Page_032thm.jpg
b8696277d26d5a929cc4cb1fc44a7413
ff81be033dfd613834d502b600496878d1a4370f
66858 F20101206_AACREX smith_j_Page_010.pro
9f653fd219f70b32a7faf38fa15084a9
ca1121e0f2d3396399b534348d74035605029bb8
28547 F20101206_AACQZR smith_j_Page_120.QC.jpg
f3601dd75ea1520d3f14edd686eea3d2
26391e2ba8fbfc77e8ddf440003a96cd3f62409a
F20101206_AACQCJ smith_j_Page_068.jp2
e1bb7e532f0f8ee6f06fdd3c49c72b17
b1ddb38f5e7828a341dd96d81c48100ddb4fd98c
913403 F20101206_AACRFL smith_j_Page_102.jp2
76b4e9dc35078cac00cada25ec2fea88
cb11b0186fb56c1b5a695e3f9943b9099a5f9575
91916 F20101206_AACRGB smith_j_Page_029.jpg
8832ae56ade22d2e11db5d112cf576e8
de9ffbbe382e01b79940012ab8e73c25e9fec1d9
27919 F20101206_AACQBW smith_j_Page_099.pro
a955c25e1d0ec7180b1a10dfab8a4a13
dd8504f23dbf895b57297ed29c9583a827bd4b89
4214 F20101206_AACREY smith_j_Page_014thm.jpg
578367da7935334141b9f2568706e479
d712f0c4e1d6080062aef163696ad4eafe78ba14
F20101206_AACQZS smith_j_Page_066.tif
c4921f8a1b32e7913d336c006a64cdee
022c955a2292b900be66fc6e273e35ba934da804
F20101206_AACQCK smith_j_Page_108.tif
6dac6827574636743825bc8d643a1ce6
5fd8d50cc2be8c401c3bbef5ffd71df07c7b5f88
110026 F20101206_AACRFM smith_j_Page_135.jp2
48be34d1f344e7b3ffd7ee712ce95e29
8d87f592a6871d25d46ef01a03c9030c00f2e09f
52091 F20101206_AACQBX smith_j_Page_042.pro
1b42980800a9cbec999272b35d5a95be
20cb3ef7dcfc83c28c77bf64b62a43280a082828
87808 F20101206_AACREZ smith_j_Page_088.jpg
f78fe5b1c4ad4381c1d2b6167a6ea8fa
e526fbdba64cfdbb38eaa453bc3fbb76f5cf0553
82486 F20101206_AACQZT smith_j_Page_077.jpg
df4355c616e15e0a77db1673033aefe1
850e621a472050439da5538ae2a2538bf1813592
60646 F20101206_AACQCL smith_j_Page_116.pro
4f20374c8f6940fdad092cc802686112
54d0f69aa67c601831695364da9235e537df14a4
F20101206_AACRFN smith_j_Page_007.tif
7dc1e2dd901b804037d651fd8685160d
75382a507a17adea5fd6341a47ef460d1035f26a
13665 F20101206_AACRGC smith_j_Page_113.QC.jpg
78a03c94fdd8849e82dc4aa16385d6df
7db0d5ca5b888cb4ffa8d71bd1ba08ec4be132b5
52857 F20101206_AACQBY smith_j_Page_053.pro
f90152f3a4ef6022692c8d12941750b2
511b45d60c83a5201d67010766e9ccdfeb076474
F20101206_AACQZU smith_j_Page_024.tif
00c3cd3064a8d453bf71f6d87dbe377e
e3d5cc27ddbfba191207d286ca2069a5869c6523
1043967 F20101206_AACQDA smith_j_Page_063.jp2
40a559cfc0b71fa509e07407213c2856
db55e00f8ff56dff41f52452ded4467a5c46cf50
7698 F20101206_AACQCM smith_j_Page_058.QC.jpg
05a38edc98db832bb91ea6f8e38b2258
a86ed20704d80f7342959a3d15bed51ffad4d859
2672 F20101206_AACRFO smith_j_Page_008.txt
d9d7e1fa1e7f5aa54db204230f344cf8
69fee2ad11e399909ce6927e2aa4312aa7682135
48728 F20101206_AACRGD smith_j_Page_020.pro
d7cf8b26f9bdb416150230194b174128
8015dd632013cd7dbd0301a4336890073cd2627b
1051965 F20101206_AACQBZ smith_j_Page_022.jp2
40a78a7949f7fdd7f40de4e705c5a75b
a792cd4d1476ab0882044c9f6e7d18a8398521c4
2268 F20101206_AACQZV smith_j_Page_073.txt
f00ee0b51958cbbc1290d3630c50685e
0c8d2951bee7b4ffc122a0436455fef068f2d22c
6965 F20101206_AACQDB smith_j_Page_116thm.jpg
0144208919a8cf3ba10b6d1fdb481aa4
1f2f8ed5648063bdf9d782a4aa2eb8365d626316
32185 F20101206_AACQCN smith_j_Page_067.pro
3740b910c6090674be38b817a204a68b
71e81f53dc7e2956c99c7946939d13561cc021f5
F20101206_AACRFP smith_j_Page_017.tif
209ced99d7810b1e528f32bab2dcd85f
60df3e710052527f614a538cf26e38d51718b1cf
84670 F20101206_AACRGE smith_j_Page_032.jpg
7ecf493a6a6d1b8d739c0beb560d323c
05d3e5f0420adc44af3362cea620c6e9a352bc18
983 F20101206_AACQDC smith_j_Page_063.txt
663124858915452f0a27bbe20d13efa7
0e5f15c5200aaacb52218fd51b9effa4d4e9c9c9
88835 F20101206_AACQCO smith_j_Page_017.jpg
316df973506ca46301cb2a6c4cde825d
3424ff183bb7d85c2892058a53b4f27ecc52d267
51608 F20101206_AACRFQ smith_j_Page_063.jpg
024731adef8f56f182be7eec1f9c225f
8afb10b69495797ec00956f0bc145ad3b9e0bf86
F20101206_AACRGF smith_j_Page_022.tif
787534388495617c48ef09c33224b91d
53544182086d067d75ffc9e361186c34cb81e5a5
2098 F20101206_AACQZW smith_j_Page_009.txt
0178a481f3512881ea813bb68b832638
1edc64f5e1ba9e8a384ebd3ed41dd6500e92fd4c
7133 F20101206_AACQDD smith_j_Page_001.QC.jpg
6e3f878e234833943dfcef9355ede3cd
bdd33a62d2b6ba4dfc181ce0a66bdccf1bf5bd8e
F20101206_AACQCP smith_j_Page_101.tif
9c30ef410a5bb4368bf3c73c200cc00f
340608d22b97c37a569910a45d46521557603c2b
1158 F20101206_AACRFR smith_j_Page_064.txt
1ea4425b3aad496ff9fd29d5b588c99b
858e556e5d3202a57bad676a4003b943e4c48e08
57654 F20101206_AACRGG smith_j_Page_140.pro
52c30a029499465d72652bedfe5457f5
2ab84c0e758a4bd844b56ec2729d37116a79708e
4636 F20101206_AACQZX smith_j_Page_119thm.jpg
9f2e96cb630ac83ce4541eeb3e2d5199
6cfb8471d1bd4aba695487e1c351bd28081b2867
773 F20101206_AACQDE smith_j_Page_023.txt
d012bd149d2b42cc70f7ec43e06be85c
5eacd7d66791fb767d551dcce475d492559428a8
5651 F20101206_AACQCQ smith_j_Page_044thm.jpg
f69fc8be0e7acf07a0aa405f34298aaa
194961c9a6af5a3111c8da8fed51a565ecafc929
27151 F20101206_AACRFS smith_j_Page_019.QC.jpg
0a66825a6b8da29b7e873568a2c01737
494757a75b1b939f31e4da9e26b06946d302b95c
28803 F20101206_AACRGH smith_j_Page_076.QC.jpg
95c4d0ed537a8bb6392ba1e2a1e48a3a
29f6109419c555516da545f566fa0e0833793e80
26486 F20101206_AACQZY smith_j_Page_119.pro
491dce29b0a1c35b277709dd827f4ece
98eb52990074f9df4b09d019fbb7672046d88517
6945 F20101206_AACQDF smith_j_Page_024thm.jpg
e9798b57f65e6e1ee535ddff4f4a8607
fceda7817608a5eb4d15989e08153b3ae923fe3d
F20101206_AACQCR smith_j_Page_089.tif
2a3517119be6870a1420610a6d7d4786
7dbe3a81013b495a9b803df3b671ba9cfa68c10a
1912 F20101206_AACRFT smith_j_Page_034.txt
852cf50e22547b9140827dc82c654654
08ac336269c3263053134ad52572283aa2f7b432
61990 F20101206_AACRGI smith_j_Page_127.jpg
71a97f4c66fa1a6aa0b63d30484872fd
512006f1d6cffa234d7796bd09e4b10ee59d2d44
F20101206_AACQZZ smith_j_Page_064.tif
b85c594b5c290ec60d492e40059dabfa
963eb1838945bc24afccdc86824e68b93a2131e9
42742 F20101206_AACQDG smith_j_Page_121.jpg
709b19460dca24db75b7688e414dd2ea
662cd0a937b1bae275fac53d54a6a60b45bd24e7
32564 F20101206_AACQCS smith_j_Page_128.jpg
84af7b40660ababfc02cbed5a44dde04
c4ba87f96ec14712d3192217c4be732852e18441
2222 F20101206_AACRFU smith_j_Page_035.txt
a463cd3444b7135a430a8decfecc309b
e4776ca47bd496165f2e34d3dccac67e6362b935
80121 F20101206_AACRGJ smith_j_Page_054.jpg
a6c694060f02773ee553530ca38cef88
5b3dab00efe7f6709c861d49b094dc5e71f5e749
27777 F20101206_AACQDH smith_j_Page_013.QC.jpg
b68b17c5e4a84207b1a0bf37c5ef85ae
bb4f8f4f89b9d2f47fb860cc4f20991499681a9e
42105 F20101206_AACQCT smith_j_Page_027.pro
9fdf59b7ef33918f934838e0f1a4d60f
5dbdbb5a08448286aef090b96e07a710196f18a1
7028 F20101206_AACRFV smith_j_Page_022thm.jpg
b1f02cbb19753374bf1cc93aa95a77aa
a5a4e4607b40294a53f65616597a8869e9691709
6351 F20101206_AACRGK smith_j_Page_065thm.jpg
5a32456397194e072ee40f295238eea2
5d73c363ea5a6c966ec344833a7eac45cdd04002
7094 F20101206_AACRFW smith_j_Page_013thm.jpg
7bfbc4c311c8a8261a236ca2312d5632
b2b888f0d038b6e28becd1984c49ba617f9c7562
6914 F20101206_AACQDI smith_j_Page_105thm.jpg
9f14cef6b334193f3f1868950704e74a
1f0fa0ed234138e2708da6068b1905672a74dc51
3610 F20101206_AACQCU smith_j_Page_046thm.jpg
ead11290add5016d922161067c38d71d
914ed330f8e92c97eb27d2fd48b215b8efedb966
574 F20101206_AACRHA smith_j_Page_046.txt
e78058095b0ba8fff4bdd07c06902f58
78d422b4606237230fef07af50e4428c16e98f53
6710 F20101206_AACRGL smith_j_Page_050thm.jpg
24ea0241e86e016880faf555c9ebed00
b5ae4714a25189a457155ba43cdd8bc85a9e0d96
45530 F20101206_AACRFX smith_j_Page_100.pro
b9fa56cefd60dc008ac047187bfc9c5e
6401fcc27a690a2c94c721d28b6408594aca50e5
561 F20101206_AACQDJ smith_j_Page_087.txt
89790905f284f0d2bde0bb8a1ee14add
6d2b5aef5c16093560fe038cd73f0622b883b896
232 F20101206_AACQCV smith_j_Page_012.txt
87cc7c8ab4d681f91b7bafc99c8841d0
b0c6dd907a8b3e0b38b347763e3389f2d27d9797
729 F20101206_AACRHB smith_j_Page_117.txt
b8ee7d4c3bc0368608749a0836ba8e76
79015425d51bffea9a6857efbcfff1f3c642bd99
2158 F20101206_AACRGM smith_j_Page_084.txt
deb89974d1ef05e601af48a175605181
f89dd9476f7a1af43041bc51b5a4332af426ed11
88801 F20101206_AACRFY smith_j_Page_105.jpg
b6aa86e7cc670ebb1211624971327d6a
364c9802b5004fbfdb9c3007dd5766af4197b8f7
6962 F20101206_AACQDK smith_j_Page_120thm.jpg
d1d16add02c77cd6375ecc3da6b01822
f5d9c1790fe70bf6c80d9a17bfc7cb31dc4f9d3a
6830 F20101206_AACQCW smith_j_Page_045thm.jpg
3f0567821f73280fe98cd4e17ace4634
6ae6b73324daf43ebb20d8d6990ea52fa88447aa
12410 F20101206_AACRHC smith_j_Page_023.QC.jpg
47c1931514cea4fe45edf9d10fd1237f
e6a5b673db5565a7ad7b4ba001e439261cbea324
962699 F20101206_AACRGN smith_j_Page_104.jp2
76995aa374f8d4fd2040d939c7b04b90
99bf37179b0b9ccc9168b6e2515d9698693b781d
1051975 F20101206_AACRFZ smith_j_Page_077.jp2
8f961f9b2f8a30129387e71a6e51e618
36e77e9f097db0065d0bb06b7d05afbf945f5b94
20438 F20101206_AACQEA smith_j_Page_134.QC.jpg
3a32457875f0482a2f23f68ce37ce403
8d3178f7bc70f92cce8e6bceef307f4fbeb30f28
26694 F20101206_AACQDL smith_j_Page_045.QC.jpg
ca47e437db247f34ddd097d2f1d60f77
1e7bc90d6efbb1fd3f5aa6044fb6ba2a0339c200
2236 F20101206_AACQCX smith_j_Page_015.txt
0395f9f7c0eee749e0db0a6c2b1bcc68
5981d3b96a19e740aea70cac5a20dad964bfc8cb
60932 F20101206_AACRGO smith_j_Page_044.jpg
e7d5b627da4d7e6e184eecc6f73191ea
5512b1174d4809170d5d33500552bdaca38f1dfc
F20101206_AACQDM smith_j_Page_140.tif
baae79dac9cca49c5d345566ee832b75
7de155dc2d5febcf0838d2fac284d20ecc3a2ed3
F20101206_AACQCY smith_j_Page_077.tif
2e710a89159740f0e77ab0164ebe3b46
13e76531297282ad408a7340f22f3d9204ae711a
92261 F20101206_AACRHD smith_j_Page_076.jpg
ffdd1508fd25a7969b3463a93da3fd8e
ac37969ab486d86a8749a7d32471ff5743f39bd9
6855 F20101206_AACRGP smith_j_Page_010thm.jpg
94821133a61bfeccfa552b69c0689e59
678f6d2da78943abab3862fdb32637fb70828220
F20101206_AACQEB smith_j_Page_102.tif
6e5f8246f22f092c49837e5c1c53e41d
a9fae445f59566438a4d9dadad6f76f875e95fe3
F20101206_AACQDN smith_j_Page_112.jp2
f8f3001898d746159577c0fb2791fb6d
367f4e492c3212a529fcadea0bffda6848f6fc8d
6051 F20101206_AACQCZ smith_j_Page_083thm.jpg
5030f4437a18bb3352e85f9ab9cf19af
ce72fef054922f95c325381d24fb5a54c0f7c620
14818 F20101206_AACRHE smith_j_Page_066.QC.jpg
0e0e5a90cb5191b10b5c689f486ca3d7
2efa12c3fc855ce7a7b3a83b5590ab0c09fcb1c9
75286 F20101206_AACRGQ smith_j_Page_014.jp2
6384d0000a4ce6b7b4affede8b9fa14a
a90af32702d0e57b09f15c35c26265f373b00580
F20101206_AACQEC smith_j_Page_098.tif
0733bff9720453443c4405c41893509c
0029333bff156f755160212f35f2d3e53593b1d7
1023098 F20101206_AACQDO smith_j_Page_092.jp2
99bb2814294264b0195538a8e5956faa
d75cac4961484bfdb03973db5c6c96cc56d41798
719 F20101206_AACRHF smith_j_Page_115.txt
9a1f1b1151ac9586643d6bbddf86b333
23091e23a1f7447e9ecb955f3e04168bd931232c
1895 F20101206_AACRGR smith_j_Page_134.txt
552736d9e9ac5ab544bc3526b2406b9b
26de1fff673d5a18f426af671da34d93af795f45
F20101206_AACQED smith_j_Page_127.tif
251ec2b3bbdba3518d817abebfbe0804
7630e03ec207bc55381c3fed688ee2dc522aea32
F20101206_AACQDP smith_j_Page_060.jp2
b66bb0acc5dbb5fd93ebc36b415c8638
75c19647bb7e2c57b74e6cc5cb9da1a5960f9768
73719 F20101206_AACRHG smith_j_Page_135.jpg
77385e6cb68883e9f009bb2792206ae1
7b1bed021a717ef5b4eb0897adc3e4b5a95f542b
F20101206_AACRGS smith_j_Page_139.tif
455d4289b030074a46e0591f8812f6fb
1dda539080c2771e80b5e1be7aed3cd7299e6480
1764 F20101206_AACQEE smith_j_Page_044.txt
a60395afadcb1dca5d61da39373242ae
0795c0669f2a00e61ff66f56790038670caee447
F20101206_AACQDQ smith_j_Page_047.txt
ba7df2bef215ea1812957fca2395f9c6
4d5e33ba09e4da689630fa0a75023392b2edd107
6348 F20101206_AACRHH smith_j_Page_068thm.jpg
a51100bdbd521dc7fbc28c78a92a8299
86b3eef25d63fe51694e32d70aa8d245a5069f73
F20101206_AACRGT smith_j_Page_073.jp2
993c6a6ce6385c1a67418ad63d7f8bb3
471d16ad4fb60c4db569506aa23aa09938fd12fc
922904 F20101206_AACQEF smith_j_Page_119.jp2
ff24b3a1c59065c645ba6636e656902b
e3b1d70753de3f4f1e3de83545a91ed0695c9c5a
F20101206_AACQDR smith_j_Page_117.tif
44a58d85690acd5f98ea3b766045161a
12b38f71d86ab83222a9d6fb39bd4c576ebf2fa9
1051961 F20101206_AACRHI smith_j_Page_110.jp2
ab2e5486e1f0ca99268252bc7239f702
f8ea5f59f31f082b7f9714e7a1890bca8706b54f
23736 F20101206_AACRGU smith_j_Page_057.QC.jpg
917f0d9408a5e5507ac878e13c8ea2c7
9a50c7d38c6fdff844c7fc5c49ba5f5db3db3b04
21041 F20101206_AACQEG smith_j_Page_104.QC.jpg
de86fe6d6de1ad365fd84d90e6b2358f
cecafb04a1bd28fab20fa7d38ca07cf88c00a4b6
18694 F20101206_AACQDS smith_j_Page_039.pro
3895b639f55765443dbd2bf8c3256c36
a09be8bf80b5d07b7205bf031d476a37cb325a56
F20101206_AACRHJ smith_j_Page_136.tif
213ce4c759024cf974242d99d14e1a7d
6c28659eadbee275ea2eeb3212507246aee6d2fe
92 F20101206_AACRGV smith_j_Page_002.txt
e7188df5a2c41eea09a480fcdd118d52
051dd95e87dfc448740bf4686992879d6bb1f3c4
F20101206_AACQEH smith_j_Page_012.tif
5b7d353b85adc5b57f1d14a031985cf1
397c5c7f68b468afea42b31b3300066e36ca891f
2290 F20101206_AACQDT smith_j_Page_021.txt
5bf4431029e2a0b068ead4aab70b930b
3f7eb741e3ba72316f3505f5d9a0667ee94e25ad
6954 F20101206_AACRHK smith_j_Page_076thm.jpg
8d4cfc3ebec04b6029e61d4706abb63e
d612ebe96eefde29e81b97a6d700262071986c4e
89051 F20101206_AACRGW smith_j_Page_110.jpg
272b73ccf5504d053e96642281a1e4c3
4ebe6149bf0a5357550e8270da769dfe3dd65dbc
23315 F20101206_AACQEI smith_j_Page_137.QC.jpg
0826cbe0798ea4e2601c2eb61d0cbd42
6953734b9a873a04d354b95d7d627c22bc3dfade
25905 F20101206_AACQDU smith_j_Page_061.QC.jpg
3bed3088f518639737b81b22528e4f22
bf8ec63f3512c9cd79dd43480e6eadfe84f19bbc
576 F20101206_AACRIA smith_j_Page_004.txt
e1478b39d4cd6f3cc4ff9c9e9472b5a2
00a2c0e50ed43dbf9010251be0bda9258d482c5c
17517 F20101206_AACRHL smith_j_Page_138.jpg
c7bbbfc2ad01d42406140814f76051dd
56d399eb118993c04bb65291d1a0a7bdcaf40d66
23903 F20101206_AACRGX smith_j_Page_132.QC.jpg
cee463362776a1db2b2e26620da600ef
3112935c82f0f58b78dd99a2aeffcadc7207babf
26341 F20101206_AACQEJ smith_j_Page_072.pro
ff1c0fc82cc650e4a4b1ddb3d0b852b7
624bdbc6d75e864642093fd5a8cd5aea86b46add
6729 F20101206_AACQDV smith_j_Page_118thm.jpg
b02080cc6722e060788a387ff65c7eae
f38efe8c37aef326dcdafaa85cdce1b4e9a5d9f4
F20101206_AACRIB smith_j_Page_131.tif
cd53d3c21d068ee27e378fbb9959b124
b57688909283222a09e000d21d1831cb8e19b48b
F20101206_AACRHM smith_j_Page_030.tif
f76be580dd723ae58b233e61eb209276
98c432a7573d8ca5a22389fee13a555aaf95a404
6900 F20101206_AACRGY smith_j_Page_036thm.jpg
a919f19480a92424741fb619996b34d3
f90824e59e55a5468d7710ac63f478ccd95bc802
38176 F20101206_AACQEK smith_j_Page_125.jpg
563b17c1d9e6ad3f69d5683c51d24af3
461ecf25218c62764f79d257152dafc71991a615
F20101206_AACQDW smith_j_Page_025.tif
524081a122a9ed5cd16b9410a4823fb2
ee87640fda65ddb0cf2475ccdbcbc96fa4a9f821
58656 F20101206_AACRIC smith_j_Page_110.pro
f650bdfff8b87be4724ad94552758145
056d0edc346ba6e59b50f3b733ba8a200cfb24e9
F20101206_AACRHN smith_j_Page_020.tif
b226257736e906838f746122f7cca0ca
fff7a691d50e30b0def0a9e7c269fc8374460475
F20101206_AACRGZ smith_j_Page_105.jp2
b4b0655803748188b2862d11470b2c99
74a637be0bc3cc8d54e9407c13d46c5fc1fbb9c6
6671 F20101206_AACQEL smith_j_Page_110thm.jpg
22166328d1c49026abd3233e4f25a581
cce3dc4c8a1cc746c290c4b286c5e3bd9368a470
32259 F20101206_AACQDX smith_j_Page_004.jp2
53a97d408c40dd96831d6739dc5ba26a
0f071579fd56b77d3681fd8974a4f0f708a38cf0
6314 F20101206_AACQFA smith_j_Page_051thm.jpg
60eb61e989f9c720fc27de454260fe93
dae2e0a639aa820acc982f86117c9b23ea9cfdea
1673 F20101206_AACRID smith_j_Page_112.txt
b67a8214a1a10423710905347bbf16c1
0e728556bca314b7714ce3721037980b1928b245
30838 F20101206_AACRHO smith_j_Page_043.pro
ba80431ac97e94b290c040a4a52376c4
aae0c273ef14f9af6d415494dc5d3c8f02b0ecf0
2342 F20101206_AACQEM smith_j_Page_019.txt
61cbf8af7200e43257bae44ea40b4b5f
3edaeb67de60030835891c4a7bede32c18cef8b9
88326 F20101206_AACQDY smith_j_Page_124.jpg
fc24f8d184183f9b102f4097ca699450
c2b30c002975939ef584c59bdf76a4cddd46f26f
F20101206_AACQFB smith_j_Page_055.jp2
4fec090443e66902edccdcd8646cf427
39e082666df2964e82ed0f8a5e680e07460061fd
4474 F20101206_AACRHP smith_j_Page_071thm.jpg
35cad908a275e65d53fd23b7ef1cfde3
a927f0c7ad3331cdf83e98a3f002a6e2c124a63c
74452 F20101206_AACQEN smith_j_Page_051.jpg
612950ab6943640bb3d349959650191c
a96162117217bdd298bcddc401931c1a3d141e38
73656 F20101206_AACQDZ smith_j_Page_067.jpg
992e773d0c6ea99f433a2379cb5e656d
7511302f8538e78b406bbd863e10453a6d743720
2194 F20101206_AACRIE smith_j_Page_011.txt
7c1294e019b29dff59bf9ec1edbb6dc1
185df544c2b92efc0c3e49aee7101305295923d6
6878 F20101206_AACRHQ smith_j_Page_073thm.jpg
469f49c334ab7a61175da1b9bb67ba30
bcbd59dfd9a6243aac1ec087b11c6e2c39baf435
89944 F20101206_AACQEO smith_j_Page_060.jpg
12f4cdaa3535b1e58fca6e8385f1e736
c2ab2847156bdfccfc2eab83b270f7b735bbfce2
6513 F20101206_AACQFC smith_j_Page_061thm.jpg
59a497583ebcd2df042f36d0984c92ba
118f84276a28803bfd76210b10633f534181b692
F20101206_AACRIF smith_j_Page_050.tif
e5192d98aadf3e60696293eb2c271e6b
3ea6a87041e65f79aa406f0ffb1e7a5ab9d99b23
72650 F20101206_AACRHR smith_j_Page_052.jpg
3a12befe90d7b8a0d862a21401d3fa32
4ef225dfa37a2e16ad683e46ecc51fde5c77979d
4982 F20101206_AACQEP smith_j_Page_131thm.jpg
21f17ca702a1d56df516b8a6473af4aa
35696195af5ad7aa1d0aa2f9b6e0df68d5021b15
F20101206_AACQFD smith_j_Page_009.tif
807e15216e3d6fc9fd41330529a8463d
9ba79ffa30cb343e01478daf1797d095e8bdc3b2
82872 F20101206_AACRIG smith_j_Page_132.jpg
c1cdd584fc7a13fd403ecf5b80d17dfd
7416b0178243fc19c33607e3a525a571873a2e81
26631 F20101206_AACRHS smith_j_Page_050.QC.jpg
5d10874efbfc56e05aacdcb71b55b303
bc6b5b61b04fc31837e82f9162255c5ef696c471
1404 F20101206_AACQEQ smith_j_Page_131.txt
e36e053bd4d894cc4f44202eb4287855
e52daf36831f1a6761451a24a56c14beed373d2b
F20101206_AACQFE smith_j_Page_086.tif
2dc97250cf19b2146b939077a494e3a6
ac32a04dd428b44a183d4b9e7c7535d804a794c1
80493 F20101206_AACRIH smith_j_Page_061.jpg
4e9814134b6fa0e7ad82dc32d1f7afb1
c00af67d7ce26da8ad370f2b1e3f43ca67507011
1051942 F20101206_AACRHT smith_j_Page_008.jp2
98010e2d24ae337f830f01935053f5b4
18634b3c264108475f2552a49629a55ba6ef1eaf
42174 F20101206_AACQER smith_j_Page_068.pro
3f71f3d85cf46c7c6549c76c5442c93e
843fa1638985deac8ccbe14bd016300f3ab0c25d
682297 F20101206_AACQFF smith_j_Page_049.jp2
dc38ba8cbded864e4b13a702548db5d5
bae008d275adf9b4895b06230319d2c84581ce07
F20101206_AACRII smith_j_Page_120.tif
f4eb10902033f2a7650d59197fe52dfd
ac7aa93f80596a45007faaf188812d05ffa8a22f
F20101206_AACRHU smith_j_Page_029.jp2
d4cd2808e6409e8461428ec71c130cc0
78e7588618700aed7099f6a050c7c3a40cf3716d
4267 F20101206_AACQES smith_j_Page_049thm.jpg
385a9da35f0e4815cb225136619c7c1d
202611dab0f09f767a1e8d9e0aea545a0f801193
2325 F20101206_AACQFG smith_j_Page_088.txt
9582b615e1c01329ae7b314641ed5bb1
d754b1ecccda92894c3ac4835dd2d10a286d1628
20130 F20101206_AACRIJ smith_j_Page_069.QC.jpg
dbbdef3ecd9a02a220fbbeeff5802dca
ab0bd4a274e181356e46e0d22c9572a926f5b260
448 F20101206_AACRHV smith_j_Page_001.txt
3cf42d5ca0a1990c53060a4e3278ea3f
758491d57e8a94ee51de70d5655076b57f338db0
51644 F20101206_AACQET smith_j_Page_061.pro
e9eda37ec7b7d93f23e73d5ae60bff24
65c72a2cc2ae9bc5ff90766fbd1f7a185ed59a28
F20101206_AACQFH smith_j_Page_090.tif
f93a9068924a598124df01dbb216282c
cd2ebad92edbf4b33c3e83230513fd748fe602a6
F20101206_AACRIK smith_j_Page_019.jp2
7e5278b16a04733bd82146979823e4db
343aac42e0ee8352a545d163d6349b837c16fff7
5835 F20101206_AACRHW smith_j_Page_136thm.jpg
59eaeaf6af0ad5cc298b192a9743ae3a
5952ff5e202e870c162a7a666ad812349750fc19
52702 F20101206_AACQEU smith_j_Page_119.jpg
157c68d52af2071363335eed41e6cc33
b1a39655d0f495771301b991883a56836844581e
46200 F20101206_AACQFI smith_j_Page_056.pro
5e10d3c0db4c7fbe2b03411419d0f27e
afe5849052a153a3cb436ba4789663704df50d91
1031167 F20101206_AACRJA smith_j_Page_043.jp2
65ff925ad9f64181230743da97816021
b2abc8f3a02c31f3bc7321465fc86e78b39ac2e1
F20101206_AACRIL smith_j_Page_035.tif
899385fba90196048bbbfeaab03430bb
3bbdb8b8c351f5a8d05e74df53edad791e1b3be8
19257 F20101206_AACRHX smith_j_Page_107.QC.jpg
db5ac50868420fbe06fb6650607c1c9a
4c6c20971498e707a635df9fa05ba6326f45c3c6
121851 F20101206_AACQEV smith_j_Page_026.jp2
381626193b1b160a466c4df0ff9ea36a
a0a693773e3433a7e8f65771bb026001943c220f
F20101206_AACQFJ smith_j_Page_072.tif
71b218078eb31e161f048cdea85accbe
76955b86e5c31005778c4033b6b673fce2e9c8ea
26543 F20101206_AACRJB smith_j_Page_122.QC.jpg
2a897a9698ea1517a46a016e71e6686e
406dc28e2d948da40e1c8ac5b32e874707b14937
F20101206_AACRIM smith_j_Page_110.tif
8201d6df5c420919e0197addf05210ba
c837f10e3c28f6e0a1ecb12c469ff7bf089840f7
3768 F20101206_AACRHY smith_j_Page_128thm.jpg
f31dd93b36b0ac7586e3a377ac69c1f4
7019ed78e2d7c515c6ca02f815ace2b8136da907
F20101206_AACQFK smith_j_Page_091.jp2
9b83abd5c357bd5bd727ae61862f26cb
477ee52ab880a758e16496043234d9d016e955ad
12801 F20101206_AACQEW smith_j_Page_121.pro
4f813375d859421bc3fd521180a1b30c
31f1ac82b76ce44604bd3a18d2894f7d3a2d73eb
2131 F20101206_AACRJC smith_j_Page_040.txt
21f2be5f4d1e6a32cc0bca6952194523
96afa6588078a384284e0b302105d5c4652ee4c6
5089 F20101206_AACRIN smith_j_Page_027thm.jpg
f16e06bbd7c3902d638041edd37070b9
3d1fd6afea19b472fced200d670bda27daf1dd52
16008 F20101206_AACRHZ smith_j_Page_014.QC.jpg
7ad48918261b63651ddc658ebaef80c6
a01b47e241af56cc0a0619ac6a4ad86e8209f27a
2104 F20101206_AACQGA smith_j_Page_061.txt
f33e729e67774ecae7d5bd8ac049d357
f3667aaca63c194642f908ec7907cf979574402e
11597 F20101206_AACQFL smith_j_Page_111.pro
266c43066b9f2bf942fe522adbfa2bfd
b56317398e3a9f3490b70ae8ad26d5f1f8b778ec
84510 F20101206_AACQEX smith_j_Page_106.jpg
d63a4bea968555efb3ea7e56866f8b7f
dd82718510a976d39930c4c4713bff48427a7df4
657 F20101206_AACRJD smith_j_Page_111.txt
7040dbaaef920aeb5687a654b6751c74
60c4b9a0534940a936ef855093476c10d4d78b28
2227 F20101206_AACRIO smith_j_Page_106.txt
080fcdff1a92c0fcfc59fb1f7d5da6d5
3eab6be2940c4c1ca116af1e5dfed344515234e3
12842 F20101206_AACQGB smith_j_Page_072.QC.jpg
4ed94215ca8ee7871b34e1290bcbc3a1
0a6d5f380eb58ec4a8dd7e6b9102cb8be666563f
5543 F20101206_AACQFM smith_j_Page_033thm.jpg
596ea7e0492880f06f264462cae9fd38
ea21ae81ea9bcd466f0270410693331c0c4fb8c4
56856 F20101206_AACQEY smith_j_Page_018.pro
00d2897458f2f4467e3b8fa06348fa5a
48058472b58a5d65d21638d80607d0a2bb0aa43a
26336 F20101206_AACRJE smith_j_Page_114.QC.jpg
91703e6318c2840d967d2e06f111d787
9f5fbe80998c306cbb8c0bd4e7a8744a6d221d01
F20101206_AACRIP smith_j_Page_065.tif
560d7b28ff86b02d33d397378defba67
66ec0b869f01f05abfacea97c9ead64840561795
2147 F20101206_AACQGC smith_j_Page_053.txt
eae4a31033ca63ffecb7748ca6f35f80
275112219bbec643ed38b42af6cf3a1fb6eade7b
F20101206_AACQFN smith_j_Page_028thm.jpg
1a75c459c07c30a2aaf2298253ceb4ef
290f2780b3dae545cc0c8c1c1a2a44a495fdef60
F20101206_AACQEZ smith_j_Page_057.tif
c258097ac2e4e1762c3231d557749272
f2ea675ca7409d5681528a03a8fe6ebdffa4cebc
4362 F20101206_AACRIQ smith_j_Page_121thm.jpg
6f0823d0022aa05c548f37bc970d57b8
10a48e20a0215ec6631386e6bc333f34d6e31cc0
1738 F20101206_AACQFO smith_j_Page_043.txt
8ae36f00d31e1136e6c257b9b816fecf
b7f00b72e1e370b00894511406e0be2aa2ac2064
1226 F20101206_AACRJF smith_j_Page_007.txt
410c53b94cc2f05da863f1ef9e9df701
24eac4a4b72535a9603096520fa80c58e5989cfc
F20101206_AACRIR smith_j_Page_087.tif
b773ab19e08e3afd1de607de029ff829
dcbf37b3d39427c8344bd4830bf62e1bb422d0cb
2343 F20101206_AACQGD smith_j_Page_118.txt
8bb58442636cb4066a8efb8d4fd385ea
579c37bb1fed6721a1f82e7528b8a3301ab29a4c
59541 F20101206_AACQFP smith_j_Page_019.pro
283ce7cbdbf63604022ff431735ecb4e
7aa50698cf909fb6130a76cd9e1156027bb59cf4
F20101206_AACRJG smith_j_Page_104.tif
7c9531d283d52acd7a6b9728b077f451
4c822480ded309de7c93fc8d62e4f682c6a83f8c
F20101206_AACRIS smith_j_Page_085.jp2
2d882af8f57749e42b2aff8229b4cbb3
cb9e86ae134a2b46851c879233fdf75ba8bcbabd
22415 F20101206_AACQGE smith_j_Page_100.QC.jpg
dfde6bb5a12b4c2b9621e57522ebf4fb
27c1e20600fc385653c6d787c162341ef348a6b0
27020 F20101206_AACQFQ smith_j_Page_103.QC.jpg
1faf4fffdbe542eb8dbde8d0c385da64
ff52494c52d49438c737af5db0b47f8f445b5198
46770 F20101206_AACRJH smith_j_Page_041.pro
8a2d7bf57e6f7c020b83f0d33485955c
f12d95c0e6a376b1733b2f4689e83c9f48112a34
882 F20101206_AACRIT smith_j_Page_107.txt
99c1b669ca8e86b46ff584930a5e82a9
25375c753ced32b60eb56d0d04651e5ea63d3eb8
648892 F20101206_AACQGF smith_j_Page_113.jp2
6ae7ac0126d69de2fa8befaf2db431c5
e12a784c3b3e3d7688dc5ef0dd7f87fa9910f14c
79869 F20101206_AACQFR smith_j_Page_097.jpg
fd0dc1d0b92cc554e137a96b941b6f0c
abdccae927146fffb4ac81c41be7026ef8a81210
76099 F20101206_AACRJI smith_j_Page_040.jpg
d1dfbc0f44e70a7c012e0497b922f784
31c659215f1e0d9af4275939da376ce955af4282
73172 F20101206_AACRIU smith_j_Page_112.jpg
559e6f61555d3092127286e111110149
ad69035994278966aae5d2cb9b29841ccb710231
F20101206_AACQGG smith_j_Page_118.jp2
1609ae1b991752c71d82fa70a988956e
e04d42c3e96da977ef1a1ed56fe8ce7c871fa116
1916 F20101206_AACQFS smith_j_Page_056.txt
856f6a89475ca0e25cd15968d337e5e1
9134c8c12a1d8fcc0e4768cc831d89f75e740728
F20101206_AACRJJ smith_j_Page_056.tif
d0405655a1c1a85f0c3caf25b78dac10
24b795d1dd3b9a2837cdc5a1585ca9f2192fd975
F20101206_AACRIV smith_j_Page_107.tif
ebbb970745d7c9365a29524ca8b0d437
d43b6498f4b39249cc23303c4b13888d90b3ac81
2351 F20101206_AACQGH smith_j_Page_024.txt
8b1d2340c3f214fced9fd7bd9debb519
1cfe3cb0f52fe325119db54543bdd9b7c45e2fb1
6373 F20101206_AACQFT smith_j_Page_077thm.jpg
4706fdfb7a7914e627b0a6173d4bab3b
34f4271b0283ed956d89c2456dedd23607b019b3
1280 F20101206_AACRJK smith_j_Page_057.txt
91b94dca355196146ec35609857407d0
7c541574a867bef29f2d8e97135f07cd3a150edc
1692 F20101206_AACRIW smith_j_Page_081.txt
440fe8cc2d919b9d1d8a0107f0b9dca1
12f344167cc470154cbf63e51ce4486edbc117c0
12161 F20101206_AACQGI smith_j_Page_063.pro
26ce1e64261692c97e1f7edb1fae69f6
29e3313186b38e52c21936e71cf28a35a3718c6b
F20101206_AACQFU smith_j_Page_101thm.jpg
4df3f4000b26e06d126f6a067781d00c
f696fe7792e627695ee1bfa84effabaa995e4702
24016 F20101206_AACRKA smith_j_Page_054.QC.jpg
e68a5dcffb46cc4d000a480fe18fb9ae
90aad968eb0b840b3ecf242bfaa3aede8a5ee8c7
83913 F20101206_AACRJL smith_j_Page_042.jpg
af92cbe8ba81f07590e4e7ddd8d81007
f750374981c728c7d95b367190a794a76375b47f
21916 F20101206_AACRIX smith_j_Page_043.QC.jpg
7215e0e4732d2502ff0624c2d339cedd
e1c23b41ade19738a765be7d7f4cb2e3562961a5
2269 F20101206_AACQGJ smith_j_Page_140.txt
b2bb2f5ccfa9af6bb47853021b8ce52f
a80427d28ef4a365ec789bce95564a923039dd95
F20101206_AACQFV smith_j_Page_114.jp2
054f7c19b066a7a6d6993e6b395eb10e
c919fcab15a4bfe2644c04329dbca2af1901f556
F20101206_AACRKB smith_j_Page_139.txt
7f9138780845d9958b3e7338d86eff21
f61f2b04517e6fec293cc610c1910a93e742f164
F20101206_AACRJM smith_j_Page_040.tif
087826cf9767e35e3137a780d38d63f2
4582040fa0985c5865d067554fe6c7a6a17c1246
F20101206_AACRIY smith_j_Page_091.tif
c3aebb6c92296e083278d7296c603dc6
88b5240d02d80da687ee15171f4a125fa48b4ce5
6030 F20101206_AACQGK smith_j_Page_047thm.jpg
4c733f30ae01fc94d5c0c8e9537878be
1b98c917b45fcc8a203bebbb26956593abba6f4f
1212 F20101206_AACQFW smith_j_Page_078.txt
34f925a6da2d9fa20629e96a20e626ea
e35ac98be5d55ef4730547ed00ecc7da59328600
1192 F20101206_AACRKC smith_j_Page_049.txt
cd9301d92b1fb9a89a374af851ebb1d4
867cf444bbe492e72ecf84e85f46d97dc89e0ca8
5747 F20101206_AACRJN smith_j_Page_138.QC.jpg
3d919c64c9176afee1baf9132c7b83bc
7dc9d433080de1f9faedc90a2b28676414525926
38415 F20101206_AACRIZ smith_j_Page_044.pro
d859e8a8736c1ec53776d1c05417b722
0d37e9e9a67fd0411423dc2ae5ebd7ca2dfe22b0
2458 F20101206_AACQGL smith_j_Page_058thm.jpg
246584e32c5919e7dcbb1fa9e08b83af
e7f4739b2a646c9e9870a58b5e6324866e81c13a
6563 F20101206_AACQFX smith_j_Page_042thm.jpg
51bb7f0934b5ddb7acf5ed8801901928
be9e7dca7df4d11dbbc1d75c6c4e3ac6d594b218
F20101206_AACQHA smith_j_Page_109.tif
fcc0cf1068e50de2b312bd51c5ac9081
e7ea40e8489dde317d4666bb6a82e37cfe2fd2fe
50287 F20101206_AACRKD smith_j_Page_038.pro
6c0fb5935802cfd69e37ea8b54498fc6
35260f7406ba4c0710b26520b86d5e261f59c985
76749 F20101206_AACRJO smith_j_Page_015.jpg
1aa4e63d0d6e4658ab2920aeff23ae59
70163af00e126c27e9ed734ce47c84d093ce7aae
53530 F20101206_AACQGM smith_j_Page_077.pro
324bd18903e7699e6137ac387e86c062
1efa96f0424a42aa1e544ae642cbc07cb8854ae8
1897 F20101206_AACQFY smith_j_Page_080.txt
51f495961869599b591c7b3aa5dc93f2
c61a4b4b2909c02b940bc8adbc9bb7cb2bf94f20
4881 F20101206_AACQHB smith_j_Page_099thm.jpg
3ba0760158c8acab4dc0dcc50f82c390
955fd17759536b15f51415c307411bd6c670f664
F20101206_AACRKE smith_j_Page_063.tif
76d7529dcce7a4cd5e0680ad39746c93
12e8c119175751de4d1e04de522657668aa7b72c
22383 F20101206_AACRJP smith_j_Page_135.QC.jpg
c11a01e737ba032b92507a61eec84a6e
de792a7888a2d46a20bac23e2d8b40f0db9f22d2
61644 F20101206_AACQGN smith_j_Page_120.pro
dcd31576d1b04fc632d849fdd00c5b29
09ad9b38dc2bdb5dfcc88658c47c32a027758e50
F20101206_AACQFZ smith_j_Page_124.tif
fc7b1dcb45f7afce2dc4120f0f0c2cf4
726853aa4aa1b3d7368af9424e6b09f5edab7aa8
40317 F20101206_AACQHC smith_j_Page_072.jpg
7b1cbf3d4a2575601c01f82a59610eb3
61aa2bba200359b194afdf3ebd6003105908e8a3
55093 F20101206_AACRKF smith_j_Page_035.pro
e79274dae9c271ac4117d2908acdb019
aec39dd1db65897526f207dafccc806129713034
438 F20101206_AACRJQ smith_j_Page_066.txt
645c0ee88215f3f7a84f5e35116a3b16
63af66e8d4bb7a90e34b9c09ac76d63e2c1933ce
11277 F20101206_AACQGO smith_j_Page_003.jp2
1719ce811f12c7458b0a581d530b382b
f1ae923d331735f30e329ecd2172ec793f001c46
57716 F20101206_AACQHD smith_j_Page_074.pro
cf4ffeeb902294baa150951dcc2af209
13324cbe2f25c013997393cc22da7c17f4061b08
F20101206_AACRJR smith_j_Page_106.tif
df94be491db41b7b278610976cf040b8
e1b609f90ad783f023f738faeb4f22c7c4e71f45
23319 F20101206_AACQGP smith_j_Page_112.QC.jpg
1fd1f3c4248e9cde57c90f341068f453
a9eef77d22c2ab7b7b6fe74bead9aa43f69392fb
F20101206_AACRKG smith_j_Page_123.tif
91dc7a8bc21c69e7792c9c48ec98fa61
4ca0df35984c1867c5313135252546f7770fcd7a
57336 F20101206_AACRJS smith_j_Page_045.pro
228d639a2bf22869fde834d5c6bede99
db526462e0f755816525103e051d1065b5139d33
6802 F20101206_AACQGQ smith_j_Page_025thm.jpg
453c11f589e3e00c99eb91163148f11a
48bb5b40feaaa977c14c174a7d4d96644b2928be
1791 F20101206_AACQHE smith_j_Page_003thm.jpg
a4ce6884115d479bb3dc6bcc55cc4bce
b51bdd608f29f6078e11eebe965481e77f9da8a5
6342 F20101206_AACRKH smith_j_Page_037thm.jpg
9a9f0c95929c98f7d6b0bd7671e0e84c
739b5b236e2e795df9d2deef2c6d9197e017bc3f
F20101206_AACRJT smith_j_Page_138.tif
2a1dcacbce88e1948e97108e3795452c
760e2200ebf734de1ed35e7c1da9dff6e70f0e76
24513 F20101206_AACQGR smith_j_Page_062.QC.jpg
3c5769ec9570fcfb55df83e3fdcd9ab8
85cd3a28b8ec0bcf0d67d285cb43b3d2a2545bd8
2034 F20101206_AACQHF smith_j_Page_136.txt
9f99c1ace078abcaddf1621704f81be0
db01ca12991b9251a22423d99cec10d197285f90
1552 F20101206_AACRKI smith_j_Page_129.txt
784f90e564558dfff7830d3ac03d7408
e79c492998efd72ce65133b82bc67215666daad3
21705 F20101206_AACRJU smith_j_Page_011.QC.jpg
5cf0a322731ca3ec09e807af6cb0c7ec
ce0c67ed090fe2ba2bbefc96236519fb4f3d5d30
81964 F20101206_AACQGS smith_j_Page_016.jpg
fc6704aafa555b97d6e829392c0512d0
6708450c2df80821376561b5c288cdb5c7429221
1006752 F20101206_AACQHG smith_j_Page_111.jp2
991d7bc239ae752206bdb48d722b0207
e06b4d32949976f38bbfd613505728e642bd533f
5694 F20101206_AACRKJ smith_j_Page_012.pro
9874b5d78cfafc1a657b876b1fd4658d
05c1c6b277c8f547462a1e93e24450da48673a8a
F20101206_AACRJV smith_j_Page_010.jp2
ba8043449d10c6d06aa50545443feaf4
2970419c75b4cb1f0e79b72c1223caa25a511a48
80814 F20101206_AACQGT smith_j_Page_130.jpg
3d6d795123aca735ad7a910599b28fac
b4fedb9d3c71ae3603463d5fe2321acef2c79d2f
75068 F20101206_AACQHH smith_j_Page_100.jpg
9c785dcf62da7c517f8705e6e9d858f9
ef9171f89aa7a6dd47ebfc85b0e9069aea54cbb4
2312 F20101206_AACRKK smith_j_Page_045.txt
41c0bfc56470e3694364c8ecb84e69c1
4cfb6e2100fe63ced1a7f08f528bc23eed0172e0
79984 F20101206_AACRJW smith_j_Page_008.jpg
169b2d60dd8e3e8e7e9960dca22cd572
9672a5becb2c1c5486e706d66578132664d641c6
2411 F20101206_AACQGU smith_j_Page_124.txt
148821ba83dc4c11b465e9a3ad30a8de
6f750bd665ff3c4f3bafc5d956132be0a4474896
6017 F20101206_AACQHI smith_j_Page_132thm.jpg
09ffb3be73eeddc8578eaf0be2aadd5e
4d359e80529798d4a7ab638ef16eea8fb070cab6
88673 F20101206_AACRLA smith_j_Page_045.jpg
310b6c27a9c6c52908b9ff7d78c79281
96283608c51e0b37720123668deb071e70ca89d4
54473 F20101206_AACRKL smith_j_Page_084.pro
6bfb28d128b2b0d9e0e5ea042526294a
318c3ab85d750d063d0acd643397fdb62c509809
2413 F20101206_AACRJX smith_j_Page_132.txt
4e54d2208ae34092847ab2d58918b0ce
2c528d09fa7cfa130347aec4cec9cc16890938b5
88191 F20101206_AACQGV smith_j_Page_079.jpg
a6cfb6c75bc82b754ecc70b9f900f71e
d053cb8db4180d7b24a13280e927db03a142833d
6817 F20101206_AACQHJ smith_j_Page_009thm.jpg
3b01828b19d970b4d1d55fda82b36a95
826841c892232643f14e5c7ac5b18a66e954995c
51122 F20101206_AACRLB smith_j_Page_136.pro
a7aeb5b7115a030e102beb4cfc5f6b16
78e5770bb018b4f7e4fb7f76767f0cc2e86a15f9
F20101206_AACRKM smith_j_Page_003.tif
9a8a1efc8904fbcbd6eeb0c8488cd577
38fec34114036fe6a2da59af6f5d2ca770e87051
77310 F20101206_AACRJY smith_j_Page_026.jpg
e2f691486e588743840a46855d0df544
fa8ddfe6af6116f2509b68e0c00b17b9579c9f86
F20101206_AACQGW smith_j_Page_005.jp2
be8d8d50cc14ed8eea95da4b5bafd3f1
59fec393512d3d0247c8759ba228d4430d90f6b5
59497 F20101206_AACQHK smith_j_Page_075.pro
6f830dd33ade47f401f57fdd8cf8e367
7c3fddaa3780c3508017627c45fd5b1c91bc60f2
27192 F20101206_AACRLC smith_j_Page_096.QC.jpg
32ca9e713a3cbd2bcc5b9b89699aa6f8
9b1d666b353de12b2164404f3a28bf2c101ed885
20976 F20101206_AACRKN smith_j_Page_109.QC.jpg
c4d78bafb8a9a9b423fc5ef9033e8407
223f256fb371d7649be17b92757c1884ca569567
6094 F20101206_AACRJZ smith_j_Page_092thm.jpg
c0f60df20d15b45d627323e9f40eccec
9c04ed23da135ec816dd39805c990d88df0b88b3
2330 F20101206_AACQGX smith_j_Page_018.txt
a053999ec3ab9dc983ee92a42c095896
f5a462e4cd67289a5d5d8b25d65afb3889ac6665
F20101206_AACQIA smith_j_Page_026.tif
bfdba436f5be29fddb5377b5684e86c9
5b9162ab87ed07440666c121c8ac04964bc6ca56
F20101206_AACQHL smith_j_Page_086.jp2
b8c79430037686c2bc892584b6fcde01
14fbaf3191a0de44c246dbdfeb5b18152006b4d5
F20101206_AACRLD smith_j_Page_009.jp2
c2a5586bcb0946be8f5a1ea0631b8471
e830b304718e0178f0063e4533485b514a52f788
24347 F20101206_AACRKO smith_j_Page_140.QC.jpg
69283b3000c788c80dd520c16dcf36b1
65860052e05b7cbd3bb13e786227604e52c2e0e4
F20101206_AACQGY smith_j_Page_006.jp2
a8fcb8f2874d45b5445d527bc6fe9f4e
e958de48769d38c260fa333482585ea57799ca46
6499 F20101206_AACQIB smith_j_Page_054thm.jpg
d8dbdc70ac60e16526bf482155d3c2eb
808732676b628e50187c2a21fdf8fde50eefcc8d
F20101206_AACQHM smith_j_Page_041.jp2
e68899ce47c23cdbcff799c8afafedc0
09810bdea41527977c2f660c671660a115bc03bc
6056 F20101206_AACRLE smith_j_Page_137thm.jpg
f9731433932ecefedafa16e97f0e4706
cad45d6afe9ad3135f574f8c31956b9d5babe860
F20101206_AACRKP smith_j_Page_115.tif
6ca0ce56b1a659f247c0dd7416c7428b
5550ac051bbe653b68581855ece070c938c73f67
6778 F20101206_AACQGZ smith_j_Page_016thm.jpg
7eaf7fba556608161d852e672291bbbf
2fab2dd482504a89396e10288412a2117c5e6cff
F20101206_AACQIC smith_j_Page_067.jp2
f4b9633757ef661649ba7543fa909845
b7d78d4db2496dd0c676db676abe3b9b9897d999
11254 F20101206_AACQHN smith_j_Page_046.QC.jpg
0c61a301b85f702633932531c47a8079
cf236b52faac2e12723874ffaa4b0ff903894f04
1364 F20101206_AACRLF smith_j_Page_002thm.jpg
49fd876b4f9d522b050dd5346b0a73dc
ddf1f19033da08ec5cab5d0915d0b4a0f006ea0d
2368 F20101206_AACRKQ smith_j_Page_028.txt
4f1b559a80046d5e3ba5b89cf824b8ea
f59e4379d0ab5ecdc2bf8a692260aa189af08623
F20101206_AACQID smith_j_Page_071.tif
12ebf145e057b8611aaea89c2b65023b
c431da063dbd7d815bc47fc99813f7c872757bdc
F20101206_AACQHO smith_j_Page_080thm.jpg
45e1e0c234734f4098d55472d333acf8
10a8ab279e47ab567d996f3905cc279b5b4572e8
F20101206_AACRLG smith_j_Page_045.tif
4588b12476ca35514a2eb8d7da05ab38
98e6e278d2cac4acbac7357cd6a5ff4113bb84e5
F20101206_AACRKR smith_j_Page_088.tif
ccee0fe2d0d2ba9aaa5ae92ef3373385
1961f0913f7df35a591229973a5723414ed0bd64
F20101206_AACQIE smith_j_Page_018.jp2
715e49632750c22567b9d16e377174fd
7f0ae04c11603345f6ea77d09bb880a8e4556d71
2270 F20101206_AACQHP smith_j_Page_016.txt
72ab7e2035248d787df6f5a46ace0758
ab42d7e4c4c60a45eb83426f12a3396c9372324e







EVALUATION OF ELECTROCHEMICAL PROCESSES
OCCURRING IN THE CATHODIC REACTION OF SOFCS



















By
JEREMIAH R,. SMITH


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

2007



































S2007 Jeremiah R. Smith



































To my grandpa, Collier Carethers, and the rest of my family and friends who encouraged

and supported me in good times and bad









ACKENOWLED GMENTS

I thank my friends and family for continually supporting me. I thank my classmates

and colleagues for helping me acquire all the information needed to make this possible.

I thank Dr. Eric Wachsman and Dr. K~eith Duncan for providing the direction needed

for my work. I thank Dr. K~evin Jones, Dr. Mark Orazem and Dr. Juan Nino for many

helpful -II_ -r;un~- I also thank the United States Department of Energy for funding

under project number DE-FC26-02NT41562 and DE-ACO5-76RL01830.











TABLE OF CONTENTS

page

ACK(NOWLEDGMENTS ......... ... .. 4

LIST OF TABLES ......... ..... .. 7

LIST OF FIGURES ......... .... .. 8

ABSTRACT ......... ...... 11

CHAPTER

1 INTRODUCTION ......... ... .. 1:3

2 BACK(GROUND . ..._.. ...... .. 15

2.1 Solid Oxide Fuel Cell Basics ......... .. 15
2.2 Materials of Interest ........ .. .. 16
2.2.1 Yttria Stabilized Zirconia as an Electrolyte .. .. .. 16
2.2.2 Lanthanum Strontium Alangfanite as a Cathode .. .. .. 18
2.2.3 Lanthanum Strontium Cobalt Iron Oxide as a Cathode .. .. .. 20
2.3 The Cathodic Reaction ......... .. 21
2.4 Impedance Spectroscopy ......... ... 25
2.4.1 Measurement Details ......... .. 25
2.4.2 Data All lli;; . ...... ... 28

:3 ERROR ANALYSIS ......... .. .. :31

:3.1 Introduction ......... . .. .. :31
:3.2 Experimental ......... ... .. :32
:3.3 Results and Discussion ....... .. .. :34
:3.3.1 High-frequency Artifacts in Impedance Data ... .. .. .. :34
:3.3.2 Correction of High-frequency Artifacts in Impedance Data .. .. :38
:3.3.3 Repeatability of Measurements ..... .. . 45
:3.4 Conclusion ......... ... .. 48

4 ELECTROCHEMICAL PROCESS IDENTIFICATION .. .. .. .. 50

4.1 Introduction ......... . .. .. 50
4.2 Experimental ......... ... .. 51
4.3 Results and Discussion ......... .. 52
4.4 Conclusion ......... .... .. 57

5 TERTIARY PHASE FORMATION ....... .. .. 59

5.1 Introduction ......... . ... .. 59
5.2 Tertiary Phase Formation ......... ... 59
5.3 Experimental ......... .. .. 61











5.4 Results and Discussion . . . .. 61
5.4.1 Electrochemical and Microstructural C'I I) Il-terization .. .. .. 61
5.4.2 Compositional(l CI. .) .terization ..... ... .. 69
5.5 Conclusion ......... .... .. 70

6 THE RELATIONSHIP BETWEEN CATHODE MICROSTRITCTITRE AND
ELECTROCHEMICAL PERFORMANCE .... ...._ .. 7:3

6.1 Introduction ......... . ... .. 7:3
6.2 Experimental ......... ... .. 77
6.3 Results and Discussion ......... .. 78
6.3.1 Effect of Sinteringf on Microstructure .... ... .. 78
6.3.2 Effect of Sintering on Impedance .... ... .. 81
6.:3.3 Effect of Microstructure on Impedance ... .. .. 88
6.:3.3.1 Series model evaluation ... .. .. 88
6.:3.3.2 ?.-1. .1 model evaluation .. .... .. 95
6.:3.3.3 ?-. -i. I1 relation to pore surface area .. .. .. . 100
6.4 Conclusion ......... ... .. 10:3

7 EVALUATION FOR LANTHANITA STRONTIUM COBALT IRON OXIDE .105

7.1 Introduction ......... . .. .. 105
7.2 Experimental ......... .. .. 106
7.3 Results and Discussion ......... .. .. 106
7.4 Conclusion ......... ... .. 122

8 CONCLUSIONS ......... ... .. 124

APPENDIX

A ANALYSIS USING TUNNELING ELECTRON MICROSCOPY (TEM) .. 126

B FOCUSED ION BEAM/SCANNING ELECTRON MICROSCOPY (FIB/SEM)
ANALYSIS . .. ..... .. .. 129

REFERENCES ............ ........... 1:32

BIOGRAPHICAL SK(ETCH ......... .. .. 1:39










LIST OF TABLES

Table page

:3-1 R and -r values with their respective errors (standard deviation, a) for raw data
measured at 900 oC'. ......... .. . :38

:3-2 R and -r values with their respective errors (aR~,) for high-frequency corrected
data measured at 900 ol'. ......... . .. 42

:3-3 Ch1 I4;-~ in polarization resistance, Rp, constant phase element coefficients Q
and c0, and time constant, -r front deconvolution of raw and corrected data for
adsorption (1) and charge transfer (2) processes. .. .. .. 45

4-1 Select elementary steps of the cathodic reaction in samples sintered at 1100 oC'. 54

7-1 Polarization resistance values in as for various elementary steps of the cathodic
reaction in lanthanum strontium cobalt iron oxide samples sintered at 950 oC'
and measured at 700 o"C at various oxygen partial pressures. .. .. .. .. 11:3

7-2 Properties of the various cathodic processes in lanthanum strontium cobalt iron
oxide. ................ ... .. 119










LIST OF FIGURES

Figure page

2-1 Possible reaction pathi-ws- for a platinum cathode on electrolyte system. .. 23

3-1 Photograph of a typical sample. .. ... .. 33

3-2 Raw impedance of lanthanum strontium manganite (LSM) measured at 900 oC
in air. ......... ..... . 34

3-3 Model fit, including the 95' confidence intervals, of the raw data for an 1100 oC
sintered sample. ......... .. 37

3-4 Standard deviation (o ,,) versus frequency determined from six replicates of data. 39

3-5 Raw and high-frequency corrected data for LSM on yttria stabilized zirconia
(YSZ) measured at 900 oC in air. ........ ... .. 40

3-6 Real fit generated from imaginary impedance data and KKE relations for LSM
on YSZ measured at 900 oC in air. . ...... .. 41

3-7 Real variance / imaginary variance for raw and high-frequency corrected data. .43

3-8 Effective capacitance calculated for LSM sintered at 1100 oC. .. .. .. .. 44

3-9 Modeled electrochemical process occurring in LSM on YSZ measured at 900 oC
in air. .. ....... .... . .... 46

3-10 Repetitions of an impedance measurement taken at 800 oC in air. .. .. .. 47

3-11 Resistance values for repetitions of an impedance measurement taken at 800 oC
in air. .. ....... .... . .... 48

4-1 Impedance spectra for 1100 oC sample at various measurement temperatures in
air. .. ........ ..... 51

4-2 Deconvolution of an imaginary impedance versus frequency profile into various
individual contributing processes. ... ... .. 52

4-3 Temperature dependence of the separated contributions in LSM on YSZ sintered
at 1100 oC for 1 h in air. .. ... . .. 54

4-4 Impedance response of an LSM cathode measured at 900 oC with pO2 (atm) aS
a parameter. . .. .. 55

4-5 Dependence of cathodic polarization resistances in an 1100 oC sintered sample
on pO2. .... ........ .............. 56

4-6 Impedance data measured at various temperatures for an 1100 oC, 1 h sintered
sample at 0.002 02. ......... . .. 57










5-1 Complex plane plots measured at 600 oC of symmetrical LSM on YSZ samples
as sintered and after a 1400 oC, 48 h anneal.

5-2 Impedance spectra for as sintered (1100 oC) sample at various measurement
temperatures.

5-3 Impedance spectra for sample after a subsequent 1250 oC, 12 h anneal measured
at various measurement temperatures.

5-4 Scanning electron microscopy (SEM) images of the cathode/electrolyte interface
as sintered at 1100 oC.

5-5 Scanning electron microscopy images of the cathode/electrolyte interface for
samples sintered at 1250 oC.

5-6 Scanning electron microscopy images of the cathode/electrolyte interface for
samples sintered at 1400 oC.

5-7 Impedance spectra for 1400 oC, 12 h annealed sample at various measurement
temperatures.

5-8 High-frequency arc resistance measured at 400 oC versus anneal temperature
for various anneal (temperature, time) pairs.

5-9 Energy-dispersive X-Ray Spectroscopy (EDS) linescan of MnE~a intensity at
LSM/YSZ interface.

5-10 X-ray diffraction of samples subjected to post-anneal sintering..

6-1 Scanning electron microscopy (SEM) images, created using a focused ion beam
/ SEM (FIB/SEM), of LSM on YSZ sintered for 1 h at various temperatures.

6-2 Microstructural parameters as a function of sintering temperature..

6-3 Porosity and tortuosity as a function of sintering temperature.

6-4 Nyquist plots measured at 800 oC for LSM sintered at various temperatures in
air.

6-5 Imaginary impedance vs. frequency plot measured at 800 oC for LSM sintered
at various temperatures in air..

6-6 ?-. -i I1 element equivalent circuit used for fittingf.

6-7 Series Voigft element equivalent circuit used for fittingf.

6-8 Deconvolution of impedance profile from 1200 oC sintered sample, measured at
800 oC in air, using both equivalent circuit models.

6-9 Temperature dependence of polarization resistance (Rp) in air determined using
both series and nested equivalent circuits measured at 800 oC.










6-10 Relation of charge transfer and adsorption polarization resistance determined
from the series Voigft element equivalent circuit to microstructural quantities
(measured in air at 800 op) ... ...... ..... 90

6-11 Relation of charge transfer and adsorption polarization resistance determined
from the nested equivalent circuit to LTPB (measured in air at 800 op). .... 9

6-12 Relation of adsorption polarization resistance determined from a nested model
to surface area per unit volume (measured in air at 800 oC). .. .. .. .. 102

7-1 Impedance response of lanthanum strontium cobalt iron oxide (LSCF) on YSZ
in air at various sinteringf temperatures. ..... .. .. 107

7-2 Imaginary impedance versus frequency for LSCF measured at 700 oC in air at
various sintering temperatures. ......... ... .. 107

73Application of a series Voigt element based equivalent circuit to LSCF sample
sintered at 1000 oC and measured at 700 oC in air. ... .. .. .. 108

7-4 Parameters determined from equivalent circuit fitting for LSCF in air, measured
at 700 0C. ............ ............ 109

7-5 Impedance response at various oxygen partial pressures of 950 oC sintered LSCF
on YSZ measured at 700 oC. ......... .. .. 111

7-6 Application of a series Voigft element based equivalent circuit to LSCF sample
sintered at 950 oC and measured at 700 oC at 0.09 ()2. .. .. .. 112

7-7 Ohmic series polarization resistance from model fittingf for LSCF sintered at
950 oC and measured at 700 oC as a function of pl)2* * ** 113

7-8 High-frequency polarization resistances as a function of p()2 for LSCF on YSZ
sintered at 950 oC and measured at 700 oC. ..... ... . 115

7-9 Low-frequency polarization resistances as a function of p()2 for LSCF on YSZ
sintered at 950 oC and measured at 700 oC. ..... ... . 117

7-10 Parameters from model fitting for LSCF at 0.09 oxygen, measured at 700 op
as a function of sinteringf temperature. ...... .. . 119

7-11 Activation energies of various electrochemical processes for 950 oC sintered LSCF
on YSZ. .. ....... ..... . .. 121

A-1 Tunnelling electron microscopy image of LSM on YSZ interface, with Energy
dispersive spectrometry (EDS) profiles inset. .... .. .. 126

A-2 Selected-area diffraction patterns for LSM, YSZ, and the transitional region. 128

B-1 Setup of FIB/SEM indicating alignment of ion and electron beam with sample. 129









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

EVALUATION OF ELECTROCHEMICAL PROCESSES
OCCURRING IN THE CATHODIC REACTION OF SOFCS

By

Jeremiah R. Smith

December 2007

Cl.! ny~: Eric Wachsman
Major: Materials Science and Engineering

The need for high efficiency and low emissions power sources has created significant

interest in fuel cells. Solid oxide fuel cells are desirable for their fuel versatility. The

cathodic reaction is known to be one of the major causes of power losses in SOFCs, but

the exact manner in which the cathodic reaction occurs is not well understood. The

cathodic reaction was investigated using primarily lanthanum strontium manganite (LSM)

cathode / vttria-stabilized zirconia (YSZ) electrolyte symmetric cells, as LSM is one of the

most studied solid oxide cathodes and the symmetry of the sample simplifies the study.

An in-depth investigation of the cathodic properties of lanthanum strontium cobalt iron

oxide (LSCF) was also performed.

The areas of interest are identification of the individual processes occurring in the

cathodic reaction and understanding how the reaction is influenced by experimental

conditions such as temperature and pO2. Elementary steps of the cathodic reaction

can be analyzed individually using AC electrochemical impedance spectroscopy (EIS).

This characterization technique gives overall polarization impedance as a function of

applied frequency. The output spectra were analyzed giving information about each of the

significant steps of the cathodic reaction.

The effect of microstructural and interfacial changes on the cathodic reaction was

also investigated. These changes were produced by sintering at various temperatures and

times. The microstructural changes were analyzed both qualitatively and quantitatively.










Ultimately, a direct relationship was established experimentally between the cathode

microstructure and electrochemical performance. This relationship was modeled based on

theory involving reaction kinetics.









CHAPTER 1
INTRODUCTION

The turn of the twentieth century marked the beginning of a technological explosion

that has changed our world forever. From the invention of the electric light bulb to the

automobile to the aeroplane, many scientific advances were made which have vastly

improved the efficiency and convenience of life in an industrialized nation. These advances

have not come without cost. Electrical and mechanical devices are required by Newton's

Law of Conservation of Energy to derive their power from some external source. To

date, that source has primarily been fossil fuels for many industries. As technology has

increased, so has our demand for fuel; unfortunately, the amount of usable fossil fuels

available is finite and if we do not increase our efficiency of consumption and develop an

infrastructure capable of utilizing alternative, renewable sources of fuel, fossil fuel sources

will eventually simply run out. One of the most promising technological advances with the

potential to enhance the efficiency and versatility with which fossil fuels are consumed is

the fuel cell. In a fuel cell, chemicals, which are continually pumped in, participate in an

oxidation-reduction reaction resulting in the efficient production of electricity (38 .~ now

and 60 .~ by 2020 [1]).

The solid oxide fuel cell (SOFC) holds particular promise. The SOFC has been

shown to be able to produce high quality power from a variety of fuel sources including

but not limited to hydrogen, hydrocarbons and carbon monoxide with only heat, water

and CO2 as byproducts [2]. Researchers at Los Alamos National Laboratory have even

proposed techniques for zero emission coal power plants based on SOFC technology [3].

High-efficiency power generation is not enough, however, and the primary focus of SOFC

research is in decreasing the system cost of SOFCs from around 800 to 400 $/kW by 2010

[1]. The research community has adopted a two-pronged approach towards reduction of

this value. First is the development of lower cost materials for SOFC stack construction.

Among these materials are metal interconnects which are made feasible by lower operating

temperatures. Additionally, as operational temperature is reduced, less energy is required










to start up the SOFC; therefore, reducing the operational temperature will significantly

reduce the system cost of the device.

The electrochentical efficiency of SOFCs is often limited by polarization losses which

accompany the oxidation-reduction reaction. 1\any of these polarization mechanisms

have negative activation energies; thus, losses tend to be greater at lower temperatures,

placing lower limits on useful operational temperature. It is accepted that polarization

losses can he divided into ohmic, concentration, and activation losses. Collective

understanding beyond this simple detail is unsatisfactory. A powerful tool for investigating

polarization losses associated with the cathodic reaction is electrochentical impedance

spectroscopy (EIS). Impedance spectroscopy is a characterization technique which allows

for the separation of contributions to overall impedance into a frequency distribution.

Each discrete loss niechanisni is active in a different frequency regime. An improved

understanding of the polarization losses occurring during the cathodic reaction will allow

future researchers to direct their efforts as they work to develop higher kW/$ SOFCs.

Iniprovenient of the fundamental understanding of the cathodic reaction using impedance

spectroscopy is the focus of this dissertation.









CHAPTER 2
BACKGCROUJND

2.1 Solid Oxide Fuel Cell Basics

The fuel cell stack is a device which takes advantage of an oxidation/reduction

reaction to generate usable electricity. The two half-reactions are separated with reduction

of an oxidant and oxidation of a fuel occurring on the cathode and anode, respectively.

For oxygen as the oxidant and hydrogen as the fuel, the oxidation and reduction reactions

proceed according to the following respective equations.


H2 + OX 2e' + H20 + Vo" (2-1)


2e' + 02, + V" OX (2-2)

The fuel cell stack, consisting of the cathode, anode, electrolyte, and interconnect, is

constructed in various designs including, but not limited to planar, monolithic, or tubular

[4]. The anode and cathode are connected by an electrolyte which allows only the flow

of ions and an interconnect which allows only the flow of electrons. The cathode must

be electronically conducting to allow generated electrons to reach the load and porous to

allow gas to flow to the reduction sites. The cathode must also have sufficient catalytic

activity for the reduction of oxidant at operating temperatures. The anode must also be

porous and electronically conductive. The anode must also possess sufficient catalytic

activity for the electrochemical oxidation of the fuel, thus minimizing polarization. Since

the purpose of the electrolyte is to transfer ions produced at the cathode to the anode

and force electrons through the load, any electrons flowing through the electrolyte will

result in voltage loss and decrease efficiency. For this reason, the electrolyte must possess

minimal electronic conductivity while maintaining maximum ionic conductivity. Further,

an electrolyte must be impermeable to the reacting gases. As mentioned above, the

interconnect provides a path to the external load. It also joins .Il11 Il:ent fuel cells to one

another and provides a pathway for the reacting gases to reach the electrodes, while










preventing mixing of the fuel and oxidant gases. In addition to these requirements, all

components must have matching coefficients of thermal expansion, must he stable in

operating conditions, and must he non-reactive with one another.

2.2 Materials of Interest

2.2.1 Yttria Stabilized Zirconia as an Electrolyte

In an SOFC, a solid oxide such as yttria stabilized zirconia (YSZ) is used as the

electrolyte [2]. Pure zirconia (ZrO2) is chemically stable in oxidizing and reducing

environments. Pure zirconia, however, exhibits a phase transition from monoclinic to

tetragonal at 1170 oC and a change from tetragonal to cubic fluorite at 2:370 oC [5]. These

phase transitions occur in the fabrication temperature range for fuel cell devices and are

accompanied by volume changes, which are undesirable. In addition to this drawback,

the ionic conductivity of pure zirconia is too low for this material to be valuable as an

electrolyte. Both of these problems can he addressed by doping with various oxides

(CaO, Y203, MygO, Sc20:3). Of these oxides, yttria (Y20:3) is most commonly used because

of stability, conductivity, and cost [2]. Yttria doping can stabilize the cubic fluorite

structure from above fabrication and operating temperatures to room temperature while

increasing the oxygen vacancy concentration. During anpinlr the Y ions substitute on

Zr4+ cation Sites according to the following reaction written in K~rogfer-Vink notation

which is described in [6]. In this notation, the charge of a substuting species with respect

to the species for which it substitutes (given in the subscript) is indicated by a prime if it

is negative or a bullet if it is positive. If the species is neutral with respect to the typical

species for a particular lattice site, the superscript is an "\x".


Y20:3 z 3 2Yir + V," + :$Of (2-3)


This increased oxygen vacancy concentration translates into an increased ionic conductivity.

Ionic conductivity in zirconia is due primarily to the presence of oxygen ion vacancies in

the fluorite structure. Undoped zirconia has a relatively low concentration of these defects.










As yttria (or other dopants) is added, the concentration of oxygen vacancies increases

[7]. It is shown that the maximum conductivity is obtained at around 8 mol .~ which

is the minimum dopant concentration required to fully stabilize the fluorite structure

of zirconia [8]. The decrease in conductivity at higher dopant concentration is due to

defect ordering or vacancy clustering, effectively reducing the total number of active

defects. At 1000 ofC YSZ with 8 mol .~ yttria has a conductivity of 0.1 S cm-l. The

properties of YSZ are reviewed in [9]. The ionic conductivity of YSZ, being directly

proportional to the concentration and mobility of ions, is a thermally activated process.

(E, a 1.0 eV [10]). For this reason, SOFCs are limited to high operating temperatures,

approaching 1000 oC for YSZ. Other types of solid oxide electrolytes such as Gadollinia

(Gd20:3) doped ceria (C/eO2) or GDC and yttria stabilized hismuth oxide (YSB) have

shown higher conductivities at lower temperatures and are of interest for intermediate

temperature (500 700 oC) SOFCs. Unfortunately, these electrolytes are less stable, have

some electronic conductivity or are simply newer and are therefore less understood than

YSZ. The presence of yttria, which increases the oxygen vacancy concentration, extends

the acceptable oxygen partial pressure range down to 10-:so atm, which includes typical

operating conditions [2].

One of the most common methods for YSZ preparation for SOFCs is tape casting.

In tape < Ih-1;! very fine, uniform particles of yttria and zirconia particles of the desired

composition are measured out. The powder is then mixed and dispersed in a solution

containing solvents, hinders, and plasticisers. The slurry is then extruded in tapes which

are cut to the desired size. These substrates must he sintered to maximize densification.

The total conductivity of YSZ is shown to be dependent on microstructure and on the

characteristics of grain boundaries in particular. The presence of grain boundaries decrease

the total conductivity of YSZ. Several works have used impedance spectroscopy, to study

bulk and grain boundary contributions to the conductivity of YSZ [11-14].










2.2.2 Lanthanum Strontium Manganite as a Cathode

In addition to being an excellent electronic conductor and good catalyst for

reduction, the SOFC cathode must he compatible with the electrolyte of choice,

limiting the possibilities for selection. An SOFC cathode material must he stable at

the high temperatures required for solid oxide electrolyte operation and in an oxidizing

environment. These requirements limit the possible choices for cathode materials.

Fabricability, cost and thermal expansion matching further decrease the pool of materials

possibilities. Strontium doped lanthanum nianganite (LSM) is a popular choice as a

cathode material for use with YSZ electrolytes because it is a good electronic conductor at

high temperatures, has a reasonable cost, has an acceptable thermal expansion match with

YSZ, and can he engineered to have high porosity [2, 15, 16]. Extensive study of LSM has

resulted in a good understanding of its properties [17-19].

Lanthanum nianganite (Lal~nO4) has a perovskite structure. The lanthanum

and manganese ions both have a valency of :3+, while the oxygen ions valency is 2-.

La~nO: is orthorhombic at room temperature and shows an orthorhombic-tetrahedral

crystallographic transition at about :387 oC [15, 20]. Intrinsic p-type electronic conductivity

has been observed in Lal~nO:4 due to cation vacancies. The room temperature conductivity

of Lal~nO: is 10-4 (S c17-1) [21]. Substitution of strontium ions, which have a valency

of 2+, for lanthanum ions causes some of the Mn ions to shift front a :3+ to a 4+ valency.

This substitution can he achieved via the following reaction.


(1 r)Lal~nO: + rSrltn03L~no (1 r)La,, +.MrL laI :of

(2-4)

The presence of AfnkIt (j4+) 10118 enhances electronic conductivity via a polaron

conduction mechanism. As predicted by the niechanisni, LSM conductivity exhibits an

Arrhenius dependence on temperature [15]. At 20 mol .~ Sr, LSM possesses an activation

energy of 0.10 eV [18]. The conduction of LSM increases with temperature and Sr content

up to 1000 oC and about 20 mol .~ respectively [18]. At temperatures greater than









1000 oC, conductivity levels off, -II__- -ru;~!_ a semiconducting to metallic transition [18].

(11. 1 .III..: conduction behavior is characterized by a negative temperature dependence

in contrast to semiconductor conductivity which increases with temperature.) The Sr

concentration also affects the thermal expansion coefficient of LSM. It is possible to

tailor the thermal expansion of LSM to suitably match that of YSZ by modifying the Sr

concentration [22].

At 1000 oC the electronic conductivity of LSM shows little dependence on oxygen

partial pressure at higher oxygen partial pressures [15]. A critical oxygen partial pressure

exists, below which the conductivity decreases as a function of the fourth root of oxygen

partial pressure. The abrupt decrease in conductivity at the critical oxygen partial

pressure is believed to be due to the decomposition of the Lal~nO3 phase. This critical

oxygen partial pressure shifts to higher values when the temperature and/or the strontium

content are increased [2].

A variety of routes are used for LSM fabrication, including, but not limited to solid

state reaction [23, 24], sol-gel synthesis [24], laser ablation of dense LSM [25], spray

pyrolosis, auto-ignition [26-28] and co-precipitation [29]. In solid state processing, powders

containing acetates or oxides of the desired cations are mixed in the stoichiometric ratio.

The powders are then mixed and subsequently milled in a solvent (acetone) solution. The

slurry is then calcined to obtain the desired crystal structure. Sol-gel synthesis differs

from solid state processing in that a gelling agent is added to the aqueous solution before

calcination. In spray pyrolosis, aqueous solutions of nitrates with the desired cations

are mixed in the stoichiometric ratio. A fuel additive is then introduced to complete

combustion into metal oxides. The solution is then ;II1 li-- II onto a surface and dried.

Calcination is performed to produce a powder with the desired chemical structure.

Cathode deposition techniques include screen printing, plasma spraying [30], slurry coating

[31] and other techniques [32-34]. After deposition, a subsequent anneal is necessary to

ensure proper adhesion of the cathode to the electrolyte.









2.2.3 Lanthanum Strontium Cobalt Iron Oxide as a Cathode

As mentioned previously, the operation temperature of YSZ based SOFCs has

been limited to above 800 oC. To address this issue, thin electrolyte geometries and

higher conductivity electrolytes have been developed reducing the practical operation

temperature. As a result, the performance of LSM has now become the primary issue.

Thus, other cathodes must be developed to take advantage of intermediate temperature

electrolytes. One such cathode is LSCF(Lal-m SrCol_,Fe,03 -

Like LSM, LSCF owes its electronic conducting properties to its perovskite structure.

When lanthanum cobalt oxide (LaCoO3) is doped with strontium oxide (SrO), the Sr

atoms substitute on a La site, creating a hole to compensate as shown.


SrO 'aacoo3 Sr1, h* OX (2-5)


In LSM, the B site ion, Mn, changes valency from 3+ to 4+ to accommodate the charge

difference created on the A site. The 4+ oxidation state is unlikely for both Co and Fe

(2+ and 3+ are favorable); therefore, most of the negative charge created by substitution

of the dopant atoms is compensated by vacancy formation.

Lacoo3 1 1
SrO 3 Sr1, + 02+Vo" 2-6
2 2

These two reactions are related by the electroneutrality condition in LSCF, described in

Equation 2-7.
n- + Sr]= [o] (2-7)


Valency changes among the Fe and Co ions account for n = [IML and p =['] hl

[Sris] is equal to the Sr dopant concentration. The excess holes contribute to electronic

conductivity allowing the LSCF to achieve an electronic conductivity similar to that

of LSM, between 200 and 300 S cm-l at 900 oC. The maximum conductivity of LSCF;

however is reached at significantly lower temperatures depending on the concentration

of the cations [35, 36]. The ionic conductivity of LSCF, on the other hand, is several










orders of magnitude higher than that of LSM at the same temperature, (0.2 versus 10-7

S cm- ) [37-40]. The relatively large ionic conductivity of LSCF is due to the additional

vacancies which increase the oxygen ion mobility. The oxygen ion diffusion coefficient

is significantly larger in LSCF as compared to LSM (about 10-7 at 800 oC versus 10-12

cm2/s at 900 oC) [41]. At pl)2S leSs than around 0.01 atm, the oxygen ion diffusion

coefficient begins to decrease as a function of oxygen partial pressure (possibly due to

defect association) at 800 oC [42]. As a result, the ionic conductivity has a maximum at

around 0.01 atmospheres despite the increase in oxygen vacancies at low p)2S [43 .

2.3 The Cathodic Reaction

The mechanism of oxygen reduction at the cathode/electrolyte interface is quite

complex and has been the study of a multitude of works [22, 25, 34, 44-47]. The overall

cathodic reaction is made up of various individual steps. The process begins as oxygen

flows through the atmosphere to the interconnects. Next, oxygen diffuses past the

interconnects to the surface of the porous cathode. From this point on, depending on

the system, including materials, microstructure, atmosphere, and other experimental

conditions, multiple pathi--li- are possible. Before an oxygen molecule can he transformed

into an ion in the electrolyte, oxygen gas must somehow pass through the cathode. This

can he accomplished if individual gas molecules diffuse through voids in the porous

cathode to the triple phase boundary (TPB), the location where the gas phase, cathode,

and electrolyte converge. It should be pointed out that depending on the openness of the

cathode (pore size, porosity and tortuosity) a convective flow in which gas molecules flow

as groups, Fickian diffusion in which molecules diffuse by randomly bouncing up against

one another, or K~nudsen diffusion in which a wall of the cathode is most likely to cause a

change in direction of propagation can occur. So, at least three distinct possibilities exist

for the way in which gas can diffuse through a porous cathode to the TPB, the area where

the porous cathode, gas, and electrolyte meet.










The situation is further complicated by considering that adsorption may occur at

some point on the cathode surface not in the immediate vicinity of the TPB. In this

scenario, the diffusion of the gas molecule through the open pore must he followed by

surface diffusion of the adsorbed species toward the TPB. If surface diffusion is rather slow

compared to molecular adsorption, then only oxygen molecules adsorbed in the vicinity of

the TPB will become ions in the electrolyte. If surface diffusion is extremely fast, however,

molecules adsorbed over the entire surface of the cathode will participate in the cathodic

reaction. For intermediate surface diffusivities, surface diffusion itself may in fact limit the

rate of the cathodic reaction.

Once the adsorbed oxygen species has reached the TPB, a charge transfer reaction

can occur in which the adsorbed species on the cathode are converted into charge carrying

ions in the electrolyte. In order for this reaction to occur, electrons (or holes) which are

transferred through the electronically conducting cathode and oxygen vacancies (or ions)

which have traveled through the electrolyte must reach the reaction site, i.e. the TPB.

Various works have attempted to summarize the possibilities (see Figure 2-1) [48, 49].

The situation is further complicated by considering that various possibilities exist for the

structure and charge of the adsorbed species.

This discussion has thus far considered only purely electronic conducting cathodes. In

mixed ionic electronic conducting cathodes (\!IE;Cs) a bulk pathway exists in addition

to the possible pathi-- .1-< mentioned above. The oxygen molecule can adsorb and

participate in a charge transfer reaction at any point on the surface of the cathode.

The formed ion or vacancy will then diffuse through the bulk of the cathode to the

AllEC/electrolyte interface. At this point, the adsorbed ion/vacancy can he incorporated

directly into the electrolyte depending on any hIlEC/electrolyte interfacial resistance.

If the electronic conductivity of the AllEC is much greater than the ionic conductivity

(as it is in LSCF) and plenty of molecular oxygen is available at the TPB, reaction

pathi-- .1-4 involving transport of ions through the AllEC may be of higher resistance than




























Lo rg var .. >oa

O- 00>-I,,:



,-- TPB O c-

O ,, + V2 + 2e' 00
i~O, iMI + V" + 2e'~ O;


Surfacer Layer


ZIRCO)NIA v";
OXYG EN COND)UCTIOR
(YSZ)

Figure 2-1. Possible reaction pathi- ai- for a platinum cathode on electrolyte system,
illustrated by Nowotny et al. in reference [49]. Permission to reproduce was
obtained from Maney Publishing, publisher of the original figure.


H


OI ,,, + V. + 28 --eO

a" ~0'2,. + V~ + e -----------, O


r"
ri
r
r
r


ZLr Pt Y O GCompound


GAS PHASE





O o






suma


O,, + el + V

O01p e ~ V










the competing pathway localized at the TPB. As the reaction proceeds and molecular

oxygen becomes depleted in the vicinity of the TPB, ionic conduction through the AllEC

become increasingly significant despite the lower ionic conductivity in the AllEC. Further

complicating the system, the individual pathi-- .1-< must he considered to operate in

parallel; depending on the relative resistance of each path, multiple paths-- .1-< may operate

simultaneously.

The mechanism reported depends on the type of conductivity observed (electronic

versus mixed), the density of the electrode, and other materials parameters. Although

LSM is generally accepted as an electronic conductor, as recently as 2003 conflicting

reports were published concerning whether ionic conductivity plI li-< a significant role in

conduction in LSM. Fleigf reports that for dense patterned LSM microelectrodes transport

of oxide ions in the cathode is the rate determining step [50]. As proposed by S. Adler,

even a very small ionic conductivity such as that in LSM may create an active reaction

l o-;-r near the cathode/electrolyte interface in which ionic behavior is significant [51].

The thickness of the cathode examined may significantly influence whether the cathode is

perceived as a purely electronic conductor or not.

An efficient cathode should be a porous electronic conductor or MIEC. The

primary method of studying reaction mechanism is to examine the effects of changes

in polarization, oxygen partial pressure, and temperature on the occurring electrochemical

processes. In another work by Adler, the author concludes that for LSCF at 700 op

oxygen reduction at the gas/ \!1100 interface and solid state diffusion in the AllEC are

the 1 in r contributing processes [52]. A couple of works propose an oxygen reduction

mechanism that consists of three rate limiting steps for an LSM cathode. The high

frequency step is attributed to charge transfer of oxygen ions from the cathode/electrolyte

interface to oxygen ion vacancies in the electrolyte. The intermediate step is attributed

to the dissociation of adsorbed oxygen molecules into adsorbed oxygen atoms. The low

frequency step is attributed to the diffusion of oxide ions to the interface [46, 53]. In










another study, S. P. Jiang examined the polarization mechanism of the cathodic reaction

on both LSM and LSCF cathodes [54]. The author observed that for LSM cathodes,

surface dissociative adsorption and diffusion, charge transfer, and oxygen ion migration

into the electrolyte were the significant reaction steps with dissociative adsorption being

the rate limiting step at low temperatures and oxygen ion migration limiting at high

temperatures. Further, the author found that for LaSrl-mCoO3, dissociative adsorption

and bulk diffusion (or surface diffusion) processes are significant in cathodic reaction with

LSCF cathodes. Increased performance with LSCF is also observed and is attributed to its

higher oxygen ion conductivity and catalytic activity.

For purely electronic conducting electrodes, the electrochemical reaction driving fuel

cell operation is restricted to the TPB [55]. Because much of the power loss in SOFCs is

due to polarization loss at the cathode-electrolyte interface, degradation of this interface

has a deleterious effect on the performance of the cell [55]. Increasing the TPB area,

on the other hand, results in a more efficient SOFC [56] and modern SOFC structures

are engineered to maximize this area. M. Ostergard was one of the first to propose and

develop composite LSM/YSZ cathodes with the specific intention of increasing the TPB

and therefore device performance [57]. Increasing the TPB has even been shown to

increase electrochemical performance for the MIEC LSCF [58]. A quality TPB requires

high porosity in the electrode and good adhesion between the electrode and electrolyte.

Breakdown of this interface is a primary cause of device deterioration. Delamination of

the cathode is one source of degradation of this interface degradation [59]. The reaction

between electrode and electrolyte is another source of interface degradation and is the

focus of one of our studies.

2.4 Impedance Spectroscopy

2.4.1 Measurement Details

Impedance at a given frequency, Z(w), is a complex number defined by Ohm's

law as the voltage, V(w), divided by the current, I(w). AC electrochemical impedance










spectroscopy (EIS) is a characterization technique that measures voltage (or current) as an

alternating current (or voltage) is applied to the test sample over a range of frequencies.

The output is the real and imaginary components of the impedance (or equivalently a

magnitude and phase angle) at each frequency measured. The data is typically di;111 li4I

in a Nyquist or Bode plot. The Nyquist plot has real impedance (Z,) and negative

imaginary impedance (-Zj) as the x and y coordinates, respectively. The sign convention

used in the imaginary axis is a result of the fact that capacitive behavior often dominates

the processes examined in the literature. With this choice of sign convention, the 1 in .0 r~y

of phenomena of interest for SOFC applications will lie in the first quadrant. The Bode

plot puts the impedance magnitude and phase angle on the y-axis and the log of frequency

in the x-axis. Trouble can arise if data is fitted only to a Nyquist plot in that any two

RC elements with identical resistance values, but different capacitance values will appear

identical on the plot. Simultaneous use of both the Nyquist and Bode representation of

the data eliminates this ambiguity. The frequency range examined is usually in the range

of 0.001 1 x 107 Hz. The upper limit on test frequency is constrained by limitations

due to the measurement device, while the lower frequency is usually limited by the time

it takes for data acquisition. These limitations are often not a concern because many of

the phenomena studied have a characteristic frequency lying in the measurable frequency

range.

The value of impedance spectroscopy as a characterization technique is that it

produces evidence of the total polarization loss at each frequency measured. For a given

polarization process, loss only occurs if the perturbation occurs at a lower frequency than

that processes relaxation frequency. If the perturbation occurs at a higher frequency

than the relaxation frequency, the system won't have time to dissipate any power via

that mechanism. Conveniently, we can look at the entire impedance spectrum and see

the individual processes contributing and their respective significant frequency ranges.










Coupling this information with other experimental data and theory allows us to identify

the significant individual processes.

Each of the individual processes occurring has its own real and imaginary impedances,

and characteristic frequency associated with it. The capacitive impedance, Zo, the

inductive impedance, ZL, and the ohmic resistance, ZR, as a function of frequency are

given by the following relations, respectively.


Zc = 1/(jwC) (2-8)


Z, = jwL (2-9)

Z = R (2-10)

A single electrochemical process will trace a semicircle in the Nyquist plot with the

semicircle's diameter lying on the positive x-axis. This behavior can be modeled by a

series resistance connected to a resistor and capacitor in parallel (Voigt element). The

distance from the origin to the beginning of the semicircle will have a magnitude of Rs

(the resistance of the series resistor). This contribution to impedance is due to either

ohmic resistances or any process that occurs at a frequency range much higher than

the measured range. The diameter of the semicircle will have a magnitude equal to the

parallel resistance (R,muet,). The peak of the semicircle will occur at the characteristic

frequency, wo. The magnitude of the height of the circle is equal to Zo, which is related

to the parallel capacitance (Cp) by the equation above. These constants are additionally

related via the R-C time constant, -r.


-r = Rp x Op. (2-11)


The time constant and the characteristic frequency (w) are inversely related.


-r = 2xr/w. (2-12)


The time constant provides information on the kinetics of the reaction.










2.4.2 Data Analysis

Each of the processes occurring has its own resistance and capacitance and therefore,

a distinct characteristic frequency associated with it. The manifestation of each discrete

process is a semicircle in the complex plane. As mentioned previously, any processes with

a characteristic frequency much higher than the measurement range will behave like an

ohmic resistance. An actual impedance measurement usually reveals a superposition of

individual semicircles, which are caused by the individual polarization processes. If two of

the processes occurring have characteristic frequencies in close proximity, their individual

semicircles will overlap. In order to extract information about the individual processes

occurring, we must be able to separate the contributions of the various phenomena acting

from one another. Ideally, all of the processes occurring can be identified in terms of

their individual resistances, capacitances, characteristic frequencies, and time constants.

This process is, however, non-trivial and has received considerable attention and been the

subject of a multitude of works.

Impedance data is often analyzed by fitting the spectra to a model, which is based

on a priori knowledge of the system. This type of model is represented by an equivalent

circuit [60]. For SOFC applications, authors [11, 12] have proposed a model based on

the brick 1... -r model [61] which separates the contributions of the electrolyte bulk

(intragranular), electrolyte grain boundary (intergranular), and electrode effects (charge

transfer). Gas diffusion and ion migration are included among other phenomena that may

contribute to the unresolved spectra. Jamnik and Maier derived a general circuit for a

cell with a MIEC [62]. When applied to the special case of a SOFC, their model results

in an equivalent circuit nearly identical to one used by several authors which is composed

of a double 1... -r capacitance in parallel with a series connection of a resistor and a Voigt

element [63]. This equivalent circuit is the most commonly used of the nested circuits.

An inherent limitation with this method is that the fit attained is dependent on

knowledge of the processes contributing to the spectrum. Two important consequences










result from this fact. The first is that black hox type study is not feasible, i.e. a sample of

unknown composure cannot he analyzed in this manner. The second is that the degree of

certainty in model parameters attainable is limited by our confidence in the assumptions

made based on a priori information. In short, even if two authors agree that a model

must he developed from theory, they may derive two different models and therefore

yield incomparable data. Unfortunately, not even a perfect fit of the data is proof of the

correctness of the chosen model because multiple equivalent circuits differing in structure

can produce the same impedance curve [64].

A second commonly used technique minimizes ambiguity at the expense of confidence

in the model. This technique matches all semicircles present in the Nyquist plot with

individual Voigt elements (resistor and capacitor placed in parallel) connected in series.

The inl r ~ advantage of this type of analysis is that comparison of data among different

research groups with different assumptions about the mechanism of reaction is facilitated.

The primary flaw of this technique is that assignment of parameters based on the model is

difficult since the model was not constructed with specific parameters in mind. In order to

assign parameters to the model, the EIS measurement must he repeated as measurement

conditions and/or sample characteristics are varied. The resulting impedance spectra are

then modeled and changes in the attained model parameters are then compared with

theory, leading to assignment of identities of the individual parameters.

Another method sometimes called system identification [65] is based on the reasonable

assumption that the input/output behavior is dependent only on the cell to be tested [66].

In system identification, instead of modeling from physical laws, a mathematical model

is built based solely on the experimental data. System identification consists of three

steps: pre-identification, model estimation, and model validation. In pre-identification the

data is manipulated into a format suiting the chosen model. 1\odel estimation involves

determination of the various parameters of the model and model validation tests the

suitability of the model [65]. The output impedance spectrum is analyzed while input,









such as experimental conditions, is varied. A model is produced, the parameters of which

are subsequently related to understood physical processes. This type of an~ llh--; may

utilize mathematical techniques to increase frequency resolution. The strength of this type

of analysis is that black box type investigation is possible.

Mathematical techniques have been developed to aid in data analysis, and in

particular to increase the resolution of the measured data [67]. For very simple systems

with clear separation of electro-chemical processes, model parameters may be obtained

without great difficulty. For as few as two conjoined semicircles, mathematical tools have

been developed which aid in attainment of model parameters [68]. Several methods of data

analysis based on Fourier analysis of the raw data have been presented [67, 69]. These

methods increase the frequency range of the data, facilitating separation of the individual

acting processes that show similar relaxation frequencies. In this manner, processes can

be resolved which are not resolvable using conventional methods [70]. The advantages

of using this type of technique include the following: 1) time constants may be obtained

with little knowledge about the system, 2) separation of distributions that are not readily

separable in conventional impedance data is possible, 3) a reduced sensitivity to random

experimental error is gained [69]. The K~ramers-K~ronig relations are also used to reduce

error associated with the data analysis. According to the K~ramers-K~ronig relations, the

same information is contained in the real and imaginary components of the impedance

profile, therefore, in data analysis, only one of the components needs to be considered [71].

The K~ramers-K~ronig relations have also been used to identify and/or reduce error in the

acquired data [72]. This type of analysis is the focus of one chapter in this work.









CHAPTER :3
ERROR ANALYSIS

3.1 Introduction

As mentioned previously, it is well understood that cathodic polarization provides the

largest contribution to losses suffered by the traditional SOFC system under operating

conditions [7:3]. Unfortunately, considerable disagreement concerning the number and

identity of the elementary electrochentical processes steps occurring in the cathodic

reaction still exists [74]. This disagreement is due to three main factors: the method of

reduction is dependent on the materials systent, variations in sample preparation and

nicasurenient conditions affect the output, and subjectivity in analysis of impedance data

leads to inconsistent.

Cathodic reduction has been studied on several materials systems, each producing

its own results. Some of the earliest work in the area has been on platinum electrode

SOFC systems [49]. Because platinum is a purely electronic conductor, the corresponding

cathodic reaction is strongly dependent on the surface properties of the platinum. 1\ore

recently, mixed ionic electronic conductors (illlECs) have been considered as cathodes. For

these systems, the cathodic reduction reaction would involve some ionic transport through

the bulk of the electrode [54]. In short, the reduction reaction pathway depends on the

materials system chosen.

The next factor leading to confusion is variation in sample preparation and

nicasurenient conditions. Even for a given materials system and a single laboratory it is

impossible to produce identical samples for testing from batch to batch. Slight variations

in cathode thickness, cathode sintering conditions, and solid electrolyte properties may

be unavoidable. Obviously, when comparing results front laboratory to laboratory,

these variations increase. The bias history of the sample may also have some effect

on its measured properties [75]. The measurement conditions also include the type of

nicasurenient done, i.e. 2, :3, and 4 point probe impedance measurement, each producing a

different output for a single materials system.










Third, the process of impedance spectra deconvolution is shaky at best. It has

been shown that multiple RC-based models can produce a given impedance spectrum

[64]. Additionally, some significant processes may be enveloped by other processes and

depending on the chosen method of displaying the data and sensitivity of the equipment

these hidden process may be overlooked. The evaluation is further complicated by artifacts

introduced in the measurement by the measurement system. These artifacts are typically

accounted for by nullingng" (explained in greater detail below) or simply truncating the

data to limit the imaginary impedance to negative values. Little understanding exists

concerning the validity of the data points immediately after the spectrum crosses the real

impedance axis.

The goal of this chapter is to analyze the quality of impedance data, particularly

at high frequencies. This is accomplished by determining how the data deviates from a

K~ramers-K~ronig consistent model. Ultimately, a method is proposed which improves the

consistency of the high-frequency data. This method was used in chapters 6 and 7 of this

dissertation.

3.2 Experimental

Symmetrical cathode/electrolyte/cathode test samples were produced for the

work. The cathode used was LSM ((Lao~sSTO.2~l *M -nO3-6) ink supplied by Nextech

Materials, Ltd. and YSZ was used as the electrolyte in this work, prepared by tapecasting

(11 I l:a tech International, Inc.) The YSZ contained 8 mol .~ yttria and had dimensions

of 10.0 x 20.0 x 0.1 mm. The cathode was screen printed on both sides of the electrolyte

in two lIn-;-rs with a square area of 64 mm2, TOSulting in the symmetric sample shown

in Figure 3-1. A drying step was performed after the screen-printing of each Ins-,-r in a

Fisher Isotemp drying oven at 120 oC for one hour. After drying, sintering at 1100 oC and

1 hour was performed in a Lindberg/Blue high temperature box furnace. The resulting

symmetrical samples had a cathode thickness of about 20 microns.
































The samples were mounted in a quartz tube inside a Barnstead/Thermolyyne furnace

to control measurement temperature. The quartz tube consisted of an inlet and outlet

for gas flow, gold leads running through alumina rods coated with platinum for shielding,

and a pressure contact sample holder. The gold leads were connected by platinum wires

to a platinum mesh, which was used as the current collector. The pressure contact holder

was designed in a way that exposes the platinum mesh and .Il11 Il-ent cathode to the

ambient gas. Air was flowed over the samples at 40 seem. A Solartron 1260 impedance

gain analyzer was used to measure the frequency response of the prepared samples.

Electrochemical impedance spectroscopy (EIS) using a Solartron 1260 impedance

gain analyzer was performed in order to measure the frequency response of the prepared

samples. A 50 mV AC voltage was applied and the induced current was measured

to produce the impedance spectra. 1\easurement was made via a 2-point connection

to the Solartron. Auto-integration was used under "\I, long" measurement conditions

with an integration time of 60 seconds. I, long is a Zplot option in which the current is

measured for noise and an attempt is made to get consistency in the measurements with

a maximum standard deviation of 1 when possible. The active frequency range was


Figure 3-1. Photograph of a typical sample.










20 103




5 10 5 2
-80I 10 a
-10 0o- 0EH 100

-120 1


0 20 40 60 80 100 120 140 101 100 101 102 103 104 105 106 107
Z, 0 Frequency, Hz

Figure 3-2. Raw impedance of lanthanum strontium manganite (LSM) measured at 900 oC
in air. a) Complex plane plot. b) Imaginary impedance vs. frequency plot.


1.0 x 10-2 3.2 x 107 Hz. SMART, Zplot and Zview were used to acquire and di play~ the

impedance data.

3.3 Results and Discussion

3.3.1 High-frequency Artifacts in Impedance Data

1100 oC sintered LSM on YSZ was tested by AC-impedance spectroscopy under a

typical operating condition of 900 oC in air. The measurement was repeated six times

to give six replicates of the data. Complex plane and Zjversus frequency log-log plots

of the first replicate are shown in Figure 3-2. The traditional complex plane plot shows

a single large arc under these testing conditions. The Zjversus frequency plot shown in

Figure 3-2(b) is broken into two regions. In the figure, -Z is plotted for the capacitive

portion of the data (lower frequencies) and +Z is plotted for the inductive region (higher

frequencies). Di playing the data in this format directly indicates the peak frequency

and therefore the time constant of the capacitive arc shown in Figure 3-2(a). The slope

of Figure 3-2(b) in the low-frequency regime is constant and close to 1, indicating that

only one process is dominant in this region and that a constant phase element may not

be necessary for modeling in this region. As frequency increases, a high-frequency artifact

becomes more and more significant. The slope change in the higher frequencies of the










capacitive arc (103 to 104 Hz) may be indicative of multiple cathodic processes, however,

the contribution of the inductive artifact makes a determination difficult. We can see that

around 104 Hz the high-frequency artifact competes with the capacitive data leading to

uncertainty in the validity of the data. This impedance profile and its 5 replicates were

analyzed using the measurement model technique developed by Orazem [76].

The measurement model uses several RC Voigft elements connected in series to

produce a K~ramers-K~ronig (KKE) consistent model for the data. The KKE relations are

valid for systems that satisfy conditions of causality, linearity, and stability. The KKE

relations assert that if these conditions are valid, the real and imaginary components of

impedance data contain identical information, and in fact, the real part of the data can

be generated from the imaginary part, and vice-versa. The KKE relations are expressed in

Equations 3-1 and 3-2 for a function f(x) = u(x) + iv(x).



u~o)= Pdx (3-1)



Fletcher applies the KKE relations to an RC circuit in [77]. An inductive artifact is an

effect of the imperfect measurement technique and is not specific to the properties of

the sample to be tested, i.e. it is non-causal. If the magnitude of this artifact becomes

significant then the imaginary part of the data will no longer be linked to the real part

and the data is no longer KKE consistent. For this reason, a series of Voigft elements will

not produce the arc shown in Figure 3-2(a).

When dealing with impedance data, the presence of a high-frequency artifact in the

raw data is not unusual. There are three common v-wsi~ to deal with this phenomena.

The first is ignoring the high-frequency portion of the data, understanding that it is not

useful. The second is truncating the data at the intersection with the real axis. The third

is the use of "nullingt a technique in which an impedance run is made under open and

short circuit conditions and the results are subtracted from the raw impedance data. If









the first option is chosen, we must abandon any hope of recovering useful high-frequency

data. Truncating the data, which is typically done at the Z, axis, gives the impression

that the data kept are valid. Below, we show that truncation of the data at the Z, axis

leaves corrupted data in the data set. Application of nulling files often over corrects for

the inductive behavior resulting in a false semicircle in the high-frequency portion of the

impedance profile.

The measurement model concept [78-80] was developed for the identification of the

error-structure of frequeno ~l-i-domain measurements. Orazem et al. extended the model to

generalized identification of distributions of time constants and ultimately, a systematic

approach was developed for analysis of error structure in impedance data [72, 76]. One

of the primary reasons for the development of the model was in the identification of the

frequency range that was unaffected by instrumental artifacts of non-stationary behavior.

In this work, the measurement model is used to determine the extent of the corruption of

the raw data by the commonly observed high-frequency artifact di;11l-plw I in Figure 3-2(b).

Because the measurement model technique relies on a complex nonlinear least-squares

(CNLS) technique for the regression of impedance profiles, (the regression is based on

both the real and imaginary parts of the data, simultaneously) the solutions attained

must be consistent with the K~ramers-K~ronig relations. In short, the actual data is fit to

the following relationship and the parameters K(, Ro Rk,, and -rr are returned along with

corresponding standard deviations.

Rk,
Z = Ro +(3-3)
z%+C1 + jean

Where Ro describes the ohmic resistance and K( indicates the number of Voigft elements

in the model while Rk, and -rr are the polarization resistance and time constant of the kth

Voigt element, respectively. If the data is totally inconsistent with the KKE relations, the

model will fail to converge to a solution and some data points may have to be removed.

Because of the non-causal artifact shown in Figure 3-2(b), the high-frequency portion of










20 ,, 4




161






6 -1
101 100 101 102 103 104 105 101 100 101 102 103 104 105
Frequency, Hz Frequency, Hz

Figure :3-:3. 1\odel fit, including the 95' confidence intervals, of the raw data for an
1100 o"C sintered sample. a) Real part. (b) Imaginary part.


the data was not KK1 consistent, the measurement model would not converge to a solution

if the entire data set was used. To alleviate this problem, we applied the commonly

used technique of truncating the data at the Z, axis and an attempt was made to fit the

remaining data using the series Voigft element model. A convergent solution; however, was

not reached indicating that the data points that were kept still had significant contribution

from the artifact. After removing more high-frequency data points, a solution was reached,

but the model showed significant deviation from the actual data at the highest frequencies

kept, leading to increased error in the model. After more high-frequency data points were

removed, the high-frequency deviation between model and data disappeared and a quality

fit was attained. The solution had six elements with the constants given in Table :3-1.

Figure :3-3 shows Z, and Zj vs frequency plots for the fitting. The figure shows the data

with the used points as solid circles located between the vertical hars, truncated points as

hollow circles outside the vertical hars, the model as a solid line and the 95 .confidence

intervals as dashed lines. The .l-i not... r-y shown in Figure :3-:3(b) is due to the inductive

artifact di1 i-. II in Figure :3-2(b). The inset figure in Figure :3-2 shows the extent to

which data had to be removed to produce a quality, KK1 consistent solution. Data near the

Z, axis is therefore unreliable.










Tabe 31. an 7 alus wth hei repective errors (standard deviation, a) for raw data
measured at 1100 oC.

Process # R (R) an (R) -r (ms) a, (ms)
Element 1 0.883 0.040 0.064 0.002
Element 2 2.02 0.19 0.24 0.01
Element 3 5.15 0.11 0.59 0.02
Element 4 1.01 0.15 1.52 0.11
Element 5 0.111 0.012 8.7 1.1
Element 6 0.040 0.004 70 11


Constants such as those shown in Table 3-1 were obtained for each of the 6 repeats of

the impedance spectroscopy measurement. The constants obtained were compared and the

real and imaginary standard deviation (or,4) in the model was calculated and is di;11l-~ pts

in Figure 3-4(a). Figure 3-4(a) shows that a,,j has a magnitude on the order of 10-3 R

Additionally, a consistent trend in both the real and the imaginary standard deviation

values is observed, the standard deviations have their largest magnitude values in the

low-frequency regime and drop off as frequency is increased. The fact that the error has a

frequency dependence but still some randomness indicates that there are both stochastic

and non-stochastic errors present in the data. The frequency dependent non-stochastic

errors are either related to systematic experimental errors or due to imperfections in the

model.

As mentioned previously, cutting off the high-frequency data at the Z, axis is a

common way of dealing with high-frequency artifacts. We used the Measurement Model

to show that simply choosing Z, axis for a cutoff point is insufficient for avoiding the

problems caused by the common high-frequency artifact. To effectively avoid contributions

from the high-frequency artifact, data must be cut off well above the Z, axis to ensure

reliable data.

3.3.2 Correction of High-frequency Artifacts in Impedance Data

As an alternative to eliminating high-frequency data points which may contain

information (overshadowed by an artifact) the following technique is proposed. The















rawrY



on a --E- Iffagliary error
++2 +++ o a *
on ~~ + *
o+ + ++.
900 0 +++ o +
+ on
real error o,


"1 0


e- 10 "



10-4


10100 110


102 103 104 110


Frequency, Hz


10






. 3
10







10-4


cocrre cte d


++ On + + ++ + ++ +
*, oo* co o + ++
O QOOO


~inary erro~r


...
OO


+

QO


real error~ ..


0. .


I I I + 1
10 1 104 105


10 00 10


Frequency, Hz

Figure 3-4. Standard deviation (o ,,) versus frequency determined from six replicates of
data. There is no significant increase in the error of the high-frequency data,
despite its being increased an order of magnitude versus the raw data. a) Raw
and b) High-frequency corrected data.











4 Raw

-2 -316 Hz a


a2 "

0 Raw a 4
4C I ^Corre cted l ooI


6 8 10 12 14 16 18 101 10 70 0 0
Z 9 Frequency, Hz

Figure 3-5. Raw and high-frequency corrected data for LSM on yttria stabilized zirconia
(YSZ) measured at 900 oC in air. a) Nyquist plot. b) Imaginary impedance
versus frequency plot.


highest frequency portion of the imaginary impedance data is first fit to the relationship

S= j2x ~f Lexp where f is the frequency of the measurement, and Lexp and P are

constants to be determined from the fittingf. If P is equal to one, the expression reduces

to Zj= jeoLexp, an eXpression for the impedance of an ideal inductor. If p is not equal

to one, then the high-frequency artifact is not a pure inductance. This fit is valid in the

high-frequency regime because as frequency gets large, the capacitive impedance, which

is characteristic of the sample, approaches zero and the modeled artifact impedance

approaches a large value. The high-frequency portion (3 x 10s to 3 x 106 Hz) of the raw

data from Figure 3-2(b) was fitted to Zj= 2x f Lexp and it was found that P = 1.037

and that Lexp = 1.46 x 10-6. The frequency range is chosen to minimize the influence

of capacitive behavior at the lower frequencies and avoid the errors in the data that are

present at the highest frequencies. This fitting is subtracted from the raw data with the

results shown in Figure 3-5. The high-frequency tail visible in Figure 3-2(a) is diminished

and the symmetry of the imaginary impedance data di 1 l-00 +4 in Figure 3-5(b) is increased

dramatically. An analysis of the K~ramers-K~ronig (KKE) consistency of the data provides an

objective method of assessing an improvement in the data.










201 1 I I I I 4






12 I -




6 """-1
101 100 101 102 103 104 105 106 101 100 101 102 103 104 105 106
Frequency, Hz Frequency, Hz

Real fit generated from imaginary impedance data and KKE relations for LSM
on YSZ measured at 900 oC in air. a) Raw data. b) High-frequency corrected
data.


N









Figure 3-6.


Figure 3-6 shows the fit produced by the Measurement Model Toolbox in an

impedance versus frequency format. The parameters for the fit are di;111lai-rd in Table

3-2. Despite the high-frequency data manipulation, there is still some unusable data in

the high-frequency regime (approaching 106 Hz) as shown in Figure 3-5(b). To get the

measurement model to converge, we had to throw out some of the highest-frequency data,

however, as shown in Figure 3-6 the usable data extends from 0.63 Hz to 50 kHz while the

usable raw data extended to only 3.9 kHz as shown in Figure 3-3. By 10s Hz in the raw

data, the high-frequency inductance has caused the imaginary resistance value to increase

to a positive value of 2 R, making the data completely unusable. This frequency range

may be crucial when trying to deconvolute impedance spectra as some cathodic processes

are expected to have their peak frequency in this regime [81]. As with the raw data, six

replicates were used to determine the standard deviation as a function of frequency for the

high-frequency manipulated data models. Figure 3-4(b) shows the standard deviation as a

function of frequency for the high-frequency manipulated data. It should be noted that the

error shows the same general trend as the error from the raw data and is of the same order

of magnitude. The usable data was extended an order of magnitude without compromising

the quality of the data.









Tabe 32. an 7 alus wth hei repective errors (aR,,) for high-frequency corrected
data measured at 900 oC.

Process # R (R) an (R) -r (ms) a, (ms)
Element 1 0.402 0.038 0.023 0.0016
Element 2 1.120 0.081 0.100 0.0083
Element 3 3.97 0.42 0.378 0.027
Element 4 3.54 0.45 0.831 0.058
Element 5 0.303 0.062 3.41 0.53
Element 6 0.065 0.008 33.85 6.47


Figure 3-7 shows the real variance divided by the imaginary variance as a function of

frequency for both the raw and high-frequency corrected data. For the vast 1!! li G~ry of

the datan, (T/afT is desirably between 10-3 and 10-1. Below 103 H~z, there is no significant,

di;f~ference in between, the two,,, cases Forboh dlata sets the ratio ,a/,a decreases as a

function of frequency to some minimum value then increases at the highest frequencies.

The reason for the dec~rease of ,a/,a as a function of frequency is that the real part of

the data and thus a has its largest magnitude at low frequencies, while the imaginary

part of the data has its largest, magnitude at about 103 H~z. The increase in (T,/(T can

be explained in that the imaginary impedance approaches zero at the highest frequencies

used, therefore, aj is small compared to a, at the highest frequencies.

In the previous analysis, we used Voigt elements connected in series to fit the data

and analyze the error structure. The Voigt elements used were each composed of a resistor

connected in parallel to a capacitor. For some systems, a fit composed of resistors and

constant phase elements (CPEs) in parallel is more useful. The constant phase element is

a mathematical tool used to model a distributed time constant. A resistor in parallel with

a CPE produces a depressed semicircle in the complex resistance plane described by the

following equation.

Z =(3-4)
1+ RQ(jw~a
Z is the complex impedance, R is the parallel resistor, and Q and a~ are the parameters

of the constant phase element describing the frequency dependence and the depression










lU I
o raw data
100 + high-freq. corrected

10






10-4
10 100 10 10 10 10 10

Frequency, Hz

Figure 3-7. Real Variance / Imaginary Variance (@/af) for raw and high-frequency
corrected data.

of the are, respectively. For a single CPE based Voigt element, the slope in an Zy versus

frequency log-log plot at low and high frequencies tends toward +a and -a, respectively. If
a = 1, the constant phase element becomes an ideal capacitor. By examining the high and

low-frequency slope of Figure 3-2(b), we can determine if there is constant phase element

behavior. In the linear portion of the low-frequency range of Figure 3-2(b), a slope of

0.878 was calculated. This value indicates that we have CPE behavior. Because multiple

capacitor based Voigt elements are required to model a single CPE based Voigt element,
we can not infer that there are six independent electrochemical processes corresponding

to the six capacitor based Voigt elements used in the model. It should be pointed out

however, that the value of the regression an .k is in determination of the quality and

validity of the data and the focus of this work is not in identifying the number of actual

physical processes contributing to the cathodic reaction.

Using 0.878 for a, the effective capacitance (Qeff) can be calculated as a function of

frequency as follows in Equation 3-5 [82].

Qey; = sin( ) (3-5)
2 Zy(f)(2xf)a










10 .

qo *z u =0.878



Om 10- *




10

'101 100 10' 10 103 104 10s
Frequency, Hz

Figure :3-8. Effective capacitance calculated for LSM sintered at 1100 op.


In the equation, Zy ( f) is the imaginary impedance as a function of frequency. Figure :3-8

shows the effective capacitance as a function of frequency calculated for LSM sintered

at 1100 oC. The effective capacitance for the LSM sample stabilizes at around 1000 Hz

indicating a double 1 u-;-r capacitance of around 100 pF.

To further illustrate the significance of data correction, we have evaluated actual

parameters for the electrochemical processes occurring in the cathodic reaction of LSM

on YSZ in air at high temperatures. We used a series resistance and two R-CPE Voigt

elements connected in series to model the individual processes occurring in the cathodic

reaction. For each of the two Voigt elements, the returned parameters were R, Q. and c0.

These parameters can he used to calculate the time constant for each of the R-CPE Voigft

elements according to Equation :36.


7 = (R x Q)1/" (:36)


Our previous work [8:3] has shown that this model is acceptable and that the high-frequency

process can he attributed to charge transfer while the intermediate frequency process is

related to adsorption of molecular oxygen. Because of the high-frequency artifact, the

initial fittingf for the charge transfer process in the raw data returned a Q value greater









Table 3-3. C'!s lIty, in polarization resistance, Rp, constant phase element coefficients Q
and a~, and time constant, -r from deconvolution of raw and corrected data for
adsorption (1) and charge transfer (2) processes.

Rpl Qi at] 7 RP2 2 0 7
units (n) (mF) (ms) (R) (mF) (ms)
Raw 8.58 0.120 0.913 0.531 0.779 0.0908 1 0.0707
Corrected 8.59 0.120 0.912 0.531 0.849 0.0677 1 0.0575
.change a 0 O mO 0 m 8.24 34.0 n/a 23.0


than one, so we fixed this value at one, effectively modeling the charge transfer process

with an RC Voigt element. Table 3-3 gives the parameters returned from the model

and Figure 3-9 di pk.l--s the raw and corrected data along with corresponding individual

processes from the model. Both the table and figure show that the low-frequency, higher

impedance process (adsorption, labeled "1" in the table) is practically unchanged by

the performance of the high-frequency correction process, however, evaluation of the

high-frequency process (charge transfer, labeled "2") is significantly altered by the

correction. As seen in Table 3-3 the determined value for the charge transfer polarization

resistance and time constant change by 8.3 and 23 respectively. It is clear that the

high-frequency inductive feature can significantly distort determined electrochemical

parameters for high-frequency, low impedance magnitude processes. In other words,

inclusion of K~ramers-K~ronig inconsistent data in the evaluation can lead to significant

deviation of determined parameters from their actual values. Because the inductive

feature has large values at very high frequencies, performance of data correction is of

particular importance for those who have optimized their system to minimize impedance

and improve rate of reaction.

3.3.3 Repeatability of Measurements

A repeatability study was performed using samples sintered at 1100 oC for one hour

and measured at 800 oC. The first sample, sample 1, was measured three times on three

consecutive d -.-- Between measurements, the sample was cooled to room temperature,

removed from the experimental set-up, then put back in and re-measured. Two more




















g Raw Data 0 Corrcted Data



10 -Adsorption (raw eI
E, and corrected)



Charge
T ra nsfer:
-+ co~rrcted raw

102 1 I I
10' 102 103 104 105
Frequency, Hz

Figure 3-9. Modeled electrochemical process occurringf in LSM on YSZ measured at
900 oC in air.










samples were prepared and sintered at the same time as sample 1. The impedance data

for samples 1, 2, and :3 is di11 li-o I in Figure :3-10. The figure shows that for sample 1, the

-120]


-100 sample 1, run 1-
o sample 1, run 2
-80 sample 1, run 3
r sample 2
-0 v sample 3


-40-

-2 *f


0 20 40 60 80 100 120
Z', 0Z

Figure :3-10. Repetitions of an impedance measurement taken at 800 oC' in air.


total resistance increased from the first measurement to the third. It is unclear whether

the instability observed indicates that the process of impedance measurement alters

the sample, or whether the sample is unchanged and the resistance difference is due to

variations in sample orientation. The profiles of samples 2 and :3 also show variation in

total resistance. C'! Iny,. -4 of this magnitude are greater than any errors associated with the

modeling of the data or any correctable error associated with high-frequency artifacts.

The series ohmicc) resistance and total cathodic resistance of the measurements

shown in Figure :3-10 are di;1.11 i-. I in Figure :3-11. The y-axis used in Figure :3-11 is the

time constant of the peak, measured as the inverse of the angular frequency of the peak

of the profiles in Figure :3-10. As seen in Figure :3-11(a), the series resistance of sample 1

decreases from almost 6 to about 5.2 R from the first to the third run. Samples 2 and :3

had series resistances of 5.1 and 5.7 R, respectively giving a mean series resistance of 5.6 R

and a standard deviation of 0.45 R. This variation is either due to the microstructural










1 ..1
0.9 sample 1, run 3 c .
E Sample 1. run 3 *
S0.8 0.8
8 0.7 sample 1.runS 2 0.7 -sample 1. run 2
S0.6 0.6 -
8 ~sample 1, run 1 sample 1, run1
u 0.5 0.5 *
a, sample 2 a, sample 2
.E E*
p 0.4 -sample 3 0.4 sample 3
0.3 1 1 0.3
5 5.2 5.4 5.6 5.8 6 80 85 90 95 100 105 110 115
Seisrssac,0bTotal cathode resistance, O


Figure :3-11. Resistance values for repetitions of an impedance measurement taken at
800 oC in air. a) Series resistance. b) Cathodic total resistance.


variations in the samples (despite the fact that they come from the same batch and were

sintered together) or differences in the quality of the pressure contacts. Figure :3-11(b)

d1i ph the total cathodic resistance measured for each sample. This value increases from

84 to about 112 R from the first to the third measurement of sample 1. Samples 2 and

:3 had cathodic resistance values of 96 and 86 R, respectively. Samples 1, 2, and :3 had a

mean cathodic resistance of 88.5 R and a standard deviation of 6.8 R. Since -r = R C,

the slope of the data is equal to the capacitance which was calculated to be 14.3 pF. The

series resistance did not have a constant capacitance.

3.4 Conclusion

We were able to successfully apply a measurement model technique to EIS data from

an LS1\ on YSZ sample. A K~ramers-K~ronigf consistent model was used to fit the data and

analyze the error. The model used consisted of a series resistance and six Voigt elements

connected in series. As with most impedance data, a high-frequency artifact corrupted

the high frequency data, limiting the frequency range of useful data. The models attained

produced a good fit visually and the data values lied primarily within calculated 95

confidence intervals. It was found that the commonly used practice of truncating the

data at the Z, axis left KKE inconsistent data in the data set. To avoid inconsistent

data, several data points at frequencies lower than the Z,-axis had to be removed. The










consistency of the data was improved by fitting the high-frequency portion of the data

(which was shown to be inductive in nature) to Zj = jeoL and subtracting the result from

the raw data over the entire frequency range. Performing this operation increased the

amount of usable data in the high-frequency regime by an order of magnitude allowing

analysis of fast occurring electrochemical processes in our other work [83]. Additionally,

we have shown that the data exhibits CPE behavior and a model intended to describe the

physical mechanisms of the cathodic reaction should therefore be based on R-CPE type

Voigt elements. When modeling the data for the purpose of determining electrochemical

parameters, high-frequency, low impedance electrochemical process are particularly

vulnerable to the inductive artifact. A limiting double-l} m;r capacitance of around 100 pF

was found at high frequencies. A repeatability study was performed and it was found that

both series resistance and total cathodic resistance had standard deviations of around 8

which is larger that errors associated with the modeling which are generally less than 1









CHAPTER 4
ELECTROCHEMICAL PROCESS IDENTIFICATION

4.1 Introduction

The cathodic reduction reaction for LSM on YSZ fuel cells was discussed in detail in

section 2.3. As mentioned, the overall cathodic reaction is composed of many individual

steps. For an electronic conducting cathode, such as LSM, the general sequence of steps

defining the cathodic reaction is fairly well agreed on; however, the significance of each of

the individual steps is still discussed. The sequence begins with the arrival of 02 moleculeS

that have diffused through the porous cathode to the vicinity of the electrolyte. At

some point after this diffusion, the molecules are adsorbed (dissociatively or as complete

molecules) on the surface of the LSM. These adsorbed species now diffuse on the surface

of the electrode towards the triple phase boundary, the area where the cathode, electrolyte

and gas phase meet. At the same time, electrons (or holes) travel through the cathode,

while oxygen vacancies travel through the electrolyte. For a purely electronic conducting

cathode, the oxygen species are restricted from the bulk of the electrode, the oxygen

vacancies are limited to the electrolyte, and the electronic species (electrons and holes) are

limited to the cathode. For the complete cathodic reaction to occur, all of these species

must come together; therefore, the cathodic reaction is limited to the TPB.

Modeling of the cathodic reaction is often performed using an electrical circuit design

containing inductors, capacitors and resistors. A single electrochemical step with an

associated capacitance and a resistance can he modeled by a Hoigt element (a single

resistor and capacitor in parallel) [84, 85]. An assumption that the various electrochemical

steps of the cathodic reaction occur sequentially leads to a model consisting of a series

connection of Hoigt elements. This is a reasonable assumption for a purely electronic

conducting cathode in which only one reaction pathway is likely. Another possibility is

that one or more of the steps are mutually dependent. In this scenario, a simple series

connection of Hoigt elements may not he suitable.












600 oC0.21 atm



40 o 10 -
20 ,40 C 00o

0 .* 1 90 o
20 4 0 8 0 1 0 1 0 1
8 Z' 1 reunc1H
Fiue41 meaN pcr o 10o"sml tvrosmaueettmeaue
inar )Nqitpo.b mgnryipdnevru rqec lt

In, this seto mea petocp sue to00 exain the cotiu ioof h
of th siniicn elcrohmia prTse to the0 oveal ahdcrat on nsmmti
LS nYS.Asrie Vog elmn mdl(it osan hs eeet i l
of cpactor) i usd t exainethepO2depndeceactiatin eergesreltiv






Figre4-1 Ipednc secta or4.2 Exmperienat aliu esrmnttmea

The sample wereqi prepare andmgia impedance tetd nte ssam maner descri iso

I h section .2Intis hptdner spO2 was one of s thevribe examined Oxyen aotiuir, o and

aron gase s were used etoroduemce thoe desird atmeospericl ctonditions. For pO2 sym Ori


aove 0.01 atms masslwcnrleswr used to rxmn h O eednegulate athefow of rgons and air nt


atvthen apeproduing gasd flow ofene known compositio aith 40 e/min. Fo pOtrms of .00





atmo gandess a ZROXSM-Leetolssdvc a used to generate the desired amshrccniin.Frp~ to




concentration at a flow rate of 100 cc/min.












550oC *Dt
model
103C -----~Dissocati ve dorto
-----Electrolyte GB
Process 4 ----- Oxygen exchange
102

10 -, Fitting
Process 3"

100 .

Process 2
101 101 103 105 107
Frequency, Hz

Figure 4-2. Deconvolution of an imaginary impedance versus frequency profile into various
individual contributing processes.


4.3 Results and Discussion

Figure 4-1(a) and (b) are Nyquist and -Z" vs. frequency plots of symmetric LSM

samples sintered at 1100 oC for 1 h at various measurement temperatures in air. The

peaks and changes in slope of the -Z" vs. frequency plot indicate the characteristic

frequencies of the significant steps of the cathodic reaction. The 900 oC are in Figure

4-1(a) is basically semicircular, but not quite symmetrical. This geometry indicates that

there is one large resistance process enveloping another smaller resistance high frequency

process. At the lower temperatures di111lai-xd in 4-1(b), two other processes are apparent

which are attributed to oxygen vacancy diffusion through the bulk and grain boundaries of

the electrolyte [10]. These processes are smaller in magnitude than the cathode processes,

but at lower temperatures, the cathode processes are not seen in the profile. At higher

temperatures, the electrolyte processes have very small resistance and are overwhelmed by

the cathodic processes and inductive artifacts.

Each of the impedance profiles included in 4-1 was separated by a subtraction

technique into the various contributions. Because the primary focus of this work is

identification of the individual reaction steps separated by their respective time constants,

use of constant phase elements in the place of the capacitors in the Voigt elements of the









model is appropriate. Modeling with pure capacitors is limited in that multiple Voigft

elements may be required to model a single process step if the behavior is inhomogeneous,

otherwise, compromises may have to be made to match only a selected portion of the

curve. Also, if the homogeneity changes over the range of temperatures measured, the

depression angle of the arc may change, resulting in a non-Arrhenius dependence of

resistance and time constant on temperature. Figure 4-2 illustrates the separation of

the 500 oC measurement of symmetric LSM on YSZ cells sintered at 1100 oC into three

contributing electrochemical processes. Each of the indicated electrochemical processes

is modeled by a single R CPE Voigt element. This separation was repeated at all

measurement temperatures for the sample.

Figure 4-3(a) and (b) display the Arrhenius dependence of polarization resistance and

time constant for the LSM 1100 oC 1 h sample. The time constant (-r) is calculated as

described in Equation 4-1.

-r = (R x Q) (1/~) (4-1)

In Equation 4-1 R is the polarization resistance, Q and a~ are the non-exponential and

exponential terms of the constant phase element, respectively. The activation energy (E,)

for the polarization resistances of the individual reactions can be calculated as follows.


R = Ro exp (E,/kT) (4-2)


In Equation 4-2, R is the polarization resistance, Ro is a constant, k is the Boltzman

constant, and T is the temperature.

Examination of Figure 4-3(a) reveals that there is some change in slope in the high

temperature region of process 4. This is an indication that the changes in slope in the

high temperature regime were due to either changes in the homogeneity of the dominant

electrochemical process step, or a contribution from a different electrochemical process

with a different activation energy. Figure 4-3(b) indicates a somewhat smoother profile

in the same region for the same process. This supports the former of the two possibilities












air


1 /


4 ~2



S3


Process # Process Identity x in Equation 4-3 E, (eV) -r (s) at 800 oC
1 Ionic diffusion (bulk) 0.0 1.1
2 Ionic diffusion (grain boundary) 0.0 1.04
3 Charge transfer 0.0 0.97 8.5 x 10-s
4 Dissociation and surf. diff. -0.15 1.17 1.8 x 10-1
5 Gas diffusion through cathode -1.1 0.0 6.0 x 100


as the effect of varying depression angles on the resistance or time constant would be

minimized in the time constant calculation. K~nowledgfe of the Arrhenius dependence of

the polarization resistance and time constant of the individual electrochemical process

steps is all the information needed to generate the impedance profile as a function of

only temperature. The polarization resistance and time constant activation energies are

presented in Table 4-1.

Figures 4-4(a) and (b) display the influence of pO2 on the various electrochemical

processes occurring at 900 oC. pO2S Traging frOm 0.21 to 1 x 10-6 atm Were created by

mixing air with argon. The Figure indicates that a second process in the low frequency

regime gains significant magnitude at low pO2 ValueS. The process is first resolvable at

0.001 atm and grows as pO2 is further reduced. The dependence of polarization resistance

on pO2 is di1l-pl nd in Figure 4-5. The high frequency section of Figure 4-4(b) shows an


10


10


0. 10


10


air







-* 3*

*r-
.


104

10



10 2

104


10


I


101 1
0.


8 1 1.2 1.4 1.6 1.8 2 0.8 1 1.2 1.4 1.6 1.8 2
1000TT, K-1 b 1000TT, K'

Temperature dependence of the separated contributions in LSM on YSZ
sintered at 1100 oC for 1 h in air. The numbers indicate the process step
number given in Table 4-1. a) Polarization resistance. b) Time constant.

Select elementary steps of the cathodic reaction in samples sintered at 1100 oC.


Figure 4-3.




Table 4-1.












Juu C 900"C
200-
0.21 atm 1
0.06 atm
150 -* 1.0 X10 3atm
'2.0 X10 5atm =-1
N ~N .* .*.
S100 -' 1.0 X 10 atm- / *.
,*** 0.21 atm
50 = 0 *0.06 atm
50 rlo IIn =' ',.". 1.O x10 3atm **'


0 100 200 10 100 12 14 10"
a~z', a Frequency, Hz

Figure 4-4. Impedance response of an LSM cathode on electrolyte sample measured
at 900 oC with pO2 (atm) aS a parameter, a) Nyquist plot. b) Imaginary
impedance versus frequency.


increase in imaginary impedance for the lower partial pressure measurements. We believe

that the increase in imaginary impedance above 105 Hz is an experimental artifact due to

induction and not the result of some new high frequency process that is not present at the

higher pO2S. COmparison of Figures 4-5(a) and (b) shows that the low frequency process

has a significantly stronger dependence on pO2 than the other processes. The dependence

of polarization resistance can be expressed as a function of pO2 aCCOrding to the following

expression.

R oc (pO2) ( _q


The exponential value, x, is equal to the slope of the line in Figure 4-5 and its values

are indicated in 4-1. The strong dependence of this process on pO2 indicates that it is very

closely related to oxygen diffusion. Because impedance spectroscopy is an electrochemical

measurement technique, the diffusion of gas molecules can not be registered until they

are converted into a species that can participate in the electrochemical reaction [74]. A

rapidly adsorbing surface, which permits an oxygen flux equal to the flow of gas through

the open pores, would produce such an effect. The lowest frequency arc (process 5) in

the lowest partial pressure regime is due to a process limited by the bulk diffusion of










10

104 900 "C

103
~t5
102


10

101

107 10-6 105 10-4 103 102 101 1E
pO atm


10 ,,

10 ~800oC -


S 10 -"-
10 -3

,10 -

10


107 10-6 105 10-4 103 102 101 1(
a pO atm


Figure 4-5. Dependence of cathodic polarization resistances in an 1100 oC sintered sample
on pO2. The numbers indicate the process step number given in 4-1. a)
Measured at 800 oC. b) Measured at 900 oC.


oxygen molecules to the adsorbing surface. The two processes (process 1 and 2) which are

present in the high frequency regime and at low temperatures, become hidden at higher

temperatures. These processes are likely due to the electrolyte bulk and grain boundary,

a conclusion supported by their determined activation energies. This leaves two as yet

unidentified processes, which can be attributed to the cathode. One of these processes

has much smaller impedance and is located in the high frequency regime, and the other is

located in the lower frequency regime and has much larger impedance, greatly influencing

the overall shape of the Nyquist plot at high temperatures. A charge transfer located

at the TPB, would be rather fast process with weak dependence on pO2. Figure 4-5(a)

indicates that process 3 is nearly independent of pO2, While process 4 is has an exponential

dependence of -0.15. C'!I. ge transfer is two steps removed from molecular oxygen, while

adsorption directly involves molecular oxygen, therefore, process 3 is likely due to charge

transfer. Adsorption, dissociation, and surface diffusion are all possibilities for the identity

of process 4, which has a stronger dependence on temperature and pO2 than the charge

transfer reaction. The identities of the contributing processes are summarized in Table 4-1.









4.4 Conclusion

The impedance data in this chapter was fitted using a series R-CPE Voigft element

model to separate the individual processes contributing to the cathodic reaction. Three

cathodic processes were identified, two of which were present under all partial pressures

of oxygen. These three processes are measurement temperature dependent and are

indicated in Figfure 4-6. In the figure, processes 1 and 2 are due to ionic transfer through

the bulk and grain boundary of the electrolyte, respectively. (The process indicated by

0 is an artifact related to overcorrection from inductance in the system.) Numbers 3

and 4 indicate cathodic process which are related to charge transfer and dissociative

adsorption, respectively. Process 3 is of much smaller magnitude than 4 and close in

relaxation frequency and is therefore enveloped in the figure. Process 5 is only resolvable

at low pO2S and is related to bulk diffusion of oxygen gas to the reaction site. Polarization


"1 0
.. 3 0.002 % 02





10 <


0o "\ 40 oC70 O


10 ~



4-1.










resistance activation energies and time constants are generated from the model parameters

and given in Table 4-1. Since the process indicated by a 0 is not a result of any physical

phenomena associated with the cathodic reduction reaction, it is not included in the table.

It was found that charge transfer and dissociative adsorption processes were not strongly

dependent on pl)2, whereas the bulk diffusion related process was strongly dependent on

p ()2









CHAPTER 5
TERTIARY PHASE FORMATION

5.1 Introduction

For pure electronic conducting electrodes, the electrochemical reaction driving fuel cell

operation is restricted to the triple phase boundary (TPB), the area where the cathode,

electrolyte, and oxidant meet [55]. Authors have used composite cathodes to increase

the triple phase boundary length and shown that increasing the TPB area results in

reduced electrode resistance for LSM/YSZ systems [56, 57]. In dealing with single-phase

cathodes, triple phase boundary length can he maximized by ensuring high porosity in

the electrode and good adhesion between the electrode and electrolyte. Despite being

useful for high-temperature SOFCs, the performance stability of LSM on YSZ fuel cells

is an issue [2, 73, 86]. Much of the power loss in SOFCs is due to polarization loss at

the cathode/electrolyte interface, one source of which is degradation of this interface.

This degradation can he microstructural, such as severe coarsening and delamination,

or compositional as with the formation of tertiary phases. Tertiary phases, which may

form during fabrication or long term high temperature operation, are often insulating

and therefore detrimental to SOFC performance [87]. These and other effects can he

induced in a timely fashion through the use of high temperature anneals. The impact of

microstructural and interfacial changes caused by harsh anneals on the overall cathodic

reaction is studied in this chapter.

5.2 Tertiary Phase Formation

During high temperature operation, the diffusivity of the elements of the electrode

and electrolyte materials is increased. Consequently, La, hin, and Sr diffuse from the LSM

electrode into the YSZ electrolyte. Of these diffusing species, Mn has been shown to be

the fastest diffuser [68, 88, 89]. At 1000 oC, diffusion of Mn into YSZ is negligible, but

from 1200-1400 oC Mn diffusion becomes considerable [90]. A consequence of this fact is

that regions form near the interface that are deficient in Mn and therefore, high in La,

Sr, Zr, and O. If concentration of these elements becomes sufficiently high, formation









of lanthanum-zirconate (LZ, La2Zr207i) and strontium-zirconate (SZ, SrZrO3) may

become favorable. Formation of secondary phases can be predicted from chemical potential

diagrams [91]. Of these two phases, the one that materializes is determined by the Sr

dopant concentration in the LSM and the temperature of anneal [2, 89, 92, 93]. Both of

these phases have higher resistivities than YSZ so their presence is deleterious to device

performance [94-97]. Chiodelli and Scagliotti have measured the conductivity of the

LaZr207i 111-;-r via impedance spectroscopy [98]. The authors efficiently formed an LZ

1... -r by solid state reaction of lanthanum oxide and a single crystal YSZ substrate. The

work reports conductivities of 2 x 10-4 and 8.6 x 10-2 S cm-l for LZ and 9.5 mol .~ yttria

YSZ, respectively at 1000 oC. Additionally, it is reported that the conductivity difference

increases as temperatures are lowered. Y and Zr will also diffuse from YSZ into the LSM,

however, this diffusion occurs to a lesser extent as shown by Yang et al. [86]. The presence

of Sr in the LSM has been shown to suppress the diffusion of Mn into the YSZ [99]. This

leads to a phase composition and reaction 111-;-r thickness that is dependent on Sr content

[68]. K~enjo and van Roosmalen have independently reported that LZ formation can be

suppressed by using non-stoichiometric LSM [92, 96].

Taimatsu et al. found that there is a limited temperature range, in which secondary

phases form from solid state reactions, suitable for study. The work found that above

1450 oC, liquid phases were ahr-l- .- formed at Lal~nO3/YSZ interfaces and near 1300 oC,

reactions were too slow for their processes to be examined within a few weeks. Below

1425 oC reactions proceeded in the solid state, and morphologfies of reaction zones

resembled one another. Therefore, annealingf near 1400 oC was used to examine the

reactivity of Lal~nO3 With YSZ by the authors [100]. Other works have also used 1400 oC

anneals to produce reaction 111-;- rs [86, 88]. Yang et al. formed 3-4 micron thick reaction

1... ris consisting of SZ and LZ with a 1400 oC 48 hr firing of a screen printed LSM (0.3 Sr)

on YSZ interface.










In addition to being dependent on temperature and time, the formation of secondary

phases is also dependent on the interfacial area and type. It has been shown by K~leveland

et al. that secondary phases which form after 70 hr anneal at 1200 oC at a powder

mixture with large interface area won't form after 120 hr at a diffusion couple type

interface (which is similar to a single-phase electrode/electrolyte interface) and that a

harsher anneal is necessary to produce secondary phases at the diffusion couple interface

[68].

5.3 Experimental

The samples used in this chapter were prepared and impedance tested in the same

manner described is section 3.2. In this chapter, however, LSM powder provided by

Nextech was used. The powder was mixed with bonders, plasticizers, and solvents to

produce an ink of the desired viscosity. The LSM powder was stoichiometric with a 1:4

Sr to La ratio. After sintering at 1100 oC, samples were subjected to high-temperature

post-anneals intended to simulate the possible effects of long-term operation in a timely

fashion.

A JEOL 1400 SEM (scanning electron microscope) equipped with energy-dispersive

X-ray spectroscopy (EDS) was used for sample imaging. X-ray diffraction (XRD) was

performed on a Philips APD 3720 XRD. TEM analysis was performed using a JEOL

200CX TEM (tunneling electron microscope) by Mark Clark and is included in Appendix

A. All microstructural characterization was performed at the Major Analytical Instrument

Center (jl AIC) at the University of Florida.

5.4 Results and Discussion

5.4.1 Electrochemical and Microstructural Characterization

An 1100 oC, 1 h sintered sample was impedance tested and then subjected to

a 1400 oC 48 h anneal and retested. The effect of the 1400 oC 48 h anneal on the

electrochemical behavior of the symmetric sample is di;1l-p Iv4 in Figure 5-1. In this

figure a 600 oC measurement temperature was used in the frequency response analysis.









.......................................
600nT "C


2.5


With post-anneal



1 3.2 Hz

0.5 -
As sintered
0 0.5 11.5 2 2.5 3 3.5 4

Z', kaZ

Figure 5-1. Complex plane plots measured at 600 oC of symmetrical LSM on YSZ samples
as sintered and after a 1400 oC, 48 h anneal.

Upon comparison of the profile before and after the 1400 oC anneal, it is apparent that

one process with a characteristic frequency of 3.2 Hz is present in both spectra. Examining

the low-frequency regime reveals drastic change as the system exhibits linear behavior

with an angle close to 450. Such behavior has been reported in [64] and is described as

a Warburg impedance caused by a diffusion limitation. In the high-frequency regime, a

process appears in the post-annealed sample that was not apparent in the as sintered

sample. The causes and significance of these changes are discussed below.

Figures 5-2 and 5-3 display EIS profiles for the sample before and after the application
of a subsequent 1250 oC 48 h anneal, respectively. The EIS data was taken at low

measurement temperatures where the high-frequency behavior is emphasized. Comparison

of Figure 5-3(a) to 5-2(a) and 5-3(c) to 5-2(c) reveals little change in the high and

intermediate-frequency processes (electrolyte bulk and grain-boundary polarization

resistance, respectively) after the 1250 oC anneal. In the low-frequency regime the 1250 oC

anneal seems to have a clear effect. The low-frequency process (related to molecular

adsorption), which is semicircular in Figure 5-2, is replaced by Warburg behavior in Figure

5-3 (especially apparent at 500 oC). As explained by Macdonald in reference [64], Warburg
















*300 o


800
4 104 cl 400o~C
C 300 OC
N1 500 oC
~400
2 10 -

350 oC~ .-250 OC

0 2 10" 4 10 6 104 0 400 800 1200
a Z', at b z', a


250 o

10t ~r 300 OC
c350 u

10 r
450 oC
500 oC
101
10 101 10" 105 10T
Frequency, Hz

Figure 5-2. Impedance spectra for as sintered (1100 oC) sample at various measurement
temperatures. a) Complex plane plot. b) High-frequency view of the complex
plane plot. c) Imaginary impedance vs. frequency plot.
















6 10'~ 600 ~~0O
450 OC

4 300 OC 350oC.
4 10 400--


2 104 .250 OC '200 50/ 50 OC.

300 uC: 40~~~~ 0 OC

0 2 104 4 104 6 "104 0204060
a Z', 0Z b 2 zI, 40a 0



104 250 o

3 ~300 oC
mi103
~1 350 oC
'10 -
400 oC
10 -

500 o
10101 103 105 10T
C ~Freqluency, Hz

Figure 5-3. Impedance spectra for sample after a subsequent 1250 oC, 12 h anneal
measured at various measurement temperatures. a) Complex plane plot. (b)
High frequency view of the complex plane plot. c) Imaginary impedance vs.
frequency log-log plot.

























Figure 5-4. Scanning electron microscopy (SEM) images of the cathode/electrolyte
interface as sintered at 1100 oC.

behavior is generally a consequence of a diffusion limitation, thus the 1250 oC 12 h anneal

likely hinders the diffusion of ambient oxygen to the reaction site, significantly impeding

the cathodic reaction. This is due to coarseningf of the porous LSM to such an extent

that the diffusion of oxygen through the cathode is reduced or eliminated. Additionally,

coarsening of the LSM could greatly reduce or destroy the TPB, which will also inhibit the

cathodic reaction.

Figures 5-4 through 5-6 di pl w~ the microstructure of the symmetric samples at the

LSM/YSZ interface. Figure 5-4 shows the as sintered microstructure after an 1100 oC

1h anneal, while Figure 5-5(a) shows the microstructure after a 1250 oC 12 h anneal.

Comparison of the figures illustrates the coarsening of the LSM microstructure that occurs

with an anneal of this severity. This microstructural change forecasted by the changes in

the impedance profile with the harsh anneal is verified by the SEM images.

Figure 5-5(a) shows coarsening of the LSM microstructure after a 1250 oC anneal

of 12 h. After 48 h the coarsening becomes more pronounced (Figure 5-5(b)). The

coarsening of the LSM after 1400 oC anneal is more complete (Figure 5-6). In fact,

the coarsening after only 1 hr at 1400 oC occurs to a greater degree than after 48 h at

1250 oC. Focusing on the LSM/YSZ interface, it is noticed that for the 1400 oC annealed






















a b


Figure 5-5. Scanning electron microscopy images of the LS1\/YSZ interface for samples
sintered at 1250 oC'. a) Sintered for 12 h. b) Sintered for 48 h.


















a b

Figure 5-6. Scanning electron microscopy images of the cathode/electrolyte interface for
samples sintered at 1400 oC'. a) Sintered for 1 h. b) Sintered for 12 h.










6 10"


4 450 OC

4 10-
C 3400 o

2 1 -350oC 300 oC2- 30
1 210 50 oC



O 2 10" 4 105 6105 1 2 3 4 5
8 Z', a Z', kaZ


10 -3c 250 oC

10 300 oC





101
101 100 101 102 103 104 105 106 10'
C Frequency, Hz

Figure 5-7. Impedance spectra for 1400 oC, 12 h annealed sample at various measurement
temperatures. a) Complex plane plot. b) High-frequency vies of the complex
plane plot. c) Imaginary impedance vs. frequency plot.


samples, a coalescence of vacancies into extended interfacial pores occurs that is not

present in the 1250 oC annealed samples. The effect of this microstructural change is

inhibition of the cathodic reaction, by blocking electrons from reaching the TPB.

The low temperature frequency response of a test sample subjected to a 1400 oC

12 h anneal is di;11 pli a in Figure 5-7(a-c). Comparing Figure 5-7 with Figure 5-3

illuminates the differences between anneal at 1250 and 1400 oC on the impedance

spectra. The electrolyte bulk and grain-boundary processes present in Figure 5-3(a)

are replaced by one high-frequency process in Figure 5-7(a). The impedance magnitude

of the high-frequency process visible in Figure 5-7(a) is more than an order magnitude









greater that the high-frequency process of Figure 5-3(a). This increase is due to the

microstructural and interfacial changes that occur with the 1400 oC anneal including

the formation of insulating tertiary phases, eradication of the TPB, and the formation

of extended interfacial pores. Significant coarsening occurs at 1250 oC as di;1l-pl wd in

Figure 5-5, however it is clear from the impedance profiles that 1400 oC sintered sample

is significantly more degraded. Focusing on Figures 5-3(c) and 5-7(c), we see that the

low-frequency behavior of both systems is similar with the exception of the 500 oC

measurement, yet different from Figure 5-2(c). Although the 500 oC impedance profile

in Figure 5-6(c) appears to begin to close off at low-frequencies, it is likely that were the

low-frequency limit of the EIS decreased, a diffusion limitation tail would appear. This is

an expected result of the coarseningf of the LSM microstructure, which prevents free flow

of molecular oxygen to the electrolyte.

1000
1

CO A 24h
OIO + 48h


LL 400oC0

200 -


1050 1150 1250 13Z50 1450

Sintering Temperature (oqj

Figure 5-8. High-frequency arc resistance versus anneal temperature for various anneal
(temperature, time) pairs measured at 400 oC.


Figure 5-8 illustrates the dependence of the high-frequency EIS contribution on

anneal temperature and time. The figure di pl .m-~a gradual increase of high-frequency

arc resistance with anneal at temperatures below 1400 oC. In this temperature range

there is no clear dependence on anneal time. At 1400 oC, however, the impedance










increases dramatically and a positive dependence on time appeal rs. This change occurs

despite Figure 5-6 which shows that by 1 h the microstructure has become nearly

dense, leaving little possibility of increased coarsening with longer time anneals. Factors

other than microstructure must he considered when explaining the observed increase in

high-frequency impedance.

5.4.2 Compositional Characterization


Ct

r
CI
r









r
CI
r


Figure 5-9. Energy-dispersive X-Ray Spectroscopy (EDS) linescan of hinE~n intensity at
LSM/YSZ interface. a) As sintered. b) After 1400 oC', 48 h anneal.


Figure 5-9 shows EDS linescans of the manganese Knc intensity at the LSM/YSZ

interface of as sintered samples with and without a 1400 oC' 48 h anneal. Despite the fact

that Mn is known to be a fast diffuser at high temperatures [88, 90], the profile is more

abrupt in the harshly annealed sample. This apparent contradiction can he explained by


M n-Koc


O 10 20 30

Distance from interface, pLm










considering the interfacial effects that occur in the harshly annealed sample. Extended

pores at the cathode/electrolyte interface will physically hinder or prevent diffusion of Mn

from the cathode to the electrolyte, effectively creating an abrupt interface. Alternatively,

the formation of a Mn free phase such as lanthanum or strontium zirconate would also

produce a more abrupt profile due to the expulsion of Mn from the region of formation

during growth. Additionally, as Mn diffusion into the zirconate is blocked, a buildup of

Mn at the edges of the zirconates is likely, resulting in a more abrupt Mn profile.

The 1250 oC, 1325 oC, and 1400 oC, 12 h annealed samples were examined by

XRD as di;111 i-, II in Figures 5-10(a)-(c), respectively. In all samples, energy peaks

characteristic of LSM and YSZ were observed. In addition to these peaks, there is

evidence of a tertiary phase present at the interface of the 1325 oC and 1400 oC annealed

samples. The proximity of some of the tertiary phase peaks to the peaks of LSM hinder

the analysis. To reduce the complexity of the XRD spectra, the samples were bathed in

highly concentrated hydrochloric acid to remove the LSM 1.>. -r. XRD of the LSM stripped

samples clearly show a set of peaks corresponding to YSZ and another set of peaks for the

1325 oC and 1400 oC annealed samples. The second set of peaks matches those belonging

to a known lanthanum zirconate (La2Zr207i) sample. Additional compositional analysis

using tunneling electron microscopy was performed by Dr. Mark Clark and is included

in Appendix A. These results show that a 0.2 micron thick transitional region rich in

lanthanum and zirconium forms at the LSM/YSZ interface after annealing at 1400 for

48 hours. The interfacial region was shown to have a similar crystal structure as YSZ,

indicating that manganese diffuses into the electrolyte as the tertiary phase forms.

5.5 Conclusion

Electrochemical and compositional analysis has been used in this chapter to

investigate the effects of harsh anneals on the electrochemical processes contributing

to the cathodic reaction occurring at the LSM/YSZ interface. The electrochemical

evaluation focused on low measurement temperatures to emphasize high-frequency



























132 oC Stipe S







05



20 30 40 50 60
b An gle / 2 0 d eg ree s

35
1400 *C Stripped LSM



15 -
-11

.As iter nedl



CU Angl 120degree

Fiue5-0 -rydfrato of sape ujce ops-nelsneig oLek
are obere fo th 20o itrdsml.)Snee t15 C b
Sitrda 12 C ) itrdat10 C










processes. We have seen three arcs which are related to the electrolyte bulk, electrolyte

grain-houndary, and an adsorption process in the previous chapter. Application of a harsh

anneal of 1250 to 1400 oC' increases the impedance of the electrode dependent are. This

increase is attributed to changes in the microstructure of the LSM, which are observed

via SENT and formation of the tertiary phase, lanthanum zirconate. Application of a

1400 oC' anneal destroys the TPB leading to a high-frequency are in the EIS spectra

indicating that the cathodic reaction is hindered. This are is an order of magnitude

greater in impedance magnitude than the high-frequency processes present in samples

annealed at lesser temperatures. Through EIS, it is found that the impedance of this are

increases with time annealed at 1400 oC'. It is found that sinteringf above 1325 oC' produces

compositional changes that are consistent with the formation of lanthanum zirconate at

the cathode/electrolyte interface.









CHAPTER 6
THE RELATIONSHIP BETWEEN CATHODE MICROSTRUCTURE AND
ELECTROCHEMICAL PERFORMANCE

6.1 Introduction

During fabrication or longterm high-temperature operation, microstructural changes

occur which affect performance of LSM on YSZ devices [81]. Because high temperatures

are required for fabrication, to some extent these changes can not be avoided. The overall

result is that the cathodic reaction, which is dependent on oxygen gas flowing through

the pores, diffusing towards the reacting site, and being transferred to the electrolyte

is affected by the altered microstructure. The impact of microstructural and interfacial

changes on the electrochemical steps contributing to the overall cathodic reaction is

investigated in this chapter.

In a previous chapter, the impact of very high-temperature anneals on the cathodic

reaction was examined. Both dramatic changes in microstructure were produced and

tertiary phases were formed at the cathode/electrolyte interface, altering the cathodic

reaction. In this chapter, microstructural changes are produced by sintering at lower

temperatures in an attempt to decouple microstructural changes and tertiary phase

formation. Additionally, non-stoichiometric LSM ((Lao~sSTO.2~l *M -nO3-6) WaS used which

has been shown to decrease formation of tertiary phases [92, 96]. Establishing a direct

relationship between cathode microstructure and electrochemical performance will clarify

the cathodic reduction reaction pathway and aid in identification of the rate-limiting step,

which is not known conclusively [73, 74].

Several works have been performed with the goal of establishing a relationship

between LTPB and electrochemical properties. Typically, the DC resistance, or the

entire cathodic resistance determined from impedance spectroscopy are related to

electrochemical performance. One of the most frequently cited works in the area was

performed by Mizusaki et al. who found that total cathode conductivity, measured by

impedance spectroscopy at 1000 oC, is essentially proportional to (LapB)-l for drip










pyrolosis prepared Lao~sCa0.4A~Ons, LC11l/YSZ/platinum cells sintered between 1100

and 1200 oC [101]. The results from this work work are somewhat questionable since

only two data points are plotted to give the observed dependence. Additionally, when a

different fabrication technique was used, a non-linear power dependence of total cathodic

polarization resistance on LTPB was reported.

Another frequently cited work was performed by K~uznecov et al. [102], who derived

a relationship successfully explaining the linear dependence reported by Mizusaki i.e.

R,.athoa, oc (Lapg)-l. In the model, the author assumes that surface diffusion of adsorbed

species towards the TPB dominates Rp and that the DC resistance can be modeled by

considering this flow of adsorbed species to the TPB. The model basically relates the

cathodic resistance to the flux of adsorbed species on the surface of the cathode. As

reported by Macdonald et al. a surface diffusion limitation is often manifested by Warburg

behavior in the impedance profile [110]; however, we did not see Warburg behavior at

high-frequencies and so the development described in the work may not he appropriate for

our data.

In a later work by K~uznecov et al. it is reported that for kinetics controlled by bulk

diffusion through the cathode, Re-onzoa, oc(illlEC/YSZ contact area xLaps)-o5 [10:3] This

development is based on the work of Adler et al. [52]. The model of Adler et al. was

intended to describe cathodes with significant ionic conductivity and is reported by Adler

et al. to be inconsistent with LSM behavior. This model considers flow of ionic species

through the cathode bulk to be the limiting factor and calculates the conductivity from

this flow. From the data of K~uznecov et al., a power dependence of -0.39 can he calulated

(from only two data points) when relating total Re-athode to LTPB for LSM measured at

950 oC [10:3].

In another work, Fleig showed that for well defined, dense patterned LSM microelectrodes

of circular geometry, the total cathodic resistance is proportional to electrode diameter

(D) to the -2.1 power when measured at 800 oC. In this geometry, LTPB is equal to the









circumference (C) where C = xrx D and therefore, Rp oc (LapB)-2.1. Additionally, Fleig

reported that the resistance scales almost linearly with thickness and that application of

a bias can change the exponent of that relationship [50, 104]. Fleig concluded that since

total cathodic resistance scales inversely with electrode contact area (area = 0.25 xx XD2)

and linearly with thickness, a bulk path through the electrode determines the oxygen

reduction rate, with transport of oxide ions in LSM being the rate-determining step. It

should be pointed out that the dense circular electrodes used by Fleig had thicknesses

of only 0.1 to 0.25 pm and diameter on the order of 60 pm, so ionic transport through

the electrode could be appreciable. In contrast, in the present work we used porous

electrodes with thickness on the order of 20 pm and individual particle sizes of around 1

p-m. Therefore, different reaction pathi-ws- are likely. For example, if an oxygen molecule

is adsorbed at the center of the top of one of the circular disks, the adsorbed species must

travel over 30 pm to reach the electrolyte via a surface path, but only 0.1 pm to reach the

electrolyte through the cathode bulk. On the other hand, for relatively spherical particles,

the path to the electrolyte via the surface will be around the same distance for both

surface and bulk diffusion. Obviously, both the shape of the cathode and the relative ease

of transport through the bulk versus on the surface will determine which path is favorable

for an adsorbed species. Additionally, surface area (not volume normalized surface area)

also scales linearly with cathode thickness and so additional evidence is needed to support

the bulk transport conclusion.

One of the first authors to utilize knowledge of a relationship between polarization

resistance and LTPB was Ostergard et. al. who decreased polarization resistance by

forming composite cathodes which have increased LTPB [57]. It has been reported that

the dependence of polarization resistance on composite cathodes thickness depends on the

measurement temperature and that as composite cathode thickness increases, polarization

resistance decreases until gas diffusion effects become important. [105, 106]










Because the electrolyte is sintered to a dense state at temperatures higher than

the operation temperature or any other fabrication steps, its microstructure is generally

considered to be stable. However, the porous microstructure required for the cathode is

sensitive to the sintering of the cathode and possibly operating conditions as well. A great

deal of microstructural analysis has been performed, however most of the previous work is

either a surface technique such as BET (Brunauer, Eninett, Teller adsorption technique),

which is usually used for powder samples or a two-dintensional microscopy technique such

as conventional SENT [75]. In contrast, this work makes use of a dual beam FIB/SEM

(Focused ion heant/scanning electron microscope) for microstructural characterization that

allows 3-D reconstruction. Of the microstructural parameters coninonly studied, four are

considered to be the most critical to electrochentical efficiency. These include pore surface

area, triple phase boundary length (LTPB), porosity, and tortuosity.

An oxygen molecule, which has diffused through the gas phase to the cathode, must

first he adsorbed before it can participate in the reduction reaction. This adsorption can

occur very close to the TPB or further away, depending on the diffusivity of the adsorbed

species. It has been proposed in fact, that oxygen reduction in an electronic conductor can

he co-linlited by both adsorption and surface diffusion [107]. Both adsorption and surface

diffusion are dependent on the pore surface area; therefore, pore surface area is one of the

key microstructural parameters for our investigation.

As coarsening occurs, small cathode particles at the interface coalesce into larger

ones, thus reducing the total TPB length per surface area. Other authors have shown that

increasing LTPB results in reduced electrode resistance for LSM/YSZ systems [57, 108].

This reduction is a direct consequence of the fact that in pure electronic conducting

electrodes, the electrochentical reaction driving fuel cell operation is restricted to the TPB

[55] due to the exclusion of ions front the bulk of the electronically conducting cathode

and of electrons front the bulk of the electrolyte. We therefore can anticipate an increase

in charge transfer resistance as sinteringf temperature is increased.










Porosity is the ratio of the void space in the microstructure to total volume. Before

the cathodic reaction can occur in an electronic conducting cathode such as LSM, oxygen

molecules must first diffuse through open pores to the vicinity of the TPB, the area where

the cathode, electrolyte, and oxidant meet. An ideal microstructure has ample void space

for molecular gas diffusion, while a partially dense microstructure impedes the flow of

molecules to the TPB, thus inhibiting the cathodic reaction.

Tortuosity is a property that quantifies the complexity of the path through which a

diffusing particle must travel in order to reach a desired destination. In terms of SOFCs,

tortuosity is a unitless parameter defined as the distance traveled by a molecule exiting an

impinging gas flow as it travels through the porous cathode to reach the solid electrolyte,

divided by the straight-line distance. A large tortuosity corresponds to a convoluted

path for a given gas molecule to traverse in order to go front the gas stream to the TPB.

Because data in three dimensions is necessary for a true tortuosity analysis, very little

work is published for actual systems. The dual beam FIB gives us the three-dintensional

data necessary for the niathentatical evaluation. We expect that cathode microstructures

with a large tortuosity will show an increase in gas diffusion polarization resistance and

related electrochentical properties.

6.2 Experimental

The samples were prepared and impedance tested in the same manner described

is section :3.2. In this chapter, samples were sintered at temperatures ranging front

950 to 1:325 oC for one hour. The sinteringf temperature range was chosen to produce

microstructural changes which could be quantified and compared to changes in the

electrochentical behavior of the samples. Microstructural images were attained using a

dual beam focused ion beam / scanning electron microscope (FIB/SEM, FEl Strata DB

2:35) by Aijie C'I, .. The FIB/SEM setup is described in Appendix B.





































Figure 6-1. Scanning electron microscopy (SEM) images, created using a focused ion beam
/ SEM (FIB/SEM), of LSM on YSZ sintered for 1 h at various temperatures.
a) Sintered at 1100 oC. b) Sintered at 1200 oC. c) Sintered at 1300 oC.

6.3 Results and Discussion

6.3.1 Effect of Sintering on Microstructure

Figures 6-1(a) 6-1(c) show interfacial cross sections of LSM on YSZ sintered at 1100,

1200, and 1300 oC, respectively. It is easy to see that by 1300 oC, the microstructure

changes drastically. Comparison of figures 6-1(a) and (b) reveals more subtle differences.

The 1100 oC annealed sample appears to be slightly more porous than the 1200 oC

annealed sample. This apparent difference is so slight that to conclusively ?-w there is any

change in porosity requires quantitative calculations. High-frequency artifacts in the data

were accounted for as described in section 3.3.2

Visually, the most noticeable difference between the 1100 and 1200 oC sintered

samples is in the connectivity. The 1100 oC sintered sample shows many round shaped










particles with near point-to-point inter-particular unions. The individual nature of

many of the particles is maintained at 1100 oC. By 1200 oC, the individual nature of

the particles is compromised. The inter-particular unions have become primarily of the

face-to-face variety with face diameter almost equal to particle diameter; the connectivity

has increased. By 1300 oC, the coarsening has progressed to the point that individual

particles no longer exist; the material is now very connected. As the particles coalesce into

large particles, the number of pores decrease and only a smaller number of large pores are

left, effectively doubling the average pore diameter between 1200 and 1300 oC.

Because the TPB is of particular importance to the cathodic reaction, we closely

examined the cathode/electrolyte interface. On initial inspection, it appears that the

cathode to electrolyte contact surface is larger for the 1200 oC sintered sample than

the 1100 oC sintered sample. Several of the particles close to the electrolyte for the

1100 oC sintered sample do not appear to contact the electrolyte. In contrast, for the

1200 oC sintered sample face-to-face contacts have been formed between the cathode

and the electrolyte. The nature of interfacial voids also changes significantly between

1200 and 1300 oC. Figure 6-1(c) shows the formation of large interfacial voids at the

cathode/electrolyte interface. These large interfacial voids form as smaller voids coalesce

while being restricted from the dense electrolyte. Formation of these interfacial voids will

greatly reduce the measurable TPB length.

FIB/SEM was performed on samples sintered at the various temperatures and

microstructural features were quantified as described in Appendix B. Porosity (p), volume

normalized pore surface area (Sv), and TPB length values were calculated at each

temperature, while tortuosity (-r) values were calculated from the 3-D data attained at

selected temperatures. Porosity was calculated from the pore area/total area in each SEM

image. The calculation was repeated for all slices in the sample and an average porosity

was attained. These results are plotted in Figures 6-2 and 6-3.










61 1.6
5 -1 11.4-




3 1-

a, n0.2
0) 0 0.

1150 1200 1250 1300 1350 1150 1200 1250 1300 1350
a Sintering Temperature, 0C Sintering Templerature, 0C

Figure 6-2. Pore surface area and LTPB as a function of sintering temperature. a) Pore
surface area. b) LTPB.


Fr-om Figure 6-1, we can see that as the sintering temperature increases from 1100 to

1300 oC, the microstructure changes from one with small pores to a microstructure with

large pores. From elementary geometry, we would expect a microstructure with many

small pores to have a larger surface area than one with large pores. Figure 6-2(a) confirms

this findings and shows that the volume normalized pore surface area decreases as sinteringf

temperature increases from 1150 to 1325 oC.

Figure 6-2(b), shows that LTPB decreases linearly as sintering temperature is

increased in the temperature range di;11l-phi I. There are outliers at 1225 and 1250 oC.

These deviations could be caused by localized interfacial voids, an unusually fine

interfacial microstructure, or a lower than anticipated sinter. A determination as to

whether the cause of the outlying points is a local anomaly or a characteristic of the bulk

sample can be made by studying the electrochemical behavior of the samples in question,

which is performed later.

Figure 6-3(a) shows that the porosity (calculated by Aijie C'I. 1.) starts at about t:I' .

for a 950 oC sintered sample and increases slightly with increasing sintering temperature

to 1200 oC and then begins to drop off. By 1400 oC (not shown), the porosity has dropped

to less than 5' indicating an almost dense cathode 1 s. r. This trend is supported by our










40 4.5

35 4 -



20-
15 2 .5 -
10 25
90 00 10 20 10 40 10 15 1015 2015 3015







YSZ isonte rero 1450 oC0 [100]. 0 0015 10 1010 15 3015


Thgue Prst n tortuosity w as caclae (clulatedn by Aijteie I. o slc temperature. )Prsit.





Tortuosity values of 3.23, 2.18, and 4.27 were calculated for the 1100, 1200, and 1300 oC

annealed samples, respectively. The minimum tortuosity occurs at about 1200 oC

indicating that the gas molecules have the most direct path to the interface. Opposing

trends accounts for the minimum that is observed. At low sintering temperatures, particle

size remains small, and gas molecules are redirected many times as they traverse the path

to the LSM/YSZ interface. At higher sintering temperatures the pores are large, however,

some of the paths may become closed off, limiting the number of available pathi- .--s.

6.3.2 Effect of Sintering on Impedance

Figures 6-4 and 6-5 show 800 oC impedance measurements of LSM on YSZ sintered

at various temperatures in air. Figures 6-4(a) and (b) are Nyquist plots covering the

entire frequency range and high-frequencies only, respectively. The profiles shown in

Figure 6-4(a) are generally .I-i-inin.! r lical in the high-frequency regime. The effect is

more pronounced in Figure 6-4(b), which only shows the highest frequency portion of

the data. The cause of this .I-i-mmetry is the presence of multiple processes occurring














Z', O2
150 200


-'150

-"100


-40


-32


1175 OC 4 ,'^





10 8 16 24
Z', st


Figure 6-4. Nyquist plots measured at 800 oC for LSM sintered at various temperatures in
air. a) All frequencies included. b) High frequencies only.














OC I ,@/4."** 10
12750 C 10
11150 O
1225 OC .
-101
1200 oC


101 100 10 10 10 1104 Os 10
Frequency, Hz

Figure 6-5. Imaginary impedance vs. frequency plot measured at 800 oC for LSM sintered
at various temperatures in air.


over the frequency range examined. As sintering temperature is increased, the presence

of the high-frequency process becomes more pronounced as seen in Figure 6-4(b). Figure

6-5 di pl wa~ the frequency dependence of the imaginary impedance. Dj~pl w~ing the data

in this format makes apparent the decrease in peak frequency of the overall reaction as

sintering temperature is increased. The 1325 o"C sintered sample shows a change in slope

at about 1 kHz. An inflection is only observed when two or more electrochemical processes

are significant.

Impedance Spectroscopy of LSM cathodes on YSZ substrates has been the subject

of a multitude of works [53, 109]. Most authors agree that two noticeable processes occur

in optimally sintered LSM on YSZ at high measurement temperatures in oxygen rich

atmospheres. At low oxygen partial pressures a third process related to the diffusion

of oxygen gas molecules through the open pores of the cathode to the active region is

observed.

Unfortunately, agreement on the isolation and identification of the high and

intermediate-frequency processes has not been as complete. Reasons for disagreement

include 1) the mechanism of reaction is dependent on measurement conditions, 2) the

mechanism of reaction is dependent on the sample preparation and sample history,










CPE1


Figure 6-6. ?-. -i. I1 element equivalent circuit used for fitting. Zhf represents features
occurring at too high a frequency to be analyzed. CPE1 is associated with
the double 1.,-< c capacitance, R1 is the charge transfer resistance, and R2 and
CPE2 are related to adsorption.


and :3) no consensus is reached for evaluation of impedance data. To overcome the

first two problems it is important for authors to specify as completely as possible all

experimental details, particularly when microstructure is not analyzed. In this work, we

have characterized the microstructure and will relate electrochemical properties directly to

the microstructure of each sample. The third problem is not easily solved.

Typically, impedance data is >.1, llh-. I1 by fitting the data to an equivalent circuit.

One school of thought proposes developing a model which is based on a priori knowledge

of the system. Several authors have >.Is llh-. 1 LS1\ on YSZ using this method. The most

often used circuit contains a double 1 e -< c capacitance in parallel with a series connection

of a charge transfer resistance and a mass transfer related element. For electronic

conductors, the mass transfer interpretation is replaced by adsorption and/or surface

diffusion. The mass transfer related element is either a Voigt element, a finite-length

Warburg element, or some general diffusion element that is not easily defined in terms of

circuit elements. Additionally, all capacitors may be replaced by constant phase elements

to account for inhomogeneities in the system. This type of circuit with slight variations

has been used by several authors and is depicted in Figure 6-6 [46, 48, 6:3, 112, 11:3].

The 1!! I i .r drawback of this model is that each author typically has their own variation

of the model making comparison of parameters attained between groups difficult. The

commonly used nested circuit shown in Figure 6-6 (with capacitors instead of constant

phase elements) was produced from a more general model in a work by Jamnik and











**


Figure 6-7. Series Voigt element equivalent circuit used for fitting. Zhf represents features
occurring at too high a frequency to be analyzed. Each Voigt element is
composed of a resistor and a constant phase element.


M.~ i-n r [62]. In the model, CPE1 represents the double 1 ore-r capacitance, R1 represents

the charge transfer resistance, and R2 and CPE2 represent a mass transfer phenomenon.

Henceforth, we will treat LSM as an electronic conductor and therefore replace the mass

transfer process by adsorption and/or surface diffusion. Macdonald explains that when

surface diffusion is significant the Randles equivalent circuit is expected; however, if no

significantly diffusing intermediates are present the diffusional impedance is replaced by a

resistor and capacitor in parallel [110]. In this work, a Warburg type slope was not seen

at high-frequencies; therefore, it is likely that adsorption is more significant than surface

diffusion.

An alternative equivalent circuit based on a series connection of Voigt elements

is also used by many authors and di;11l li-- 4I in Figure 6-7 [54, 84, 114, 115]. In this

type of model, assignment of identities to the individual processes is accomplished by

identification of activation energies, pO2 dependence, bias voltage dependence, and

other circumstantial evidence. The 1!!I i ~r advantage of modeling in this fashion is that

comparison of efforts between different groups is facilitated; however, because the model

is not derived specifically for the system, confidence is diminished. Jiang et al. has used

error structure an~ lli--; to show that both of these models can accurately produce the

desired response [116]. In our previous work, we examined activation energies, pO2

dependence and other evidence and concluded in agreement with others that charge

transfer was the high and dissociative adsorption was the intermediate frequency processes

[111]. In both models, Zhf represents the total impedance of all processes occurring at










too high a frequency to be represented in the frequency response. These processes include

electrolyte resistance, residual inductive artifacts and any ohmic resistances.

Looking back at Figures 6-4 and 6-5 we see that the intermediate frequency process

has a larger polarization resistance magnitude but that the high frequency process

increases in relative magnitude as sinteringf temperature is increased. It should be

pointed out that a larger magnitude means a larger power consumption due to that

mechanism, but does not necessarily mean that that mechanism is the rate-limiting

step. An increase in charge transfer resistance is evidence of a decrease in the quality

of the cathode/electrolyte interface, where charge transfer occurs. ('I! Iage transfer

polarization resistance becomes more significant at higher sinteringf temperatures due to

the deterioration of the triple phase boundary. As sintering temperature is decreased, this

high-frequency process becomes less pronounced and inductive artifacts become significant

in the high-frequency portion of the data.

Both equivalent circuit models were used to fit impedance profiles such as the ones

shown in Figure 6-4. Figure 6-8 is included as an example illustrating the deconvolution

of the data. The impedance profile shown is for the 12000C' sintered sample measured at

800 o"C air. Figure 6-8(a) shows the data and the fitting obtained using the nested model,

while Figure 6-8(b) shows the data along with the fitting (solid line) from the series

model. Additionally, Figure 6-8(b) shows the individual components which are summed

to produce the series model fitting. Because both models accurately fit the data, more

analysis is necessary to determine which of the two is more appropriate.

Since the measurement was done in air, the polarization resistance due to bulk gas

diffusion is negligible and only two Voigt elements are necessary in the series fitting, one

for adsorption (dashed line) and one for charge transfer (dotted line). As can he seen in

Figure 6-8(b), a single process with relatively large magnitude (adsorption) provides the

1! in r contribution to the profile. Above 104 Hz, charge transfer becomes significant and

causes the overall profile to deviate from the symmetric contribution due to adsorption.













"102


100 -Fitting


102

a 10o ~ 100 1 0 1 02 10 104 10 106
Frequency, Hz


"102


10'


'100


'1 0 ~ 1 ******' ******' *****' ******' "** "" ** ** ** ***
b0 lo- loo o'1 o2 103 1(4 105 106
Frequency, Hz

Deconvolution of impedance profile from 1200 oC sintered sample, measured at
800 oC' in air, using both equivalent circuit models. a) ?-. -ib 1 model. b) Series
model.


Figure 6-8.










From each of the Hoigt elements used, polarization resistance (Rp) values and constant

phase element parameters were attained for charge transfer and adsorption. From the

fitting using the nested equivalent circuit, double-l} u. r capacitance, charge transfer

resistance, and parameters associated with adsorption were attained.

The process was repeated at various sinteringf temperatures ranging between 1150 and

1325 oC'. The sinteringf temperature range was chosen to begin above temperatures where

sinteringf is incomplete and end below the melting temperature of LSM on YSZ, which

is around 1450 oC' [100]. Previous research shows that by 1400 oC', the charge transfer

resistance has increased dramatically because the LSM 1w-;r is fully dense, effectively

(1. -1 i~elim;~! any triple phase boundaries [81]. The impedance was performed at 800 oC' in

air. In future work, we will examine the cathodic reaction in low oxygen partial pressure

regime and relate the bulk gas diffusion polarization resistance to porosity and tortuosity.

Figure 6-9(a) shows the sintering temperature dependence of charge transfer Rp

determined from both models. For both models, the charge transfer polarization resistance

increases exponentially as sintering temperature is increased. The individual nature of the

Hoigt elements in the series model may contribute to the improved fit for the series model

as compared to the nested model for charge transfer resistance. The relatively large scatter

in the charge transfer data for the nested model is related to the fact that charge transfer

Rp is an order of magnitude smaller than the adsorption Rp, and the two processes are

solved for simultaneously. In contrast, a subtraction technique which removed processes

individually was used in the deconvolution for the series model. Figure 6-9(b) shows the

dependence of adsorption Rp on sinteringf temperature.

6.3.3 Effect of Microstructure on Impedance

6.3.3.1 Series model evaluation

Figure 6-10 relates the change in electrochemical performance caused by varying

sintering temperature to the corresponding microstructural changes by showing the

influence of TPB length on charge transfer Rp and the influence of pore surface area

























n
rr *10
a,
e
u,
r=


a,
3
r
O


1LI
1100


1150 1200 1250 1300

Sintering Templeraturel oC


*1350


40 i'
1100


1150 1200 1250 1300

Sintering Temperarture 0C


1350


Figure 6-9. Temperature dependence of polarization resistance (Rp) in air determined
using both series and nested equivalent circuits measured at 800 oC'. a) ('I! Ivge
transfer Rp. b) Adsorption Rp.
















102 -\I~ lCII



\LCharge Transfer vs. L


-1225 OC


100
0.3 13 10



Figure 6-10. Relation of charge transfer and adsorption polarization resistance determined
front the series Hoigft element equivalent circuit to microstructural quantities
(measured in air at 800 oC).


on adsorption Rp with electrochentical parameters determined front the series Hoigft

element model. Focusing first on charge transfer, we see that the charge transfer resistance

increases as triple phase boundary length decreases. In the LTPB vs. sintering temperature

plot shown in Figure 6-2(b), we noted that there were two outlier points located at 1225

and 1250 oC' and proposed reasons for their presence. When relating charge transfer Rp

to the actual microstructure, LTPB, we observe only one outlier located at 1225 oC'. The

cause of the outlier at 1250 oC' must he a bulk characteristic because the rise in LTPB waS

accompanied by a corresponding drop in charge transfer Rp.

Turning our attention to the data point for the 1225 oC' sintered sample, we see

that the lowered LTPB value is not accompanied by a corresponding increase in charge

transfer Rp. In fact, the charge transfer Rp for the 1225 oC' sintered sample is similar in

magnitude to the 1200 and 1250 oC' sintered samples. Since the low LTPB value seen at

1225 oC' is not accompanied by an increase in charge transfer Rp, we can conclude that

the low LTPB value is caused by a localized phenomena such as an interfacial gap and is










not due to a characteristic of the extended microstructure. It should also be pointed out

that despite this interfacial gap, the 1225 oC point fits the model when considering the

error bars.

Curve fitting of the data indicates that there is a power-law dependence of charge

transfer Rp on LTPB at an 800 oC measurement temperature given by Equation 6-1.


Rp = 2.93(LapB)-3.5 (6-1)


Other authors have shown a dependence of overall polarization resistance on the

inverse of LTPB at a measurement temperature of 1000 oC. Mizusaki et al. assumed

that charge transfer in not the rate-determining reaction in the development of their

model [34]. K~uznecov et al. developed a model which assumes that oxygen reduction

takes place everywhere on the LSM surface, i.e. charge transfer is not limited to the

triple phase boundary [102, 103]. In this scenario, polarization resistance is predicted to

be proportional to (Laps)-l. This dependence was explained using models which were

based on surface diffusion limitation. At lower operating temperatures, reaction kinetics

associated with the charge transfer reaction may become the rate limiting step. A power

law dependence can also be predicted if we consider the reaction kinetics associated with

the charge transfer reaction. For a given chemical reaction, the rate of the reaction, v, is

dependent on the concentration of the reacting species ca4 and CbB.


v= k(cA a CB b (6-2)


After adsorption of gas molecules on the surface of the cathode, a charge transfer reaction

occurs at the cathode/electrolyte interface. J. Nowotny et al. outlined the various possible

adsorption-charge transfer reaction combinations in reference [49] (see Figure 2-1). Let

us assume, for now, that adsorption and charge transfer occur according to the following,

respective, reactions.

02 + S 02,ads (6-3)










2e' + ,02,ads + V" O + s (6-4)

In Equation 6-4, s is a surface site. The previous equations describe the molecular

adsorption of oxygen followed by a separate charge transfer step. The rate of the charge

transfer reaction indicates how quickly charged species are being transferred across the

cathode/electrolyte interface. This exchange of charged species determines the exchange

current, Io. It is convenient to consider the exchange current density (io), which is the

exchange current per unit area and is given by the following relationship.


Io = io x Alne (6-5)


In Equation 6-5, Aint is the planar area of the cathode/electrolyte interface, i.e. 64 mm2

in this work.

If the individual species of the charge transfer reaction are treated as reactants and

products, then the exchange current density can be expressed in Equation 6-6.


io= Q (kyS(ce)m(cozns, )"(c1vo.)p k,.(co,)v(c,)') (6-6i)


In Equation 6-6, the forward and reverse rate constants are given by kf and k,, and

Q accounts for balance of units. In Equation 6-6 ce,, coz,,, C, co., and c, are the

concentration of electrons, adsorbed oxygen, oxygen vacancies, and surface sites on

the cathode able to participate in the cathodic reaction, respectively. The second term

on the right side of the equation describes the rate of the reverse reaction. Because the

electrochemical measurement was performed in an oxygen rich atmosphere we will assume

this term can be neglected for simplicity.

The interfacial reaction is impeded by a charge transfer resistance, Ret, which under

equilibrium conditions has been shown to be inversely proportional to lo.


Ret =9R\ (l (6-7)


In Equation 6-7, R is the gas constant, T, n, and F have their usual meaning.









Substituting Equation 6-5 into the charge transfer equation, Equation 6-7, gives the

following expression for Re-.

Ret =( .Tn ji (6-8)

Using io from Equation 6-6, we can express Ror as described in Equation 6-9.


Ret = 'y(ce )pmB (coz,d, a)T~B (cV.)- (6-9)


In Equation 6-9, Q' = Q Aine nF/RT.

The amount of species "\i" available to react per unit area ((c,)TPB) is limited by

LaPB per unit area. In order to relate the number of species in the vicinity of the TPB

to the bulk concentration of species in their respective phases, we need to multiply the

concentration of species by the TPB volume of each respective phase.


ci = (c,)r PB = (ci) VTPB,i = (ci) (L, PB) Are,i (6-10)


In Equation 6-10, Aren,i is the cross-sectional area of the TPB for each of the respective

"\i" phases. Substituting the effective concentrations of active species [i]TPB into Equation

6-9 gives the following expression for the dependence of Rct on LTPB.

R et = n g [( e/>) (A r e n,e/ ) ]- m [(c o a, ,,, ) ( A r eB o,,, as)] [ c ( r a ] ( 1

x(Lapg)-(n+m~p)

The exponential quantity (n+m+p) gives the reaction order dependence on LTPB.

In chemical reactions the reaction order can be given by the coefficients in the balanced

chemical equation. If we assume that the charge transfer reaction is of the form of

Equation (6-4) and that the exponential terms, m, n, and p are given by the coefficients

in Equation 6-4, then a reaction order of -3.5 is predicted and Ret oc (Laps)-3.5 Which is

exactly what was observed.

Thus far in this work, we have referred to the intermediate frequency process as

adsorption rather than "\dissociative adsorption" as often reported in the field. The










exponential dependence of 3.5 can only be predicted if complete adsorbed oxygen

species participate in the charge transfer reaction, therefore, in this work, we refer to

the large magnitude, intermediate frequency process as adsorption. A relationship between

adsorption Rp and volume normalized pore surface area is also established as shown in

Figure 6-10. The data was fit to a power-law relationship resulting in Equation 6-12.


R, =1025(Sv)-1.76 (6-12)


Because impedance spectroscopy is an electrochemical technique, it can not detect the

presence of adsorbed species unless they participate in the electrochemical reaction. If we

assume that the charge transfer reaction, Equation 6-4, is the rate limiting step then the

species generated in the adsorption can only be detected after the charge transfer reaction

occurs, i.e. the rate we detect generation of adsorbed species is limited by the rate of

the charge transfer reaction. Every time adsorption (Equation 6-3) occurs, an 02,ads is

generated. For each 02,ads generated, however, the charge transfer reaction (Equation

6-4) can occur twice. We can conclude that if the reaction is charge transfer limited,

and adsorption does not occur dissociatively, the rate of the adsorption reaction is one

half that of the charge transfer reaction. We expect the reaction order dependence of

adsorption R, on LTPB to be smaller than -3.5 and in fact we find it to be -1.76. It should

be pointed out that the process we are referring to as adsorption is difficult to distinguish

from arrival of adsorbed molecular oxygen to the triple phase boundary by other means. If

surface diffusion is significant at this measurement temperature, oxygen molecules can not

only arrive at the reaction site by adsorption, but also by surface diffusion after adsorption

elsewhere on the cathode surface, coupling adsorption and surface diffusion. Other authors

have reported a dependence of overall polarization resistance on (Lrps)-l when charge

transfer is not the rate limiting step. Our dependence of 1.76 indicates that the reaction

of adsorbed molecules is reduced when charge transfer is rate limiting. Our findings

are not inconsistent with others in that our measurements were carried out at 800 oC
















10 .






-Charge Transfer .


100
0.4 0.65 0.8 12
L per surface area, onm~
TPB

Figure 6-11. Relation of charge transfer and adsorption polarization resistance determined
from the nested equivalent circuit to LTPB (measured in air at 800 op).


while much of the previous work has been carried out at near 1000 oC. At the lower

measurement temperature, the additional polarization components become larger aiding

in deconvolution and the charge transfer reaction may be slowed, effectively changing the

rate limiting step. Varying oxygen partial pressure and temperature will have an effect of

changing the rate limiting step and future work will investigate the influence of oxygen

partial pressure and temperature on the determined reaction order.

6.3.3.2 Nested model evaluation

Figure 6-11 shows the dependence of R 1 (charge transfer) and R 2 (adsorption related)

from Figure 6-6 on LTPs. Curve fitting of the data revealed power dependencies.


R1 = 7.43(LapB)-1. (6-13)


R2 = 86.1(LapB)-2.1 64










The power dependence of Roy on LTPB from the nested model, -1.6, is significantly

different from that determined from the series model, -3.5. The power dependence of R2

on LTPB (-2.1) is consistent with the value reported by Fleig [50] for the dependence of

total resistance at 800 oC. This result is not unexpected since R2 has the larger magnitude

of the two processes and makes up the ill I iG~~y of the cathodic impedance.

Previously, we assumed that molecular adsorption led to a chargeless adsorbed species

participating in the charge transfer reaction occurring at the TPB. We now consider the

possibility that oxygen adsorption leads to a negatively charged intermediate which is

one of the many possible reactions proposed by Nowotny et al. [49]. The corresponding

adsorption and charge transfer reactions are expressed in Equations 6-15 and 6-16,

respectively.

02,g S + e' O',ads (6-15)

2O'~,,4, + Vo" + e' O 0 + s (6-16)

In these reactions, the adsorbed species possesses a negative charge. Because of this

(and any lattice distortions associated with adsorption), the individual adsorbed species

are repelled from one another. For this reason, the amount of low energy sites available

may be reduced as compared to an uncharged adsorbed species. The concentration of

adsorption sites may directly influence the rate of the reaction. In a work by Mizusaki

et al. sites (s) were used in a model which relates the rate of the dissociative adsorption

reaction to the current density of the electrode [117]. In the work, the authors were

interested in the total conductivity. In this work, both charge transfer and adsorption

are of interest and so we must consider the individual reactions corresponding to charge

transfer and adsorption independently.

In the previous section, an expression for exchange current density was derived by

assuming that the rate of reaction is directly related to the concentration of all of the

individual species involved in the reaction, including e' and h*. An alternative approach

is used in electrochemistry to link the exchange current to the rate of production of









electronic carriers by the oxidation and reduction reactions. For example, as described in

Equation 6-16, if an O',ads and a Vo" combine to produce an Og, an e' is consumed (or

alternatively, a h* is produced). This approach is based on the Butler-Volmer Equation

(Equation 6-17) which depends on both the forward and reverse reactions.

in = o exp ) exp( )crl, (6-17)
RT RT

In Equation 6-17, R, T, and F have their usual meaning, rl, describes the surface

overpotential, and as, and ac, are the anodic and cathodic apparent transfer coefficients,

respectively. The Butler-Volmer equation describes the dependence of the the current

density (i,) on an applied potential. In EIS, small AC potentials are applied at various

frequencies. In this work, the applied potential oscillates around 0 V and at 0 V, i,

approaches io, the exchange current density. For this reason, it is most appropriate to

consider io in the modeling of the system. The following approach is typically used in

aqueous electrochemistry where oxygen ions are 02- inStead of Of and vacancies and

reaction sites are not usually considered. In the following development, we will use this

formulism, but will try and incorporate the significance of vacancies and reactions sites,

which are important in solid state systems.

An expression for the exchange current density (io), using the formalism of N~i.. l!! ill

can be expressed by the following relation after rearranging terms [118].


io I I r_ = F cirodc).kiC(cteethdjC) (-a)u] (6-18)


In Equation 6-18, a represents the number of charges transferred in the step, F is

Faraday's constant, k, and k, are cathodic and anodic rate constants, ci,anodic and ci,,,thodic

are the concentrations of the anodic and cathodic species, respectively and pi and qi

are the coefficients of the anodic and cathodic species in the charge transfer reaction

(Equation 6-16), respectively. Whether an electron is consumed in the forward reaction

(hole is produced), or an electron is produced in the reverse reaction, net charge is flowing









in the same direction. For this reason, the forward and reverse reactions both add to the

net exchange current.

There is an activation energy associated with this reaction and a different activation

energy is associated with the reverse reaction, i.e. the production of O' s,, and Vo" from

the dissolution of an Og. Even with no applied potential, these reactions may occur due

to the internal energy of the system. If a potential is applied, however, it will affect the

two activation energies in different manners, depending on the direction of the bias and

the particulars of the system. For instance, in (Lao~sSTO.2~l *M -nO3-6, there are about
four times as many M ~', as there- are Mu-'M Because of the availability of Ms0,~ to

change valencies, a bias which favors the production of holes will more efficiently create

current than the reverse. Such an influence can be accounted for by the utilization of

the symmetry factor, P. Upon examination of Equation 6-18, we see that if P = 0, then

the anodic reactants are disregarded in the calculation of the exchange current density.

This situation corresponds to Equation 6-16 proceeding only in the forward direction. A

Value of 0.5 represents both the forward and reverse reactions occurring in unison, an

equilibrium condition.

And so, from Equations 6-18 and 6-16 we have the following relations.

If = 0: io = nlFke[(c.o~~ )oC.s
(6-19)
S= 0.5~ : io = nFk ".kji [ (co;,d )0.25 Cyo* )0.5 COX )0.5 Cs): 0.5

Combining equation 6-19 and equation 6-8, we have expressions for Rev.

1 F>RT 1,

(6-20)
S= 0.5 : Ret = (it(F2 kk )o~s[(co:, d)-0.25 (Cyo)-0.5 (Cgf -0.5 Cs -0.5


The quantity ci describes the concentration of the i'thr species: which is available to

participate in the interfacial reaction. The concentration of the i'thr species: per- unit~ a~rea









available to participate in the interfacial reaction is limited by the amount of LTPB in that

unit area and is described in Equation 6-10.

Application of Equation 6-10 to Equation 6-20 gives a direct relationship between

Roy and LTPs. So we have: if P = 0,




iRet = x.5

R~t=~(A(nF) (nF)2 ke


CO ,as Arno; ) [c.Aps,.) (ap)-






COGendenc (A ReTno )TP -0.25 ete Cy.(Ap,.)] nd-o..s [hs co nsstn (Artho thos[c(es~)-





observed trend of the nested data, Roy oc(LapB)-1.6

The adsorption reaction which produces the intermediate species 02~,ad, iS expressed

in Equation 6-15. Application of the exchange current equation (Equation 6-18) to the
adsorption reaction gives the following relations.



Ifp = 0 : io = nFke[(coz,,)(c,)]
(6-23)
= 0.5 : i = 11F (keks~)o. [(coa,,Oj(,)o.5 os(cog, :d)o.s

Applying these exchange current densities and the concentration equation (Equation

6-10) to the charge transfer equation (Equation 6-8) gives the following relationships

between R, and LTPB for the adsorption reaction given in Equation 6-15.












1L ~- n)RT k,1
Aine (F)2 ke(6-24)

([co,,,(Airn~o,,,)] [cr(AiRrpss)] x (LTPR)-2

if 4 = 0.5,

1 RT 1 0.
Ras=Aint (nF)2 kek,)'

[c,,(An ~ [,)]o [c(ATPp,s)]-l co,(Areds )-~ (6-25)

x (LapB)-1.s

Fr-om data deconvolution using the nested equivalent circuit, a power dependence

of adsorption polarization resistance on LypB)-2.1 WaS observed. This dependence was

identical to the power dependence reported by Fleig for total cathodic resistance [50]. The

observed power dependence, -2.1, most closely matches the P = 0 case which predicts a

power dependence of -2. For many chemical reactions, P has a value close to 0.5. For this

p, a power dependence of -1.5 is predicted for the adsorption related process. If P = 0, the

dissociative adsorption reaction described in Equation 6-15 is not in equilibrium, which is

consistent with the idea that dissociative adsorption is the rate limiting step as reported

by others [119].

6.3.3.3 Nested relation to pore surface area

As described previously by Fleig, total cathodic polarization resistance (and thus

the resistance of the adsorption related process) scales inversely with cathode/electrolyte

contact area and linearly with cathode thickness [50]. This tendency was explained by

Fleig by considering bulk transport of ionic species through the cathode. For this bulk

transport to occur, adsorption must occur on either the entire pore surface area or an

active region of the pore surface area which is within some critical distance, 6 of the

cathode/electrolyte interface. In either case, the area on which adsorption can occur is

directly proportional to the volume normalized pore surface area, Sv. Alternatively, if


If p = 0,










surface diffusion rather than bulk diffusion dominates, adsorption may still occur either

over the entire surface area or some portion of this area within a distance, 5, of the

cathode/electrolyte interface. In each of these scenarios, the area on which adsorption can

occur will be limited by the surface area per unit volume (Sv).

For simplicity, we will treat only one of the possibilities listed in the previous

paragraph here, however, adsorption for each of the scenarios should be limited by Sv.

For the moment, assume adsorption occurs according to Equation 6-15 over the entire

surface area of the cathode. In addition, assume that surface diffusion of adsorbed

intermediates is not a limiting factor. This assumption is reasonable, as Warburg behavior

was not seen in the various impedance profiles. An exchange current exists based on the

generation of the charged intermediates OG,ads, which will participate in the charge transfer

reaction at the TPB after diffusing to an active location. There is a resistance to this

electrochemical reaction which can be expressed as Reads The exchange current density

(io,ads) is not directly dependent on LTPB since adsorption is not limited to the TPB, like

charge transfer, but may occur over the entire surface area. The total exchange current,

however, is limited to the total pore surface area per unit volume. The resistance to the

adsorption reaction described in Equation 6-15, can be expressed according to Equation

6-8. Previously, for charge transfer, Aine was the planar geometric area of the cathode.

For adsorption, Aine must be replaced by Aint,pore, which represents the pore/cathode

interfacial area as described in Equation 6-26.


Aint,pore = SV x Aint x tcathose (6-26)


In Equation 6-26, Sv represents the surface area per unit volume, Aint is equal to

64 mm2, and the cathode thickness (teethode) is equal to about 20 pm. Substitution of

Equation 6-26 into Equation 6-8gives the following relation.

1 RT\ 1
Ras=SV Aint immtode n1F) o,ads 67










'10"


1I0'

Surface Area per Volume, pmn

Figure 6-12. Relation of adsorption polarization resistance determined from a nested
model to surface area per unit volume (measured in air at 800 oC). Red line
represents the actual fit and dashed line represent a power dependence of -1.

The exchange current density, i~,ads now represents the formation of the charged

intermediate, OG,ads as adsorption occurs. Figure 6-12 d~;-1i ph the dependence of Rads

on Sy. The fitting in the figure reveals Reas oc (SV)-1.3, Which is close to a power

dependence of -1, as illustrated by the dashed line in the figure.
Observation of the trend lines in Figure 6-12 reveals that all of the data points except

two lie along the dashed line. If the dependence of adsorption polarization resistance does

indeed show a dependence on pore surface area to the -1 power, then the same process

would show a dependence on LTPB to the -2 power, if the surface area is proportional to

the triple phase boundary length squared. This is a reasonable assumption if the particles

are relatively uniform in size and spherical in shape. The power dependence observed in

Equation 6-14 and Figure 6-11 can be explained using both models. Further research is

required to determine which is valid.










6.4 Conclusion

We have evaluated the effects of sinteringf temperature on both the electrochemical

and microstructural characteristics of LSM on YSZ symmetric cells. A FIB/SEM system

was used to analyze the microstructure. :3-D images were used to determine the tortuosity

at select sintering temperatures, while evenly spaced 2-D images were used for the

evaluation of triple phase boundary length, volume normalized pore surface area, and

porosity. Impedance data from LSM on YSZ symmetric samples measured at 800 oC was

fitted to two commonly used models, generating different results.

U~se of a series Voigt element model led to a power dependence of -:3.5 for charge

transfer resistance on LTPB and a dependence of -1.75 for adsorption polarization

resistance on volume normalized surface area. The exponential dependence of -:3.5

was predicted by application of principles of reaction kinetics to a charge transfer step

involving uncharged adsorbed molecular oxygen (O2,ads). In this model, we assumed

that the coefficients of the species in the charge transfer reaction determines the power

dependence of the respective species in the current exchange reaction, the concentration

of the species able to participate in the cathodic reaction are linearly dependent on the

amount of triple phase boundary length per unit area, and that the reverse reactions are

negligible.

Comparison of electrochemical parameters from the nested model to the microstructural

data revealed a dependence of -1.6 and -2.1 for charge transfer and adsorption on LTPB,

respectively. For these processes, power dependence of -1.5 and -2, respectively, were

predicted by assuming that the adsorbed intermediate is of the form Oads, the exchange

current can he expressed by Equation 6-18, the concentration of the species able to

participate in the cathodic reaction are linearly dependent on the amount of triple phase

boundary length per unit area, and that the value of /3 is equal to zero.

Since adsorption polarization resistance makes up the inl I iG~~y of the total cathodic

resistance, this individual process can he compared to the results of others who reported










only total cathodic resistance or conductivity. Our results were consistent with those of

Fleig [50] whose analysis was performed on the same material at the same measurement

temperature. Fleig concluded that since total cathodic resistance is proportional to

(Laps)-2, and scales linearly with cathode thickness, bulk conductivity through the LSM

is significant. We have proposed an alternate explanation for LSM which does not depend

on bulk ionic diffusion through LSM which has a low ionic conductivity at 800 oC'. Other

authors have used higher measurement temperatures and produced results inconsistent

with ours; however, few data points were used to demonstrate a relationship between

resistance and LTPs. Additionally, it is reported that a change from Warbug behavior to

non-Warburg behavior occurs at around 800 oC' indicating that the rate limiting step may

undergo a transition in this temperature regime [11:3] The works of K~uznecov et al. (LSM,

950 oC') and Mizusaki et al. (LC'jLl 1000 oC) were performed at higher temperatures were

faster reaction kinetics at the TPB and higher ionic conductivity in LSM are expected

[10:3].

Polarization resistance of the adsorption related process was observed to have a

power dependence of -1.3 on volume normalized pore surface area. A dependence of -1 is

predicted by both models, assuming uniform geometry of particles as previously discussed.

It was demonstrated that both interpretations of the adsorption data front the nested

model will predict the observed dependence of Rp on LTPB. For this reason, future work

is necessary to determine the correct model.









CHAPTER 7
ELECTROCHEMICAL PROCESSES IN LANTHANUM STRONTIUM COBALT IRON
OXIDE

7. 1 Introduction

Thus far, this work has focused on the cathodic reaction of LSM on YSZ. The

cahtode LSM is a purely electronic conductor and has an operating temperature from

800 to 1000 oC. Because high operating temperatures increase the $/kW system cost

of SOFCs, interest has shifted to other cathodes. Composite cathodes and mixed ionic

electronic conductors (illl;Cs) have shown promise for the intermediate temperature

range. The cathode LSCF (Lao. STO.2 00.2760o.8 03Ss) is one of the most studied MIECs

and projects an operating temperature significantly less than 800 op.

Although the ionic conductivity of LSCF is less than its electronic conductivity (0.03

vs 2.9 S cm-l at 800 oC) the ionic conductivity is significantly higher than that of LSM

(10-7 S cm- ) [38-40]. The active region for cathodic reduction is no longer restricted to

the TPB. In effect, the cathodic reaction can occur at the cathode/gas /electrolyte TP B,

the cathode/gas interface, or the current collector/gas/cathode interface. The preferred

site of the cathodic reaction depends not only on the ionic and electronic conductivity of

the cathode, but also on the catalytic nature of the AllEC and the current collector, the

thickness of the cathode, the resistance of species transferred across the various interfaces,

and the local oxygen concentration at the prospective reaction site [48]. These reaction

path--li- act in parallel and therefore, the reaction will proceed in whatever manner

minimizes total resistance. Because the electronic conductivity of LSCF is significantly

greater than the ionic conductivity, the preferred reaction pathway is with charge transfer

occurring at or near the cathode/gas/electrolyte TPB where plenty of oxygen (enough

oxygen to efficiently convert the electronic current in the cathode to ionic current in the

electrolyte) is present. If the required ionic current is greater than that which the supply

of oxygen allows (concentration polarization), then the ionic conducting properties of

LSCF become significant. The active reaction area will expand up the surface of the










MIEC to regions where more oxygen is available. As subsequent oxygen molecules are

taken further from the TPB an oxygen gradient is created [51].

For this reason, in oxygen rich atmospheres an impedance profile similar to the

case of our electronic conducting cathode (LSM) is anticipated. Our previous results

showed that for LSM, two electronic processes (charge transfer and adsorption) dominate

the cathodic reaction in air. In low partial pressures of oxygen a third electrochemical

process, related to the bulk diffusion of gaseous oxygen appears at very low frequencies.

For LSCF, in addition to these three processes, an additional arc should be present due to

ionic transport through the MIEC whenever the third arc, associated with concentration

polarization is present.

7.2 Experimental

The samples were prepared and impedance tested in the same manner described

is section 3.2. In this chapter, however, the LSM was replaced with LSCF supplied by

Nextech Materials, Ltd. Sintering was performed for one hour at temperatures ranging

from 800 to 1150 oC. Sintering at temperatures above and below 1000 oC was performed

in Lindberg/Blue high and low temperature box furnaces, respectively. Argon and air were

combined to produce a flow rate of 100 seem for pO2S leSS than 0.21 .For pO2S greater

than 0.21 oxygen mixed with argon was used rather than air. High-frequency artifacts

in the data were accounted for as described in section 3.3.2

7.3 Results and Discussion

Figures 7-1(a) and (b) show Nyquist plots measured in air at 700 oC for symmetric

LSCF on YSZ samples sintered at high and low temperatures, respectively. The

corresponding imaginary impedance vs. frequency plot is shown in Figure 7-2. For

sintering temperatures above 900 oC the magnitude of the impedance profile increases

as sintering temperature is increased, as shown in Figure 7-1(a). For temperatures

below 900 oC, the polarization resistance magnitude does not decrease significantly

as sinteringf temperature is reduced. As sinteringf temperature is reduced, the form












1'100 OC





250 300 350 400


-200 ,
700 O;C meas. .
air
-*150 -

--100 v'

-50- ""
**"a --a >050 OC
I~~90 00
0 50 100 150 200
8 Z', Oz
-2.5 ,
700 "C measurement
air


-0.5

0


9 '10 1112


Figure 7-1. Impedance response of lanthanum strontium cobalt iron oxide (LSCF) on YSZ
in air at various sintering temperatures. a) High sintering temperatures. b)
Low sintering temperatures.


103


102





'100


Figure 7-2. Imaginary impedance versus frequency for LSCF measured at
various sinteringf temperatures.


700 oC in air at


~10 0 I ~+. 8 9 0 9*94
10 1 100 101 102 103 10 i 10s
Frequency, Hz










102


"1000 OC sinter
101

a .Model
N o0 Adsorptio) -


1 "Charge TransferP 1
ii010
10 100 101 1 02 103 10 105

Frequency, Hz

Figure 7-3. Application of a series Voigft element based equivalent circuit to LSCF sample
sintered at 1000 oC and measured at 700 oC in air.

of the profile degrades and the shape becomes more depressed as shown in Figure

7-1(b). This depression is typically viewed as a distribution of time constants among
the significant electrochemical processes occurring. At the lowest sinteringf temperature,

800 oC the impedance profile may not be stable due to the proximity of sintering and

testing temperatures. The first well-defined are occurs at 950 oC and above 950 oC the

polarization resistance magnitude increases rapidly with temperature. This provides an
indication that 950 oC is the optimum sintering temperature for LSCF on YSZ. From

Figure 7-2, it is clear that the 1000 oC sample has two frequency peaks, one at around 30

Hz, and one at around 104 Hz. This trend continues as sinteringf temperature is increased
as indicated by the .I-i-isso:-~ it ic nature of the profiles. By 1150 oC both processes are

similar in magnitude and frequency and thus difficult to distinguish.

A series Voigt element model was used to fit the impedance data of the 1000 oC

sintered sample. Figure 7-3 d~;-1i ph the polarization contributions which make up

the impedance profile of the 1000 oC sintered sample, measured at 700 oC in air. Like

in LSM, there is a small magnitude high-frequency process and a larger magnitude











charge transfer
=adsorption
700 oC measurement*
102 -air






10 -


750 850 950 1050 1150
Sintering Temperature, OC

Figure 7-4. Parameters determined from equivalent circuit fittingf for LSCF in air,
measured at 700 oC.

intermediate-frequency process. A similar fitting was performed for each of the profiles

shown in Figure 7-2. At the lowest sinteringf temperatures, the fittingf was complicated

by the depressed nature of the profile, while at the highest sintering temperatures

deconvolution was difficult due to the low relaxation frequencies of some of the processes.

The shape of the profiles and the quality of the fitting, leads us to assume that two

processes contribute to the overall impedance profile when measured in air at 700 oC,

particularly at intermediate sintering temperatures such as 1000 oC.

Figure 7-4 shows the polarization resistances obtained from fittingf each of the

impedance profiles shown in Figure 7-2 with a series Voigft element model. The figure

shows a minimum in charge transfer polarization resistance at 900 oC and a minimum

in adsorption polarization resistance at 850 oC. At sintering temperatures above 900 oC,

adsorption polarization resistance becomes larger in magnitude than charge transfer, while

below 900 oC the trend is less pronounced. It is likely that by 950 oC good inter particular

adhesion between the individual LSCF particles and adhesion between the LSCF and the

YSZ has been achieved. Sintering above 950 oC only degrades the cathode as evidenced










by the increased polarization resistance. It is anticipated that this increased polarization

resistance is accompanied by microstructural and possibly phase changes which will be

verified by FIB/SEM in the future.

The 950 oC sintered sample was impedance tested in a variety of pO2S at 700 oC.

Figures 7-5(a) and (b) display the results of the impedance testing broken down into

high and lower pO2 regimes, respectively. From Figure 7-5(a), we see that there is little

change in the imaginary impedance vs. frequency plot from 58 to 10 oxygen. Evidently,

the mechanism of the cathodic reaction is unchanged by variations in pO2 COnCentratiOn

if oxygen is still quite abundant. There is a slight increase in polarization resistance as

pO2 is dropped from the high value of 58 to 10 .~ oxygen. At 3 .~ oxygen, we begin to

see a deviation from the simple two process profile occurring at higher pO2S. By 0.67

oxygen, we can clearly see the formation of two low frequency process which are not

present above 10 .Deconvolution of the sample measured at 0.09 02 is Shown in

Figure 7-6. In the figure, four cathodic processes are apparent. For LSM, we saw one new

process at low partial pressures of oxygen attributed to concentration polarization and

related to bulk gas diffusion to the reaction site. For LSCF, two low frequency processes

appear simultaneously. It is likely that one process is due to concentration polarization

created by the lack of oxygen molecules available at the reaction zone and the second

process is related to LSCF compensating for this inavailablility of oxygen molecules. The

two high-frequency cathodic processes do not appear to be significantly affected by the

change in oxygen concentration from 3 to 0.32 Below 0.32 the low-frequency

processes continues to become more pronounced, particularly the lowest frequency one

which d~;-1i ph a sharp dependence on pO2. Surprisingly, in this regime, the overall

magnitude of the highest-frequency polarization resistance (charge transfer) decreases

as pO2 decreases. This trend is opposite the effect seen when oxygen is abundant. As

concentration polarization becomes more prominent, the reaction mechanism shifts in such
































95
70







0 2


102

1i0'

"100


102
"1 0


10100 10' 102 103 104 105
Freqluency, Hz


0 OC sinter, 0.32 %~
0 OC measurement 0.09 %~
0i.03 %,'
x 0.003%"C





10100 '10' 102 103 104 105
Frequncy, Hz


102





" 1

"10

1 '


Impedance response at various oxygen partial pressures of 950 oC sintered
LSCF on YSZ measured at 700 oC. a) High oxygen partial pressures. b) Low
oxygen partial pressures.


Figure 7-5.










"102


"1 unr PU a>5 YQ uaul trumuipull alfU~LV
']1 MIEC process V Charge transfer
a Model
'10"




"1 02 1 0- 100 101 102 103 104 105
Frequency, Hz

Figure 7-6. Application of a series Voigft element based equivalent circuit to LSCF sample
sintered at 950 oC and measured at 700 oC at 0.09 02*

a way that minimizes the contribution of the original charge transfer reaction, possibly due
to V" formation.

The impedance profiles of Figures 7-5(a) and (b) were fitted using a four Voigt
element based series model with the results di;11l-ple4 in Table 7-1 and Figures 7-7 through

7-9. As mentioned previously, the bulk diffusion process is not apparent at higher pO2S
and therefore model parameters for the bulk diffusion process begin at lower pO2S. FifurtO

7-7 shows the series resistance contribution. Unlike LSM, LSCF shows an increasing

ohmic resistance (R,) as pO2 is decreased. Fitting the data points to Equation 4-3

gives a dependence of ohmic resistance on pO2 to the 0.054 power. As mentioned in the

background section, the electronic conductivity of LSCF is created by formation of holes
when Sr is incorporated on a La site in the lattice. As pO2 is reduced, an increasing
number of vacancies are formed reducing the concentration of holes, therefore, the ionic

conductivity goes up while the electronic conductivity goes down. This decrease in electric
conductivity of the MIEC causes the trend seen in Figure 7-7.

















pO2 *~~ s c Rt Iads Rion B RD
58.0 9.16 1.87 5.52
41.0 9.50 1.86 5.80
21.0 9.76 1.83 6.27
12.5 10.03 2.06 6.85
3.0 10.74 1.97 7.711 0.71
1.0 10.73 1.06 7.87 1.51 0.53
0.67 11.88 2.12 8.49 2.07 0.73
0.09 12.02 0.92 7. 75 2.44 6.24
0.003 11.95 0.6;6 5.62 3.67 6;9.6;9


Table 7-1.


Polarization resistance values in as for various elementary steps of the cathodic
reaction in lanthanum strontium cobalt iron oxide samples sintered at 950 oC
and measured at 700 oC at various oxygen partial pressures.


--I700 OC mleasurement
950 OC sinter


Ohmic resistance





10 -


n
[li


-y = 8.8627 x^(-0.054206)


R= 0.95155

*I


_1 0


7


10 4


10 "


"1 0


pO atm

Figure 7-7. Ohmic series polarization resistance from model fittingf for LSCF sintered at
950 oC and measured at 700 oC as a function of pO2*










Wang et al reported that for LSCF (La0.6STO.4oOO.8760.203-6) the hole is the 1!! linr~~

carrier for pO2S above 0.03 .~ oxygen at 800 oC [43]. The electroneutrality condition in

LSCF is given in Equation 7-1.


n + [Sr ]l = 2 [Vo"] + p (7-1)


Valency changes among the Fe and Co ions account for n = [Mi,] and p = [M' ], while

[Sr I] is equal to 0.2. Also in [43], a plot of conductivity (measured by 4 point probe) on

pO2 TOVealS a dependence of about 0.094 in the high partial pressure regime at 800 oC.

From the data of Bucher et al. reported a conductivity power dependence of 0.097

at 700 oC can be calculated [120]. Because the hole is the 1!! linr~~ carrier, the power

dependence of h* on pO2 Should also be about 0.097. The power dependence observed,

0.054, close to this value.

For charge transfer, the power dependence observed, 0.077, is also close to the power

dependence of ionic conductivity of LSCF reported by Wang. This value has greater error

due to the very low R value reported. A charge transfer resistance independent of pO2

would indicate that this process occurs in a similar fashion as for purely electronically

conducting LSM. However, Vo" formation would reduce Ret as is observed. The decrease

in Ret at low pO2S alSo may indicate that this process becomes increasingly insignificant as

the ionic pathway through the MIEC bulk becomes favorable.

The intermediate-frequency arc was previously attributed to an adsorption related

process. For LSCF, this process shows two distinct regimes. At partial pressures above

0.32 intermediate-frequency polarization resistance decreases as pO2 inCTreSeS, While

below 0.32 .~ (approaching concentration polarization), the polarization resistance

decreases as pO2 decreases. The corresponding power dependencies are -0.086 and 0.099

for the high and low pO2 regimes, respectively. For high pO2, the reactions is likely

confined to the vicinity of the TPB and therefore, we observe a negative pO2 poweT

dependence like that observed in LSM (-0.15 for adsorption/surface diffusion). In the



























I I I I


7


e
1


10







Ir







0.1


00 OC mneasuirement
150 OC sinter


Cha;rge Transfe~r


-y = 2.1007 x^(0.077022)


R= 01.47197


1i0-4


10-3
pO ,


atm


10


100


700 OC nicasurernent
950 OC sinter










-y = 5.4779 xA(-0.086384)
-y = 15.167 xA(0.0j98958)


adslorptioln


n
rr


R= 0.98226~
R= 0.96549


4L


10


10


pO atrn


"100


Figure 7-8.


High-frequency polarization resistances as a function of p()2 for LSCF on
YSZ sintered at 950 oC and measured at 700 oC. a) C'I Ilge transfer Rp. b)
Adsorption Rp.










low pO2 regime, a positive dependence or Rp on pO2 is Observed. At low pO2S, the bulk

path in the MIEC becomes active, therefore, less surface diffusion of adsorbed species is

required. The positive dependence observed may be a direct indication that adsorption

and surface diffusion are coupled in this system.

Figure 7-9(a) shows the partial pressure dependence of the ionic process. This process

was not seen in LSM at low pO2S, and so so this process must be directly related to

the ionic conductivity of the MIEC. There are three regimes, two of which are visible in

Figure 7-9: 1) pO2 > 3 .~ OXygen, 2) 3 .~ oxygen > pO2 > 0.67 .~ oxygen, and 3) pO2

< 0.67 .oxygen. In the first regime, any resistance associated with this process is two

small to analyze (therefore no data points above 3 .~ oxygen). In this regime activation

polarization dominates and the cathodic reaction is confined to the TPB, like in LSM. In

regimes two and three, the vacancy concentration in the MIEC has increased enough to

make ionic conductivity through the MIEC bulk a competitive pathway. As mentioned

near the end of Section 2.2.3, ionic conductivity in LSCF d~;-1i ph a maximum at around

10-2 atm [42]. Below this concentration, a sharp drop off in ionic conductivity exists

as reported by Wang et al. [43]. In regime 2, the ionic conductivity in the MIEC is at

a maximum and so bulk processes in the MIEC are favorable. In regime 3, the ionic

conductivity in the MIEC decreases significantly, possibly due to defect association. In

addition, the vacancy concentration at the MIEC/gas interface is high, so reaction can

occur as soon as the holes arrive at a possible reaction site. This process is limited by

the flow of h* through the MIEC. Since the hole is the primary carrier in LSCF, the

conductivity dependence of LSCF on pO2 Should match the polarization resistance of

this process. The observed dependency of Rp on pO2, 0.095, is in very close agreement

with the findings of Wang et al. and Bucheret al. from whose data power dependencies of

conductivity on pO2 of 0.094, and 0.097 can be measured, respectively.

In regime two, a very strong dependence on pO2 is Observed. This is explained

by considering that this process step effectively competes with the bulk gas diffusion











700 OC mleasurenient
950 OC sinter



Mr4IEC process




-y = 0.059108 x^(-0.70802) R= 0.99921
-y = 1.3739 xA(-0.09478) R= 0.95726


n
rr


0."1 L
10 5


10-4 10" 10 10
pO atm


190


"102


700 oC measurement
950 OC sinter



Bulk Igas
diffulsion




-y = 0.011409 xA(-0.85802) R= 0.99759


a


"100


10


10 4


1I0 "
pO2, t


"1 0


Figure 7-9. Low-frequency olarization resistance as a function of pO2 for LSCF on YSZ
sintered at 950 oC and measured at 700 oC. a) 1\lEC specific process Rp. b)
oxygen gas diffusion Rp.










related process (Figure 7-9(b)). At all pl)2s, the resistance of this process is less than

the polarization resistance associated with bulk gas diffusion as seen in Figures 7-9(a)

and (b). At low pl)2s, a concentration gradient is formed in the vicinity of the TPB

and to compensate, the bulk path through the AllEC becomes active, reducing the total

resistance. As pl)2 increases, less of a concentration gradient is formed and the bulk gas

diffusion resistance decreases; reaction via the TPB becomes favorable. If no concentration

gradient exists, the reaction should take place completely in the vicinity of the TPB and

since the electronic conductivity is much higher than ionic conductivity in LSCF and

the alternate reaction pathway (through the bulk of the AllEC) is not favorable. The

measured value of -0.86 for the dependence of bulk gas polarization resistance on p()2 is

close to the value of unity seen for the bulk diffusion process in the electronic conducting

cathode. The reason for the diminished value could be related to the fact that true

concentration polarization may not he achieved because oxygen molecules are not strictly

provided from regions in the immediate TPB vicinity due to the contribution of the ionic

process.

The dependence of the various processes on measurement temperature and oxygen

partial pressure is summarized in Table 7-2. Trending of the total resistance measured

by impedance was reported by Murray et al. and is included in the table [58]. The trend

of the total resistance follows the that of the largest Rp of the individual processes. At

low p()2s, Murray reported a discontinuity in total resistance dependence on pl)2 and

speculated that different mechanisms limit the cathode performance in different pl)2

regimes. Our results support that hypothesis.

Figure 7-10 di pl oni~ the effects of sintering temperature on the electrochemical

processes occurring at 700 oC and 0.09 .oxygen. This partial pressure was chosen to put

the cathodic reduction reaction in the the concentration polarization regime (less than

0.67 .~ oxygen) discussed above. Like in Figure 7-4 there is a high and a low sintering

temperature regime and they separate at around 950 oC. Focusing first on sintering



















Process Rp pO, > 0.01 atm pO, < 0.001 atm E, (air) E, (0.09 % Og)
Ohmic -0.054 -0.054 -0.40 -0.38
Charge transfer 0.077 0.077 -1.72 -1.63
Adsorption -0.086 0.098 -1.41 -1.56
MIEC process -0.71 -0.095 n/a -1.50
Bulk gas diffusion -0.85 n/a n/a 0.50
Total Rp Murray -0.91 -0.038 -1.63


Table 7-2.


Properties of the various cathodic processes in lanthanum strontium cobalt
iron oxide. The results of Murray et al. [58] for total resistance are included.
The small pO2 dependencies observed are close to the pO2 dependency of the
hole concentration in LSCF (0.094) reported by Wang et al. [43] and the ionic
conductivity of LSCF at higher pO2S (0.097) reported by Bucher et al. [120].


103
0 ch
O ac


102 1 bu


O

1.. 10





100


800 850


Sinter Temp, OC

Figure 7-10. Parameters from model fitting for LSCF at 0.09 .oxygen, measured at
700 oC as a function of sintering temperature.


900 950 1000 1050 1100 1150 1200










temperatures above 950 oC, we see that there are only two significant electrochemical

processes. It should be pointed out that gas diffusion and ionic conductivity through

the AllEC bulk may still occur, but the magnitude of charge transfer and adsorption

overshadow the other two processes. C'I Ivge transfer and adsorption increase with

sintering temperature at low oxygen concentrations in a manner similar to that seen

at high oxygen concentrations. In the low partial pressure regime, four processes are

apparent. The gas diffusion related process is independent of sinteringf temperature up

to 1000 oC. This may indicate that the microstructue is relatively stable at the sintering

temperatures for which this process could be observed. Formation of tertiary phases may

significantly alter the charge transfer resistance while not significantly affecting bulk gas

diffusion. The AllEC specific process is nearly independent up to 950 oC, but drops at

1000 oC. The formation of a tertiary phase at the cathode electrolyte interface could

effectively block transfer of ionic species from the AllEC to electrolyte, effectively reducing

efficiency of ionic transport in the cathode and making other reaction pathir- .va more

favorable. Alternatively, if connectivity increases with sintering, the path for diffusion of

ionic species may become more direct, thus reducing the resistance of this process. Clearly,

microstructural and compositional analysis is necessary for a more complete an~ lli--- Both

charge transfer and adsorption seem to be relatively constant below 1000 oC, with charge

transfer showing a minimum at 950 oC and adsorption showing a minimum at 900 op.

Minima for charge transfer and adsorption occur at the same sintering temperatures at

higher p()2S aS shown in Figure 7-4.

Figures 7-11(a) and (b) show the activation energies determined by varying the

measurement temperature of the 950 oC sintered sample measured in air and at 0.09

oxygen, respectively. The activation energy for charge transfer is -1.72 and -1.63 eV

in air and at 0.09 .oxygen, respectively, while the activation energy for adsorption is

-1.56 and -1.41 eV in air and at 0.09 .~ oxygen, respectively. Activation energies for the

AllEC specific (-1.50 eV) and bulk diffusion (0.50 eV) were also determined for the low











I I I .

Adsorption, Ea = -1.41 eV -






Charge Transfer, Ea = -"1.73 eV

air
950 oC sinter


I I I I


10


101


1.10


10 :



'10

"1


101


102


1.05


1.11.15
1I000/T, K1


1 .25


io~nic dliffulsion
charge transfer


-+gas diffusion


950 OC sinter
S0.09% O


1 "1.05


T), K1


1 .25


Figure 7-11. Activation energies of various electrochemical processes for 950 oC sintered
LSCF on YSZ. a) Measured in air. b) Measured at 0.09 .~ oxygen.


1.1
(1000 /










oxygen concentration measurement. Activation energy values for charge transfer and

adsorption are higher than those measured for LSM (-0.97 and -1.2 for charge transfer

and adsorption, respectively as reported in Section 4.:3. The positive correlation with

measurement temperature seen for bulk gas diffusion in Figure 7-11(b) may be caused by

an increasing oxygen gradient formed in the vicinity of the TPB since the other processes

occur more efficiently at higher temperatures.

7.4 Conclusion

LSCF behaves similarly to LSM in high p()2s. At pl)2s greater than :3.0 .oxygen

the cathodic reaction is confined to the TPB where charge transfer and adsorption

and/or surface diffusion are the only significant electrochemical processes. In air, the

activation energy of the charge transfer and adsorption related process are -1.72 and

-1.56 eV, respectively. C'I Ivge transfer shows a weak dependence on pl)2 at high oxygen

concentrations while adsorption has a pl)2 power dependence of -0.086.

As pl)2 is reduced, an alternate reaction pathway involving ionic transport through

the bulk of the AllEC becomes more significant. This change is manifested by the

formation of two additional low-frequency arcs in the impedance profile. LSM, on the

other hand, presents only one new are, related to the bulk gas diffusion of oxygen,

in the concentration polarization regime. The second are is a most likely a result of

a surface exchange process at the gas LSCF interface which leads to a pathway for

ionic conductivity through the AllEC bulk. At low pl)2s, this pathway becomes more

favorable because 1) a concentration gradient forms depleting the region near the TPB of

molecular oxygen and 2) an increased number of oxygen vacancies (leading to higher ionic

conductivity) form in the AllEC due to the low p()2 of the ambient gas. At low pl)2s, the

activation energies of charge transfer and adsorption are -1.6:3 and -1.56 eV, respectively,

while for the process specific to the AllEC and bulk gas diffusion, the activation energies

are -1.50 and 0.50 eV, respectively.










As for pO2 dependence under concentration polarization, charge transfer polarization

resistance appears to decrease slightly as partial pressure is decreased, although significant

scatter in the data leads to uncertainty in the trending. A decrease in polarization

resistance due to charge transfer could be explained by the simple fact that a smaller

percentage of the ionic current through the electrolyte is supplied by oxygen ions that

have passed through the TPB compared to the cathodic reaction at higher pO2S. FOT

this reason, there is less total power loss due to charge transfer polarization resistance

and the reduction in polarization resistance is anticipated. The power dependence

of adsorption polarization resistance on pO2, 0.099, can similarly be explained by a

decreasing percentage of the electrolyte current being supplied by adsorbed oxygen

molecules which pass through the TPB. At low pO2S, polarization resistance of the process

associated with surface exchange of oxygen molecules leading to bulk ionic transport

through the MIEC increases as pO2 decreases. Two regimes are observed, possibly

indicating the point at which ionic conductivity in the LSCF no longer limits the supply of

oxygen ions to the MIEC/electrolyte interface from the bulk of the cathode. As for bulk

diffusion, a power dependence of -0.85 is reported for low pO2S.









CHAPTER 8
CONCLUSIONS

The electrochemical properties of the cathodic reaction were investigated by

impedance spectroscopy and other characterization techniques with a goal of identifying

the significant individual processes. In order to accomplish this task, high-frequency

impedance data had to be analyzed. Unfortunately, the quality of this data was

diminished by high-frequency inductive artifacts. The influence of these artifacts on

impedance data was analyzed and it was found that several data points at frequencies

lower than the Z,-axis had to be removed from the raw data if the raw data was to be

used. The quality of the data was improved by fitting the high-frequency portion of the

data (which was shown to be inductive in nature) to Zj = jeoL and subtracting the result

from the raw data over the entire frequency range. Performing this operation increased

the amount of usable data in the high-frequency regime by an order of magnitude allowing

analysis of fast occurring electrochemical processes.

It was found that the in air at high temperatures two processes are most significant,

charge transfer and adsorption. The dependencies of these two processes on measurement

temperature and oxygen partial pressure were investigated in chapter three. It was found

that the charge transfer resistance is smaller in magnitude than the adsorption related

process. In addition to these two processes, a bulk diffusion related process becomes

significant at low partial pressures of oxygen. Polarization resistance activation energies

and time constants are generated from the model parameters and given in Table 4-1.

The cathodic reaction was found to be dramatically influenced by the sinter applied.

In addition to altering the microstructure, over-sintering can also cause the formation

of tertiary phases. The electrochemical reaction was drastically inhibited at the highest

sintering temperatures. At these sintering temperatures, it was shown that lanthanum

zirconate, an insulating lI-; r had formed at the cathode/electrolyte interface.

To separate the effects of microstructure from tertiary phase formation, sintering at

lower (but still high) temperature of non-stoichiometric LSM was performed. SEM/FIB










was used to quantitatively analyze the microstructural changes occurring. By relating

the microstructural parameters to the electrochentical changes, direct relationships were

developed which were predicted by theory.

The study was extended to LSCF in the final chapter. It was found that the cathodic

reaction on LSCF behaves similarly to LSM at high partial pressures of oxygen. In

comparison to LSM, LSCF had larger magnitude activation energies for charge transfer

and the adsorption related process. At lower partial pressures of oxygen, the ionic reaction

pathway becomes significant and an are in the impedance profile not seen in LSM appears.

LSCF samples were sintered at various temperatures and corresponding evolution of the

impedance profile was examined. The data taken in this chapter is to be compared to

microstructural information front the samples which is yet to be extracted.









APPENDIX A
COMPOSITIONAL ANALYSIS USING TUNNELING ELECTRON MICROSCOPY
(TEM)


~p


Figure A-1.


Tunnelling electron microscopy image of LSM on YSZ interface, with Energy
dispersive spectrometry (EDS) profiles inset. For EDS profiles, Blue=
Zirconium, Red = Manganese, Green = Yttrium, Purple = Lanthanum,
Yellow = Strontium. (Courtesy of Mark Clark)


The interface of an LSM/YSZ sample sintered at 1400 oC for 48 hours was analyzed

using energy dispersive spectroscopy (EDS) and tunneling electron microscopy (TEM)

by Dr. Mark Clark. A TEM image with superimposed EDS linescans of the LSM/YSZ

interface is di 1 l-p II in Figure A-1. The linescans indicate that the amount of diffusion

is element specific, interfacial pores is not likely the cause of the abrupt profile seen in

Figure 5-9(b). The abruptness of the manganese profile compared to the lanthanum

profile indicating that lanthanum, but not manganese from the electrode is diffusing

into the YSZ. Like the lanthanum profile, the concentration profile of zirconium also

exhibits a gradual decrease, though this time the concentration gradient is decreasing


LSM ------...










from the electrolyte towards the electrode. The yttrium and strontium EDS intensities are

lower than the other elements and don't provide any conclusive support, although their

concentration dropoff in the transition region appears to be more abrupt like manganese

than gradual like lanthanum and zirconium. These diffusion characteristics are evidence of

a 0.2 micron thick transitional region, which is rich in both lanthanum and zirconium, but

lacking in manganese and the other measured elements. TEM was used to investigate the

crystal structure of the electrolyte, electrode and transitional regions.

Selected-area diffraction patterns (SADPs) of the electrode, Figure A-2(a), and

electrolyte, Figure A-2(b) were attained using a 200 kV accelerating voltage, a camera

length of 20 cm. Comparison of the figures allows a clear distinction between these

regions. The electrolyte di; 1l li-- dIa diffraction pattern indicative of an fcc ( i--r I1 with a

beam direction along the [112] zone axis. The pattern of the electrode was not as trivial

due to the electrode's complex crystal structure. The pattern of the transition region,

Figure A-2(c), was identical to that of the electrolyte, -II_- _t h-r;! that the transitional

region is formed by the diffusion of lanthanum from the LSM region into the YSZ region,

preserving the crystal structure of the YSZ. Another possibility is that the transition

region exhibited a phase change, but the new phase was also of cubic body structure with

a similar lattice constant.
































a b


Figure A-2.


Selected-area diffraction patterns for LSM, YSZ, and the transitional region.
Patterns attained for the [112] beam direction with a 200 kV accelerating
voltage. a) Diffraction pattern for LSM. b) Diffraction pattern for YSZ. c)
Diffraction pattern of the transitional region.









APPENDIX B
FOCUSED ION BEAM/SCANNING ELECTRON MICROSCOPY (FIB/SEM)
ANALYSIS



















Figure B-1. Setup of FIB/SEM indicating alignment of ion and electron beam with
sample.


A dual beam FIB/SEM (FEl Strata DB 2:35) was used to create a three-dimensional

image of the microstructure of the symmetric samples as described below. The symmetric

samples were mounted on a 450 aluminum mount, which was tilted another 70 so that the

face of the symmetric cathode was parallel to the ion beam as shown in Figure B-1.

Before exposure of the sample to the ion and electron beams, the symmetric samples

were sputter coated with platinum to minimize charging and protect the sample. The

dual beam FIB/SEM was used to ablate successive 1 u. ris of specified thickness with

SENT imaging after each ablation. These uniformly spaced 2-D images were then aligned,

producing a :3-D image. Serial milling in the z-direction was conducted with steps of

about 50 nm per slice using a Ga+ ion beam. About :30 images were taken per sample.

After each FIB slice, SENT imaging was performed at a magnification of 12,000X. This

mill-image-mill procedure was repeated from the platinum protective surface coating,

through the cathode and down to the dense YSZ electrolyte. The SENT images were taken

at :380 with respect to the sample face normal resulting in elongation of the raw images in









the x direction with x,,tm; = x,,, co~s(:380). The images had to be edited using Adobe

Photoshop in order to adjust for the projection effects of the images. In this manner, a

three dimensional map of the microstructure was created and later used to quantify the

microstructural parameters of interest.

Porosity (p), volume normalized pore surface area (Sr-), and TPB length values were

calculated at each temperature, while tortuosity (-r) values were calculated from the :3-D

data attained at selected temperatures. Porosity was calculated from the pore area/total

area in each SENT image. The calculation was repeated for all slices in the sample and an

average porosity was attained. These results are plotted in Figures 6-2 and 6-:3.

The pore surface area reported is normalized per unit volume. From [121, 122], the

pore surface area (Sr-) per unit volume can he calculated according to Equation B-1.


Sr- L 2PL (B-1)


In Equation B-1, dS is the surface area element, L" is the unit volume, and PL is the

number of phase changes (gas to solid) per unit length and was counted manually from

each of the SENT images through the bulk of the cathode.

The triple phase boundary length (LTPB) was calculated by application of the

following equation [121, 122].

LaPB = (iT/2)PL (B-2)

For LTPB, PL was counted from the pore/LSM phase changes per unit length in the SENT

images at the LSM/YSZ interface. The units for the calculation of LTPB and Sr- are

p-m l, which is dimensionally accurate for a length normalized per unit surface area and

an area normalized per unit volume.

The tortuosity was the only microstructural parameter that was not attained for

each SENT and averaged making use of the uniformly spaced FIB slices. Tortuosity was

calculated by estimating the length a gas particle must travel as it departs the impinging

gas flow and travels to the electrolyte divided by the straight-line thickness. The method










used for the tortuosity calculation was based on the definition of tortuosity. The length

traveled by a gas particle was calculated by first determining x, y, and z coordinates of

the center of pores in .Il11 Il-ent slices and then tracking the changes in pore center location

from slice to slice. Equation B-3 can he used to estimate the total distance traveled by a

particle (L,) using the Pythagorean Theorem.


L, = :~+ rz2 1+ i2 Cz1-Dj)(B-3)
i=1

In Equation B-3 (r,z, Un, x,z) are the coordinates of the nth point used and there are N

total points determined. Once attained, L, can he divided by the straight-line distance to

give the tortuosity.

The volume normalized pore surface area was calculated as described in Equation

B-1 using three SENT images for each sintering temperature. For each image a line was

taken in three directions for a total of nine measurements per sintering temperature. The

temperature dependence of the average pore surface area and corresponding standard

deviation as a function of temperature is plotted in Figure 6-2(a). For the area normalized

triple phase boundary length, a single line near the interface was taken from each of three

SENT images per sintering temperature. The results were averaged and plotted in Figure

6-2(b).









REFERENCES


[1] M.C. Williams, J. Strakey, W. Sudoval, J. of Power Sources, 159 (2006) 1241.

[2] N.Q. Minh, J. of the American Ceramic Society 76 (1993) 563.

[3] H.-J. Ziock, E.J. Anthony, E.L. Brosha, F.H. Garzon, G.D. Guthrie, A.A. Johnson,
A. K~ramer, K(.S. Lackner, F. Lau, R. Mukundan, N. N ...-- .. T.W. Robinson,
B.Roop, J. Ruby, B.F. Smith, J. Wang, Technical Los Alamos National Laboratory
Publication LA-UR-02-5969, Proceedings of the 28th International Technical
Conference on Coal Utilization & Fuel Systems, Clearwater, FL, 2003.

[4] N.Q. Minh, T.R. A1lan-rlinic J.R. Esopa, J.V. Guiheen, C. Horne, J.J. Van Ackeren,
in: S.C. Singhal, H. Iwahara (Eds.), Proceedings of the 3rd International Symposium
on Solid Oxide Fuel Cells, The Electrochemical Society, Inc., Pennington, NJ, 1993,
p. 801.

[5] M. Yashima, M. K~akihana, M. Yoshimura, Solid State lonics 86/88 (1996) 1131.

[6] F.A. K~roger, H.J. Vink, in: Solid State Physics, 3rd Ed., Academic Press, New York,
NY, 1956, pl.

[7] J.F. Baumard P. Aberlard, in: N. Claussen, M. Ruhle, A.H. Heuer (Eds.), Advances
in Ceramics 12, Science and Technology of Zirconia II, American Ceramic Society,
Columbus, OH, 1984, p. 555.

[8] C.B. Choudhary, H.S. Maiti, E. C. Subbarao, in: E.C. Subbarao (Ed.), Solid
Electrolytes and Their Applications, Plenum Press, NY, 1980, p. 1.

[9] E.C. Subbarao H.S. Maiti, Solid State lonics 11 (1984) 317.

[10] X. Guo, M. Maier, J. Electrochem. Soc. 148 (2001) E121.

[11] J.E. Baurle, J. of Phys. C'I. 11. Solids 30 (1969) 2657.

[12] M. K~leitz, H. Bernard, F. Fernandez, E. Schouler, in: A.H. Heuer, L.W. Hobbs
(Eds.), Advances in Ceramics, vol. 3, Science and Technology of Zirconia. American
Ceramic Society, Columbus, OH, 1981, p. 310.

[13] M.J. Verkerk, B.J. Middelhuis, A. J. Bw .4_I1 Solid State lonics 6 (1982) 159.

[14] E.P. Butler, R.K(. Slotwinski, N. Bonanos, J. Drennan, B.C.H. Steele, in: N.
Claussen, M. Ruhle, A.H. Heuer (Eds.), Advances in Ceramics 12, Science and
Technology of Zirconia II, American Ceramic Society, Columbus, OH (1984) p. 572.

[15] J.H. K~uo, H.U. Anderson, D.M. Sparlin, J. of Solid State C!. ~!!I s-1y 87 (1990) 55.

[16] M. Kertesz., I. Riess, D.S. Tanhauser, R. L .mp .pee F.J. Rohr, J. of Solid State
C!. Ins!-l ry 42 (1982) 125.










[17] T. Hashimoto, N. Ishizawa, N Mizutani, 31. K~ato, J. of Materials Science 2:3 (1988)
1102.

[18] K(. K< lI li- Im. I T. Ishihara, H. Ohta, S. Takeuchi, Y. Esaki E. Inukai, J. of the
Ceramic Society of Japan 97 (1989) 1:324.

[19] A. Haninouche, E.L. Schouler, 31. Henault, Solid State lonics 28-:30 (1988) 1205.

[20] G.V. Subba Rao, B.M. Wanklyn, C.N. R. Rao, J. of Physical C!. Ins-tI ry :32 (1971)
:345.

[21] N.Q. Minh, T. Takahashi, in: Science and Technology of Ceramic Fuel Cells,
Anisterdan1; New York: Elsevier Science (1995).

[22] A. Haninouche, E. Siebbert, A. Janimou, Materials Research Bulletin 24 (1989) :367.

[2:3] H. Taguchi, D. Matsuda, 31. NI I.s ... Tanihata, YIT. Miyamoto, J. American
Ceramic Society 75 (1992) 201.

[24] H. Taguchi, D. Matsuda, J. of Materials Science Letters 14 (1995) 12.

[25] A. Endo, 31. Ihara, Solid State lonics 86-88 (1996) 1195.

[26] H.S. Alaiti, A. C'I I1:1 .I~orty, M.K(. Paria, in: S.C. Singhal, H. Iwahara (Eds.),
Proceedings of the :$rd International Syniposiunt on Solid Oxide Fuel Cells, The
Electrochentical Society, Inc. Pennington, NJ, 1990 p. 190.

[27] A. C'I I1:1 .I~orty, P.S. Devi, H.S. Alaiti, Materials Letters 20 (1994) 6:3.

[28] A. C'I I1:1 .I~orty, P.S. Devi, S. Roy, H.S. Alaiti, J. of Materials Research 9 (1994) 986.

[29] R. Basu, S. Pratihar, Materials Letters :32 (1997) 217.

[:30] L.W. Tai, P.A. Lessing, J. of the American Ceramic Society 74 (1991) 5.

[:31] T. Ishihara, T. K~udo, H. Matsuda, Y. Takita, J. of the American Ceramic Society 77
(1994) 1682.

[:32] B. Gharbage, 31. Henault, T. Pagnier, A. Haninon, Materials Research Bulletin 26
(1991) 1001.

[:33] L.G.J. de Haart, K(.J. de Vries, A.P.M. Carvalho, J.R. Frade, F.M.B. Marques,
Materials Research Bulletin 26 (1991) 507.

[:34] J. Mizusaki, H. Tagawa, K(. Tsuneyoshi, A. Sawata, J. of the Electrochentical Society
1:38 (1991) 1867.

[:35] L.-W. Tai, 31.3. No 1- 11 .11, H.IT. Anderson, D.M. Sparlin, S.R. Sehlin, Solid State
lonics, 76 (1995) 259.









[36] L.-W. Tai, M.M. No 1- 11 .11, H.U. Anderson, D.M. Sparlin, S.R. Sehlin, Solid State
lonics, 76 (1995) 273.

[37] Z. Li, M. Behruzi, L. Fuerst, D. Stover, in: S.C. Singhal, H. Iwahara (Eds.),
SOFC-III, PV 93-4, The Electrochemical Society, Pennington, NJ, 1993, p. 171.

[38] Yasuda, K(. Ogasawara, M. Hishinuma, T. K~awada, M. Dokiya, Solid State lonics
86-88 (1996) 1197.

[39] G.Ch. K~ostogloudis, Ch. Ftikos, Solid State lonics 126 (1999) 143.

[40] Y. Teraoka, T. Nobunga, K(. Okamoto, M. Miura, N. Yamazoe, Solid State lonics 48
(1991) 207.

[41] S. Carter, A. Selcuk, R.J. ChI II. r, J. K< .) ..1 J.A. K~ilner, B.C.H. Steele, Solid State
lonics 53-56 (1992) 597.

[42] M. K~atsuki, S. Wang, M. Dokiya, T. Hashimoto, Solid State lonics 156 (2003) 453.

[43] S. W.'11_ M. Katsuki, M. Dokiya, T. Hashimoto, Solid State lonics 159 (2003) 71.

[44] K(. Tsuneyoshi, K(. Mori, A. Sawata, J. Mizusaki, H. Tagawa, Solid State lonics 35
(1989) 263.

[45] A. Hammouche, E. Siebert, A. Hammou, M. K~leitz, A. Caneiro, J. of the
Electrochemistry Society 138 (1991) 1212.

[46] M. Ostergard, M. Mogensen, Electrochimica Acta 38 (1993) 2015.

[47] Y. Takeda, R. K~anno, M. Noda, Y. Tomita, O. Yamamoto, J. of the Electrochemical
Society 134 (1987) 2656.

[48] M. Liu, J. Electrochem. Soc. 145 (1998) 142.

[49] J. Nowotny, T. Bak, M. K(. Nowotny, anc C. C. Sorrell, Advances in Applied
Ceramics 104 (2005) 154.

[50] J. Fleig, Annu. Rev. Mater. Res. 33 (2003) 361.

[51] S. B. Adler, Solid State lonics 111 (1998) 125.

[52] S. B. Adler, J. Electrochem Soc. 143 (1996) 3554.

[53] X. J. C!. is, K(. A. Kthor, S. H. Chan, J. of Power Sources 123 (2003) 17.

[54] S. P. Jiang, Solid State lonics, 146 (2002) 1.

[55] A.J. Appleby, F.R. Foulkes, Fuel Cell Handbook, Van Nu~-i1I lIa1 Reinhold (1989).










[56] D. Herbstritt, A. Weber,E. Ivers-Tiffee, in: IT. Stinining, S.C. Singhal, (Eds.), Solid
Oxide Fuel Cells VI, PV99-19, The Electrochentical Society Proceedings Series,
Peningfton, NJ, 1999, p. 972.

[57] 31. Ostergard, C. Clausen, C. B I---- r, 3. Alogensen, Electrochinlica Acta 40 (1995)
1971.

[58] E. P. Murray, 31. J. Server, S. A. Barnett, Solid State lonics 148 (2002) 27.

[59] D. Herbstritt, A. K~rugel, A. Weber, E. Ivers-Tiffee, Electrochentical Society
Proceedings 2001-16 (2001) 94:3.

[60] B. Boukanip, Solid State lonics 20 (1986) :31.

[61] Van Dijk, A.J. Burgraaf, Phys. Status Solidi (a) 6:3 (1981) 229.

[62] J. Janinik, J. Alaier, Phys. C'I, is, C'I, is, Phys. :3 (2001) 1668.

[6:3] F. Baumann, J. Fleig, H. Habernmeier, J. Alaier, Solid State lonics 177 (2006) 177.

[64] J.R. Alacdonald, Impedance Spectroscopy, John Wiley and Sons, NY (1987).

[65] L. Ljung, System Identification, Prentice-Hall, Englewood Cliffs (1987).

[66] H. Schichlein, 31. Feuerstein, ECS proceedings 99-19 (1999) 1069.

[67] H. Schichlein, A. Muller, Applied Electrochentistry :32 (2002) 875.

[68] K(. K~leveland, 31. Einarsrud, C.S. Schmidt, S. Shanisili, S. Faaland, K(. Wiik, T.
Grande, J. of the American Ceramic Society 82 (1999) 729.

[69] A. Franklin, H.Bruin, Physical Statistical Solids 75 (198:3) 647.

[70] A. Muller, H. Schichlein, ECS proceedings 99-19 (1999) 925.

[71] D.L. Alisell, R.J. Sheppard, J. of Physics D: Applied Physics 6 (197:3) :379.

[72] 31. Orazent P. Shukla, 31. Menthrino, Electrochinlica Acta 47 (2002) 2027.

[7:3] S. Singhal, K(. K~endall, High Temperature Solid Oxide Fuel Cells: Fundamentals,
Design and Applications, Elsevier Advanced Technology, Oxford, UK(, 200:3 p. 24:3.

[74] S. Adler, C'I, in... I1 Reviews, 104 (2004) 4791.

[75] F. H. van Hueveln, H. J. Bouwnicester, J. Electrochent. Soc., 144 (1997) 1:34.

[76] 31. E. Orazent J. Electroanalyt. C'I. 11. 572 (2004) :317.

[77] G. Fletcher in: J. Duchant (Ed.), Mathematical Methods in Physics, Wm. C. Brown
Coninunications, Dubuque, IA, 1994 p. 448.










[78] P. Agarwal, 31. E. Orazent, L. H. Garcia-Rubio, J. Electrochent Soc., 1:39 (1992)
1917.

[79] P. Agarwal, O. D. Crisalle, 31. E. Orazent, L. H. Garcia-Rubio, J. Electrochent Soc.,
142 (1995) 4149.

[80] P. Agarwal, 31. E. Orazent, L. H. Garcia-Rubio, J. Electrochent Soc., 1:39 (1992)
4159.

[81] J. R. Smith, E. D. Wachsnian, Electrochinlica Acta 51 (2006) 1585.

[82] 31. E. Orazent B. Tribollet, book to be published., 2007, Chapter 19.

[8:3] J. R. Smith, A. C'I, i.~ D. Gostovic, D. Hickey, D. K~undinger, K(. L. Duncan,
R. T. Dehoff, K(. S. Jones, E. D. Wachsnian, Solid State lonics, to be published,
(2007).

[84] X. J. C'I. in~ K. A. Kthor, S. H. Chan, Solid State lonics, 167 (2004) :379.

[85] J.-D. K~in et. al., Solid State lonics, 14:3 (2001) :379.

[86] C. Yang, W. Wei, Ceramic Engineering and Science Pro. 2:3 (2002) 7:33.

[87] C. Brugoni, U. Ducati, 31. Scagliotti, Solid State lonics 76 (1995) 177.

[88] H. Tainiatsu, K(. Wada, H. K~aneko, J. of the American Ceramic Society 75 (1992)
40.

[89] K(. Wiik, C.R. Schmidt, S Faaland. S. Shanisili, M.-A. Einarsrud, T. Grande, J. of
the American Ceramic Society 82 (1999) 721.

[90] S.K(. Lau, S.C. Singhal, 1985 Fuel Cell Seminar Abstracts, (1985) p. 107.

[91] H. Tagawa, N. Sakai, T. K~awada, 31. Dokiya, Solid State lonics 40/41 (1990) :398.

[92] J.A.M. van Roosnialen, E.J.P. Cordfunke, Solid State lonics 52 (1992) :30:3.

[9:3] G. Stochinol, E. Syskakis, A. Naountidis, J. of the American Ceramic Society 78
(1995) 929.

[94] H.Y. Lee, S.M. Oh, Solid State lonics 90 (1996) 1:33.

[95] Y.C. Hsiao, J.R. Selman, Solid State lonics 98 (1997) :33.

[96] T. K~enjo, 31. Nishiya, Solid State lonics 57 (1992) 295.

[97] J. A. Labrincha, J.R. Frade, F.R.M. Marques, J. of Materials Science 28 (199:3) :3809.

[98] G. Chiodelli, 31. Scagliotti, Solid State lonics 7:3 (1994) 265.

[99] H. Yokokawa, N. Sakai, T. K~awada, 31. Dokiya, Denki K~agaku 57 (1989) 821.










[100] H. Tainiatsu, H. K~aneko, K(.Wada, J. F, in: B.V.R. C'le o.--1.1, Q. liu, L. C'I. in
(Eds.), Recent Advances in Fast lon Conducting Materials and Devices, World
Scientific, Singapore, Republic of Singapore, 1990, p. 417.

[101] J. Mizusaki, A. Tagawa, K(. Tsuneyoshi, A. Sawata, J. Electrochent. Soc. 1:38 (1991)
1867.

[102] 31. K~uznecov, P. Ostchik, K(. Eichler, W. Schaffrath, Ber. Bunsenges. Phys. Chent.
102 (1998) 1410.

[10:3] 31. K~uznecov, P. Ostchik, P. Obenaus, K(. Eichler, W. Schaffrath, Solid State lonics
157 (200:3) :371.

[104] V. Brichzin, J. Fleig, H.-U. Habernmeier, G. Cristiani, J. Alaier, Solid State lonics
152-15:3 (2002) 499.

[105] 31. Juhl, S. Prinidahl, C. Alanon, 31. Alogensen, J. of Power Sources 61 1996 17:3.

[106] T. K~enjo AI. Nishiya, Solid State lonics 57 (1992) 295.

[107] N. L. Robertson, J. N. Michaels J. Electrochent. Soc. 1:37 (1990) 129.

[108] D. Herbstritt, A. Weber, E. Ivers-Tiffee, Electrochentical Society Proceedings, 99-19
(1999) 972.

[109] S. P. Jiang, J. G. Love, Y. Ramprakash, Journal of Power Sources 147 (2000) :3195.

[110] J.R. Alacdonald, Impedance Spectroscopy, John Wiley and Sons, New York, NY,
1987, p. 74.

[111] J. R. Smith, A. C'I, i.~ K. L. Duncan, 31. E. Orazent E. D. Wachsnian,
Electrochentical Society Transactions 1 (2006) 24:3.

[112] A. Mitterdorfer, L. J. Gauckler, Solid State lonics 111 (1998) 185.

[11:3 E. P. Murray, T. Tsai, S. Barnett, Solid State lonics 110 285.

[114] F. h. Van Hueveln, H. J. 31. Bouwnicester, F. P. F. Van Berkel, J. of the
Electrochentical Society 144 (1997) 126.

[115] Y.-K(. Lee et. al., J. of Power Sources 115 (200:3) 219.

[116] S. P. Jiang, J. G. Love, Y. Ramprakash, Journal of Power Sources 110 (2002) 201.

[117] J. Mizusaki, K(. Antano, S. Yanmauchi, K(. Fueki, Solid State lonics 22 (1987) :31:3.

[118] J. ?-. i.--n! Ias K(. E. Thomas-Alyea, Electrochentical Systems, John Wiley & Sons, Inc,
Hohoken, NJ, 2004, p. 21:3.

[119] J. Mizusaki, K(. Antano, S. Yanmauchi, K(. Fueki, Solid State lonics 22 (1987) :32:3.










[120] E. Bucher, W. Sitte, G.B. Caraman, V.A. Cl.,~~ I. p,isv, T.V. Aksenova,
M.V. All li-o i., Solid State lonics, 177 (2006) 3109.

[121] S. A. Saltykov, Stereometric Metallography, 2nd ed., Metallurgizdat, Moscow, 1958,
p. 446.

[122] C. S. Smith, L. Guttman, Trans. AIME 19 (1953) 81.









BIOGRAPHICAL SKETCH

Jeremiah Robinson Smith was born at an early age in Long Beach, California to

Paul and Pezzy-- Smith. At the age of three, Jeremiah, his sister Dionne, and his brother

Paul Jr. moved along with Mr. and Mrs. Smith to San Antonio Texas. When he was

eight, the Smiths moved to Maryland. It wasn't long before Jeremiah made new friends

in Maryland, but still longed for Texas. Mrs. Powell's fourth grade class was one of the

toughest academic years of Jeremiah's life. (If his mother had not decided to do some of

Jeremiah's endless homework Jeremiah would never have made it past the fourth grade.)

As the years went by, Jeremiah began to identify with Maryland and eventually no longer

considered himself a Texan although his loyalty to the Dallas Cowboys never waivered.

While taking classes at Glenallan Elementary School (Go Gators!), Jeremiah

discovered that he was a good student. Receiving $2 for an A and $1 for a B was all

the extra motivation Jeremiah needed to become a consistent member of the honor roll.

As Jeremiah matured, he realized that a big part of his academic success was that he grew

up with two bright older siblings who exposed him to new ideas and were ahr-l-w happy to

help him with his homework. Jeremiah was happy to attend John F. K~ennedy, the same

high school attended by his brother and sister. As Jeremiah's high school years began to

come to an end, Jeremiah was sad to see his brother and sister spend less and less time

at home and eventually move out of the house. Jeremiah decided that he too would move

out and placed a premium on scholarship offers that included room and board. This was

one of the key factors that led him to choose University of Maryland, Baltimore County

(UMBC) for schooling over Howard University, where his brother and sister attended

college .

Room and board wasn't the only reason Jeremiah attended UMBC. One of Jeremiah's

regrets from high school was that very few of his high school friends were black, despite

the fact that his high school was extremely diverse. Jeremiah was recruited to UMBC

to join the M.~ i-n choff Scholarship program, a program devoted to developing black Ph.










D.s in the fields of science and engfineeringf. Although Jeremiah didn't understand what a

Ph. D. was at the time, he related to the other recruited students and decided to become

a ?1. i-n choff Scholar. To this day-, some of Jeremiah's best friends are other ? i-, i- hoff

Scholars. At UMBC, Jeremiah us! lj ured in physics partly because he had ah--.--s had an

intrinsic curiosity about the world and partly because physics was viewed as the toughest

us! lj ur. As graduation approached, the coursework became more and more abstract

and therefore less and less interesting to Jeremiah and he realized he didn't want to

pursue graduate studies in physics. Jeremiah discovered the field of materials science and

engineering by chance and elected to attend graduate school at the University of Florida.

Upon arrival at the University of Florida, Jeremiah joined the research group of

Dr. K~evin Jones. Jeremiah was happy to join Dr. Jones' group because the focus of the

research was on silicon technology, an area that had direct industrial importance, unlike

many of the fields in graduate level physics. After about three years of study, Dr. Jones

was promoted to chair of the Department of Materials Science and Engineering and was

forced to cut back on his research. Jeremiah was saddened to learn that the project with

which he was involved was not being renewed. Jeremiah was invited to join the research

group of Dr. Eric Wachsman. He was grateful and excited to accept the invititation as he

would be involved in a project focusing on fuel cell research. This research was appealing

to Jeremiah because not only did it have immediate technological significance (unlike

the abstract physics which bored him), but also had environmental impact. Jeremiah

was undeterred by those who informed him that his project was extremely difficult

and questioned if it was even possible. At several points, Jeremiah became discouraged

including when he realized his goal of graduating in two additional years was unrealistic.

Fortunately, many people encouraged and prI li-- II for Jeremiah during these times and he

never quit trying. Jeremiah graduated with his Ph. D. in 2007.





PAGE 1

1

PAGE 2

2

PAGE 3

3

PAGE 4

Ithankmyfriendsandfamilyforcontinuallysupportingme.Ithankmyclassmatesandcolleaguesforhelpingmeacquirealltheinformationneededtomakethispossible.IthankDr.EricWachsmanandDr.KeithDuncanforprovidingthedirectionneededformywork.IthankDr.KevinJones,Dr.MarkOrazemandDr.JuanNinoformanyhelpfulsuggestions.IalsothanktheUnitedStatesDepartmentofEnergyforfundingunderprojectnumberDE-FC26-02NT41562andDE-AC05-76RL01830. 4

PAGE 5

page ACKNOWLEDGMENTS ................................. 4 LISTOFTABLES ..................................... 7 LISTOFFIGURES .................................... 8 ABSTRACT ........................................ 11 CHAPTER 1INTRODUCTION .................................. 13 2BACKGROUND ................................... 15 2.1SolidOxideFuelCellBasics .......................... 15 2.2MaterialsofInterest .............................. 16 2.2.1YttriaStabilizedZirconiaasanElectrolyte .............. 16 2.2.2LanthanumStrontiumManganiteasaCathode ........... 18 2.2.3LanthanumStrontiumCobaltIronOxideasaCathode ....... 20 2.3TheCathodicReaction ............................. 21 2.4ImpedanceSpectroscopy ............................ 25 2.4.1MeasurementDetails .......................... 25 2.4.2DataAnalysis .............................. 28 3ERRORANALYSIS ................................. 31 3.1Introduction ................................... 31 3.2Experimental .................................. 32 3.3ResultsandDiscussion ............................. 34 3.3.1High-frequencyArtifactsinImpedanceData ............. 34 3.3.2CorrectionofHigh-frequencyArtifactsinImpedanceData ..... 38 3.3.3RepeatabilityofMeasurements ..................... 45 3.4Conclusion .................................... 48 4ELECTROCHEMICALPROCESSIDENTIFICATION .............. 50 4.1Introduction ................................... 50 4.2Experimental .................................. 51 4.3ResultsandDiscussion ............................. 52 4.4Conclusion .................................... 57 5TERTIARYPHASEFORMATION ......................... 59 5.1Introduction ................................... 59 5.2TertiaryPhaseFormation ........................... 59 5.3Experimental .................................. 61 5

PAGE 6

............................. 61 5.4.1ElectrochemicalandMicrostructuralCharacterization ........ 61 5.4.2CompositionalCharacterization .................... 69 5.5Conclusion .................................... 70 6THERELATIONSHIPBETWEENCATHODEMICROSTRUCTUREANDELECTROCHEMICALPERFORMANCE ..................... 73 6.1Introduction ................................... 73 6.2Experimental .................................. 77 6.3ResultsandDiscussion ............................. 78 6.3.1EectofSinteringonMicrostructure ................. 78 6.3.2EectofSinteringonImpedance .................... 81 6.3.3EectofMicrostructureonImpedance ................ 88 6.3.3.1Seriesmodelevaluation ................... 88 6.3.3.2Nestedmodelevaluation ................... 95 6.3.3.3Nestedrelationtoporesurfacearea ............. 100 6.4Conclusion .................................... 103 7EVALUATIONFORLANTHANUMSTRONTIUMCOBALTIRONOXIDE 105 7.1Introduction ................................... 105 7.2Experimental .................................. 106 7.3ResultsandDiscussion ............................. 106 7.4Conclusion .................................... 122 8CONCLUSIONS ................................... 124 APPENDIX AANALYSISUSINGTUNNELINGELECTRONMICROSCOPY(TEM) .... 126 BFOCUSEDIONBEAM/SCANNINGELECTRONMICROSCOPY(FIB/SEM)ANALYSIS ...................................... 129 REFERENCES ....................................... 132 BIOGRAPHICALSKETCH ................................ 139 6

PAGE 7

Table page 3-1Randvalueswiththeirrespectiveerrors(standarddeviation,)forrawdatameasuredat900C. ................................. 38 3-2Randvalueswiththeirrespectiveerrors(R;)forhigh-frequencycorrecteddatameasuredat900C. .............................. 42 3-3Changeinpolarizationresistance,RP,constantphaseelementcoecientsQand,andtimeconstant,fromdeconvolutionofrawandcorrecteddataforadsorption(1)andchargetransfer(2)processes. .................. 45 4-1Selectelementarystepsofthecathodicreactioninsamplessinteredat1100C. 54 7-1Polarizationresistancevaluesinsforvariouselementarystepsofthecathodicreactioninlanthanumstrontiumcobaltironoxidesamplessinteredat950Candmeasuredat700Catvariousoxygenpartialpressures. ........... 113 7-2Propertiesofthevariouscathodicprocessesinlanthanumstrontiumcobaltironoxide. ......................................... 119 7

PAGE 8

Figure page 2-1Possiblereactionpathwaysforaplatinumcathodeonelectrolytesystem. .... 23 3-1Photographofatypicalsample. ........................... 33 3-2Rawimpedanceoflanthanumstrontiummanganite(LSM)measuredat900Cinair. ......................................... 34 3-3Modelt,includingthe95%condenceintervals,oftherawdataforan1100Csinteredsample. .................................... 37 3-4Standarddeviation(r;j)versusfrequencydeterminedfromsixreplicatesofdata. 39 3-5Rawandhigh-frequencycorrecteddataforLSMonyttriastabilizedzirconia(YSZ)measuredat900Cinair. .......................... 40 3-6RealtgeneratedfromimaginaryimpedancedataandKKrelationsforLSMonYSZmeasuredat900Cinair. ......................... 41 3-7Realvariance/imaginaryvarianceforrawandhigh-frequencycorrecteddata. 43 3-8EectivecapacitancecalculatedforLSMsinteredat1100C. .......... 44 3-9ModeledelectrochemicalprocessoccurringinLSMonYSZmeasuredat900Cinair. ......................................... 46 3-10Repetitionsofanimpedancemeasurementtakenat800Cinair. ........ 47 3-11Resistancevaluesforrepetitionsofanimpedancemeasurementtakenat800Cinair. ......................................... 48 4-1Impedancespectrafor1100Csampleatvariousmeasurementtemperaturesinair. ........................................... 51 4-2Deconvolutionofanimaginaryimpedanceversusfrequencyproleintovariousindividualcontributingprocesses. .......................... 52 4-3TemperaturedependenceoftheseparatedcontributionsinLSMonYSZsinteredat1100Cfor1hinair. ............................... 54 4-4ImpedanceresponseofanLSMcathodemeasuredat900CwithpO2(atm)asaparameter. ...................................... 55 4-5Dependenceofcathodicpolarizationresistancesinan1100CsinteredsampleonpO2. ........................................ 56 4-6Impedancedatameasuredatvarioustemperaturesforan1100C,1hsinteredsampleat0.002%O2. ................................ 57 8

PAGE 9

.................... 62 5-2Impedancespectraforassintered(1100C)sampleatvariousmeasurementtemperatures. ..................................... 63 5-3Impedancespectraforsampleafterasubsequent1250C,12hannealmeasuredatvariousmeasurementtemperatures. ....................... 64 5-4Scanningelectronmicroscopy(SEM)imagesofthecathode/electrolyteinterfaceassinteredat1100C. ................................ 65 5-5Scanningelectronmicroscopyimagesofthecathode/electrolyteinterfaceforsamplessinteredat1250C. ............................. 66 5-6Scanningelectronmicroscopyimagesofthecathode/electrolyteinterfaceforsamplessinteredat1400C. ............................. 66 5-7Impedancespectrafor1400C,12hannealedsampleatvariousmeasurementtemperatures. ..................................... 67 5-8High-frequencyarcresistancemeasuredat400Cversusannealtemperatureforvariousanneal(temperature,time)pairs. .................... 68 5-9Energy-dispersiveX-RaySpectroscopy(EDS)linescanofMnKintensityatLSM/YSZinterface. ................................. 69 5-10X-raydiractionofsamplessubjectedtopost-annealsintering. .......... 71 6-1Scanningelectronmicroscopy(SEM)images,createdusingafocusedionbeam/SEM(FIB/SEM),ofLSMonYSZsinteredfor1hatvarioustemperatures. 78 6-2Microstructuralparametersasafunctionofsinteringtemperature. ........ 80 6-3Porosityandtortuosityasafunctionofsinteringtemperature. .......... 81 6-4Nyquistplotsmeasuredat800CforLSMsinteredatvarioustemperaturesinair. ........................................... 82 6-5Imaginaryimpedancevs.frequencyplotmeasuredat800CforLSMsinteredatvarioustemperaturesinair. ............................ 83 6-6Nestedelementequivalentcircuitusedfortting. ................. 84 6-7SeriesVoigtelementequivalentcircuitusedfortting. .............. 85 6-8Deconvolutionofimpedanceprolefrom1200Csinteredsample,measuredat800Cinair,usingbothequivalentcircuitmodels. ................ 87 6-9Temperaturedependenceofpolarizationresistance(RP)inairdeterminedusingbothseriesandnestedequivalentcircuitsmeasuredat800C. .......... 89 9

PAGE 10

............................. 90 6-11RelationofchargetransferandadsorptionpolarizationresistancedeterminedfromthenestedequivalentcircuittoLTPB(measuredinairat800C). ..... 95 6-12Relationofadsorptionpolarizationresistancedeterminedfromanestedmodeltosurfaceareaperunitvolume(measuredinairat800C). ........... 102 7-1Impedanceresponseoflanthanumstrontiumcobaltironoxide(LSCF)onYSZinairatvarioussinteringtemperatures. ...................... 107 7-2ImaginaryimpedanceversusfrequencyforLSCFmeasuredat700Cinairatvarioussinteringtemperatures. ........................... 107 7-3ApplicationofaseriesVoigtelementbasedequivalentcircuittoLSCFsamplesinteredat1000Candmeasuredat700Cinair. ................ 108 7-4ParametersdeterminedfromequivalentcircuitttingforLSCFinair,measuredat700C. ....................................... 109 7-5Impedanceresponseatvariousoxygenpartialpressuresof950CsinteredLSCFonYSZmeasuredat700C. ............................. 111 7-6ApplicationofaseriesVoigtelementbasedequivalentcircuittoLSCFsamplesinteredat950Candmeasuredat700Cat0.09%O2. ............. 112 7-7OhmicseriespolarizationresistancefrommodelttingforLSCFsinteredat950Candmeasuredat700CasafunctionofpO2. ............... 113 7-8High-frequencypolarizationresistancesasafunctionofpO2forLSCFonYSZsinteredat950Candmeasuredat700C. .................... 115 7-9Low-frequencypolarizationresistancesasafunctionofpO2forLSCFonYSZsinteredat950Candmeasuredat700C. .................... 117 7-10ParametersfrommodelttingforLSCFat0.09%oxygen,measuredat700Casafunctionofsinteringtemperature. ....................... 119 7-11Activationenergiesofvariouselectrochemicalprocessesfor950CsinteredLSCFonYSZ. ........................................ 121 A-1TunnellingelectronmicroscopyimageofLSMonYSZinterface,withEnergydispersivespectrometry(EDS)prolesinset. .................... 126 A-2Selected-areadiractionpatternsforLSM,YSZ,andthetransitionalregion. .. 128 B-1SetupofFIB/SEMindicatingalignmentofionandelectronbeamwithsample. 129 10

PAGE 11

Theneedforhigheciencyandlowemissionspowersourceshascreatedsignicantinterestinfuelcells.Solidoxidefuelcellsaredesirablefortheirfuelversatility.ThecathodicreactionisknowntobeoneofthemajorcausesofpowerlossesinSOFCs,buttheexactmannerinwhichthecathodicreactionoccursisnotwellunderstood.Thecathodicreactionwasinvestigatedusingprimarilylanthanumstrontiummanganite(LSM)cathode/yttria-stabilizedzirconia(YSZ)electrolytesymmetriccells,asLSMisoneofthemoststudiedsolidoxidecathodesandthesymmetryofthesamplesimpliesthestudy.Anin-depthinvestigationofthecathodicpropertiesoflanthanumstrontiumcobaltironoxide(LSCF)wasalsoperformed. TheareasofinterestareidenticationoftheindividualprocessesoccurringinthecathodicreactionandunderstandinghowthereactionisinuencedbyexperimentalconditionssuchastemperatureandpO2.ElementarystepsofthecathodicreactioncanbeanalyzedindividuallyusingACelectrochemicalimpedancespectroscopy(EIS).Thischaracterizationtechniquegivesoverallpolarizationimpedanceasafunctionofappliedfrequency.Theoutputspectrawereanalyzedgivinginformationabouteachofthesignicantstepsofthecathodicreaction. Theeectofmicrostructuralandinterfacialchangesonthecathodicreactionwasalsoinvestigated.Thesechangeswereproducedbysinteringatvarioustemperaturesandtimes.Themicrostructuralchangeswereanalyzedbothqualitativelyandquantitatively. 11

PAGE 12

12

PAGE 13

Theturnofthetwentiethcenturymarkedthebeginningofatechnologicalexplosionthathaschangedourworldforever.Fromtheinventionoftheelectriclightbulbtotheautomobiletotheaeroplane,manyscienticadvancesweremadewhichhavevastlyimprovedtheeciencyandconvenienceoflifeinanindustrializednation.Theseadvanceshavenotcomewithoutcost.ElectricalandmechanicaldevicesarerequiredbyNewton'sLawofConservationofEnergytoderivetheirpowerfromsomeexternalsource.Todate,thatsourcehasprimarilybeenfossilfuelsformanyindustries.Astechnologyhasincreased,sohasourdemandforfuel;unfortunately,theamountofusablefossilfuelsavailableisniteandifwedonotincreaseoureciencyofconsumptionanddevelopaninfrastructurecapableofutilizingalternative,renewablesourcesoffuel,fossilfuelsourceswilleventuallysimplyrunout.Oneofthemostpromisingtechnologicaladvanceswiththepotentialtoenhancetheeciencyandversatilitywithwhichfossilfuelsareconsumedisthefuelcell.Inafuelcell,chemicals,whicharecontinuallypumpedin,participateinanoxidation-reductionreactionresultingintheecientproductionofelectricity(38%nowand60%by2020[ 1 ]). Thesolidoxidefuelcell(SOFC)holdsparticularpromise.TheSOFChasbeenshowntobeabletoproducehighqualitypowerfromavarietyoffuelsourcesincludingbutnotlimitedtohydrogen,hydrocarbonsandcarbonmonoxidewithonlyheat,waterandCO2asbyproducts[ 2 ].ResearchersatLosAlamosNationalLaboratoryhaveevenproposedtechniquesforzeroemissioncoalpowerplantsbasedonSOFCtechnology[ 3 ].High-eciencypowergenerationisnotenough,however,andtheprimaryfocusofSOFCresearchisindecreasingthesystemcostofSOFCsfromaround800to400$/kWby2010[ 1 ].Theresearchcommunityhasadoptedatwo-prongedapproachtowardsreductionofthisvalue.FirstisthedevelopmentoflowercostmaterialsforSOFCstackconstruction.Amongthesematerialsaremetalinterconnectswhicharemadefeasiblebyloweroperatingtemperatures.Additionally,asoperationaltemperatureisreduced,lessenergyisrequired 13

PAGE 14

TheelectrochemicaleciencyofSOFCsisoftenlimitedbypolarizationlosseswhichaccompanytheoxidation-reductionreaction.Manyofthesepolarizationmechanismshavenegativeactivationenergies;thus,lossestendtobegreateratlowertemperatures,placinglowerlimitsonusefuloperationaltemperature.Itisacceptedthatpolarizationlossescanbedividedintoohmic,concentration,andactivationlosses.Collectiveunderstandingbeyondthissimpledetailisunsatisfactory.Apowerfultoolforinvestigatingpolarizationlossesassociatedwiththecathodicreactioniselectrochemicalimpedancespectroscopy(EIS).Impedancespectroscopyisacharacterizationtechniquewhichallowsfortheseparationofcontributionstooverallimpedanceintoafrequencydistribution.Eachdiscretelossmechanismisactiveinadierentfrequencyregime.AnimprovedunderstandingofthepolarizationlossesoccurringduringthecathodicreactionwillallowfutureresearcherstodirecttheireortsastheyworktodevelophigherkW/$SOFCs.Improvementofthefundamentalunderstandingofthecathodicreactionusingimpedancespectroscopyisthefocusofthisdissertation. 14

PAGE 15

2e0+1 2O2+VoOo(2{2) Thefuelcellstack,consistingofthecathode,anode,electrolyte,andinterconnect,isconstructedinvariousdesignsincluding,butnotlimitedtoplanar,monolithic,ortubular[ 4 ].Theanodeandcathodeareconnectedbyanelectrolytewhichallowsonlytheowofionsandaninterconnectwhichallowsonlytheowofelectrons.Thecathodemustbeelectronicallyconductingtoallowgeneratedelectronstoreachtheloadandporoustoallowgastoowtothereductionsites.Thecathodemustalsohavesucientcatalyticactivityforthereductionofoxidantatoperatingtemperatures.Theanodemustalsobeporousandelectronicallyconductive.Theanodemustalsopossesssucientcatalyticactivityfortheelectrochemicaloxidationofthefuel,thusminimizingpolarization.Sincethepurposeoftheelectrolyteistotransferionsproducedatthecathodetotheanodeandforceelectronsthroughtheload,anyelectronsowingthroughtheelectrolytewillresultinvoltagelossanddecreaseeciency.Forthisreason,theelectrolytemustpossessminimalelectronicconductivitywhilemaintainingmaximumionicconductivity.Further,anelectrolytemustbeimpermeabletothereactinggases.Asmentionedabove,theinterconnectprovidesapathtotheexternalload.Italsojoinsadjacentfuelcellstooneanotherandprovidesapathwayforthereactinggasestoreachtheelectrodes,while 15

PAGE 16

2.2.1YttriaStabilizedZirconiaasanElectrolyte 2 ].Purezirconia(ZrO2)ischemicallystableinoxidizingandreducingenvironments.Purezirconia,however,exhibitsaphasetransitionfrommonoclinictotetragonalat1170Candachangefromtetragonaltocubicuoriteat2370C[ 5 ].Thesephasetransitionsoccurinthefabricationtemperaturerangeforfuelcelldevicesandareaccompaniedbyvolumechanges,whichareundesirable.Inadditiontothisdrawback,theionicconductivityofpurezirconiaistoolowforthismaterialtobevaluableasanelectrolyte.Bothoftheseproblemscanbeaddressedbydopingwithvariousoxides(CaO;Y2O3;MgO;Sc2O3).Oftheseoxides,yttria(Y2O3)ismostcommonlyusedbecauseofstability,conductivity,andcost[ 2 ].Yttriadopingcanstabilizethecubicuoritestructurefromabovefabricationandoperatingtemperaturestoroomtemperaturewhileincreasingtheoxygenvacancyconcentration.Duringdoping,theY3+ionssubstituteonZr4+cationsitesaccordingtothefollowingreactionwritteninKroger-Vinknotationwhichisdescribedin[ 6 ].Inthisnotation,thechargeofasubstutingspecieswithrespecttothespeciesforwhichitsubstitutes(giveninthesubscript)isindicatedbyaprimeifitisnegativeorabulletifitispositive.Ifthespeciesisneutralwithrespecttothetypicalspeciesforaparticularlatticesite,thesuperscriptisan88x00. Thisincreasedoxygenvacancyconcentrationtranslatesintoanincreasedionicconductivity.Ionicconductivityinzirconiaisdueprimarilytothepresenceofoxygenionvacanciesintheuoritestructure.Undopedzirconiahasarelativelylowconcentrationofthesedefects. 16

PAGE 17

7 ].Itisshownthatthemaximumconductivityisobtainedataround8mol%whichistheminimumdopantconcentrationrequiredtofullystabilizetheuoritestructureofzirconia[ 8 ].Thedecreaseinconductivityathigherdopantconcentrationisduetodefectorderingorvacancyclustering,eectivelyreducingthetotalnumberofactivedefects.At1000,CYSZwith8mol%yttriahasaconductivityof0.1Scm1.ThepropertiesofYSZarereviewedin[ 9 ].TheionicconductivityofYSZ,beingdirectlyproportionaltotheconcentrationandmobilityofions,isathermallyactivatedprocess.(Ea1.0eV[ 10 ]).Forthisreason,SOFCsarelimitedtohighoperatingtemperatures,approaching1000CforYSZ.OthertypesofsolidoxideelectrolytessuchasGadollinia(Gd2O3)dopedceria(CeO2)orGDCandyttriastabilizedbismuthoxide(YSB)haveshownhigherconductivitiesatlowertemperaturesandareofinterestforintermediatetemperature(500-700C)SOFCs.Unfortunately,theseelectrolytesarelessstable,havesomeelectronicconductivityoraresimplynewerandarethereforelessunderstoodthanYSZ.Thepresenceofyttria,whichincreasestheoxygenvacancyconcentration,extendstheacceptableoxygenpartialpressurerangedownto1030atm,whichincludestypicaloperatingconditions[ 2 ]. OneofthemostcommonmethodsforYSZpreparationforSOFCsistapecasting.Intapecasting,veryne,uniformparticlesofyttriaandzirconiaparticlesofthedesiredcompositionaremeasuredout.Thepowderisthenmixedanddispersedinasolutioncontainingsolvents,binders,andplasticisers.Theslurryisthenextrudedintapeswhicharecuttothedesiredsize.Thesesubstratesmustbesinteredtomaximizedensication.ThetotalconductivityofYSZisshowntobedependentonmicrostructureandonthecharacteristicsofgrainboundariesinparticular.ThepresenceofgrainboundariesdecreasethetotalconductivityofYSZ.Severalworkshaveusedimpedancespectroscopy,tostudybulkandgrainboundarycontributionstotheconductivityofYSZ[ 11 { 14 ]. 17

PAGE 18

2 15 16 ].ExtensivestudyofLSMhasresultedinagoodunderstandingofitsproperties[ 17 { 19 ]. Lanthanummanganite(LaMnO3)hasaperovskitestructure.Thelanthanumandmanganeseionsbothhaveavalencyof3+,whiletheoxygenionsvalencyis2-.LaMnO3isorthorhombicatroomtemperatureandshowsanorthorhombic-tetrahedralcrystallographictransitionatabout387C[ 15 20 ].Intrinsicp-typeelectronicconductivityhasbeenobservedinLaMnO3duetocationvacancies.TheroomtemperatureconductivityofLaMnO3is104(Scm1)[ 21 ].Substitutionofstrontiumions,whichhaveavalencyof2+,forlanthanumionscausessomeoftheMnionstoshiftfroma3+toa4+valency.Thissubstitutioncanbeachievedviathefollowingreaction. (1x)LaMnO3+xSrMnO3LaMnO3!(1x)LaLa+xSr0La+(1x)MnMn+xMnMn+3Oo(2{4) ThepresenceofMnMn(Mn4+)ionsenhanceselectronicconductivityviaapolaronconductionmechanism.Aspredictedbythemechanism,LSMconductivityexhibitsanArrheniusdependenceontemperature[ 15 ].At20mol%Sr,LSMpossessesanactivationenergyof0.10eV[ 18 ].TheconductionofLSMincreaseswithtemperatureandSrcontentupto1000Candabout20mol%,respectively[ 18 ].Attemperaturesgreaterthan 18

PAGE 19

18 ].(Metallicconductionbehaviorischaracterizedbyanegativetemperaturedependenceincontrasttosemiconductorconductivitywhichincreaseswithtemperature.)TheSrconcentrationalsoaectsthethermalexpansioncoecientofLSM.ItispossibletotailorthethermalexpansionofLSMtosuitablymatchthatofYSZbymodifyingtheSrconcentration[ 22 ]. At1000CtheelectronicconductivityofLSMshowslittledependenceonoxygenpartialpressureathigheroxygenpartialpressures[ 15 ].Acriticaloxygenpartialpressureexists,belowwhichtheconductivitydecreasesasafunctionofthefourthrootofoxygenpartialpressure.TheabruptdecreaseinconductivityatthecriticaloxygenpartialpressureisbelievedtobeduetothedecompositionoftheLaMnO3phase.Thiscriticaloxygenpartialpressureshiftstohighervalueswhenthetemperatureand/orthestrontiumcontentareincreased[ 2 ]. AvarietyofroutesareusedforLSMfabrication,including,butnotlimitedtosolidstatereaction[ 23 24 ],sol-gelsynthesis[ 24 ],laserablationofdenseLSM[ 25 ],spraypyrolosis,auto-ignition[ 26 { 28 ]andco-precipitation[ 29 ].Insolidstateprocessing,powderscontainingacetatesoroxidesofthedesiredcationsaremixedinthestoichiometricratio.Thepowdersarethenmixedandsubsequentlymilledinasolvent(acetone)solution.Theslurryisthencalcinedtoobtainthedesiredcrystalstructure.Sol-gelsynthesisdiersfromsolidstateprocessinginthatagellingagentisaddedtotheaqueoussolutionbeforecalcination.Inspraypyrolosis,aqueoussolutionsofnitrateswiththedesiredcationsaremixedinthestoichiometricratio.Afueladditiveisthenintroducedtocompletecombustionintometaloxides.Thesolutionisthensprayedontoasurfaceanddried.Calcinationisperformedtoproduceapowderwiththedesiredchemicalstructure.Cathodedepositiontechniquesincludescreenprinting,plasmaspraying[ 30 ],slurrycoating[ 31 ]andothertechniques[ 32 { 34 ].Afterdeposition,asubsequentannealisnecessarytoensureproperadhesionofthecathodetotheelectrolyte. 19

PAGE 20

LikeLSM,LSCFowesitselectronicconductingpropertiestoitsperovskitestructure.Whenlanthanumcobaltoxide(LaCoO3)isdopedwithstrontiumoxide(SrO),theSratomssubstituteonaLasite,creatingaholetocompensateasshown. InLSM,theBsiteion,Mn,changesvalencyfrom3+to4+toaccommodatethechargedierencecreatedontheAsite.The4+oxidationstateisunlikelyforbothCoandFe(2+and3+arefavorable);therefore,mostofthenegativechargecreatedbysubstitutionofthedopantatomsiscompensatedbyvacancyformation. 2O2+1 2Vo(2{6) ThesetworeactionsarerelatedbytheelectroneutralityconditioninLSCF,describedinEquation 2{7 ValencychangesamongtheFeandCoionsaccountforn=[M0M]andp=[MM],while[Sr0La]isequaltotheSrdopantconcentration.TheexcessholescontributetoelectronicconductivityallowingtheLSCFtoachieveanelectronicconductivitysimilartothatofLSM,between200and300Scm1at900C.ThemaximumconductivityofLSCF;howeverisreachedatsignicantlylowertemperaturesdependingontheconcentrationofthecations[ 35 36 ].TheionicconductivityofLSCF,ontheotherhand,isseveral 20

PAGE 21

37 { 40 ].TherelativelylargeionicconductivityofLSCFisduetotheadditionalvacancieswhichincreasetheoxygenionmobility.TheoxygeniondiusioncoecientissignicantlylargerinLSCFascomparedtoLSM(about107at800Cversus1012cm2/sat900C)[ 41 ].AtpO2slessthanaround0.01atm,theoxygeniondiusioncoecientbeginstodecreaseasafunctionofoxygenpartialpressure(possiblyduetodefectassociation)at800C[ 42 ].Asaresult,theionicconductivityhasamaximumataround0.01atmospheresdespitetheincreaseinoxygenvacanciesatlowpO2s[ 43 ]. 22 25 34 44 { 47 ].Theoverallcathodicreactionismadeupofvariousindividualsteps.Theprocessbeginsasoxygenowsthroughtheatmospheretotheinterconnects.Next,oxygendiusespasttheinterconnectstothesurfaceoftheporouscathode.Fromthispointon,dependingonthesystem,includingmaterials,microstructure,atmosphere,andotherexperimentalconditions,multiplepathwaysarepossible.Beforeanoxygenmoleculecanbetransformedintoanionintheelectrolyte,oxygengasmustsomehowpassthroughthecathode.Thiscanbeaccomplishedifindividualgasmoleculesdiusethroughvoidsintheporouscathodetothetriplephaseboundary(TPB),thelocationwherethegasphase,cathode,andelectrolyteconverge.Itshouldbepointedoutthatdependingontheopennessofthecathode(poresize,porosityandtortuosity)aconvectiveowinwhichgasmoleculesowasgroups,Fickiandiusioninwhichmoleculesdiusebyrandomlybouncingupagainstoneanother,orKnudsendiusioninwhichawallofthecathodeismostlikelytocauseachangeindirectionofpropagationcanoccur.So,atleastthreedistinctpossibilitiesexistforthewayinwhichgascandiusethroughaporouscathodetotheTPB,theareawheretheporouscathode,gas,andelectrolytemeet. 21

PAGE 22

OncetheadsorbedoxygenspecieshasreachedtheTPB,achargetransferreactioncanoccurinwhichtheadsorbedspeciesonthecathodeareconvertedintochargecarryingionsintheelectrolyte.Inorderforthisreactiontooccur,electrons(orholes)whicharetransferredthroughtheelectronicallyconductingcathodeandoxygenvacancies(orions)whichhavetraveledthroughtheelectrolytemustreachthereactionsite,i.e.theTPB.Variousworkshaveattemptedtosummarizethepossibilities(seeFigure 2-1 )[ 48 49 ].Thesituationisfurthercomplicatedbyconsideringthatvariouspossibilitiesexistforthestructureandchargeoftheadsorbedspecies. Thisdiscussionhasthusfarconsideredonlypurelyelectronicconductingcathodes.Inmixedionicelectronicconductingcathodes(MIECs)abulkpathwayexistsinadditiontothepossiblepathwaysmentionedabove.Theoxygenmoleculecanadsorbandparticipateinachargetransferreactionatanypointonthesurfaceofthecathode.TheformedionorvacancywillthendiusethroughthebulkofthecathodetotheMIEC/electrolyteinterface.Atthispoint,theadsorbedion/vacancycanbeincorporateddirectlyintotheelectrolytedependingonanyMIEC/electrolyteinterfacialresistance.IftheelectronicconductivityoftheMIECismuchgreaterthantheionicconductivity(asitisinLSCF)andplentyofmolecularoxygenisavailableattheTPB,reactionpathwaysinvolvingtransportofionsthroughtheMIECmaybeofhigherresistancethan 22

PAGE 23

Possiblereactionpathwaysforaplatinumcathodeonelectrolytesystem,illustratedbyNowotnyetal.inreference[ 49 ].PermissiontoreproducewasobtainedfromManeyPublishing,publisheroftheoriginalgure. 23

PAGE 24

Themechanismreporteddependsonthetypeofconductivityobserved(electronicversusmixed),thedensityoftheelectrode,andothermaterialsparameters.AlthoughLSMisgenerallyacceptedasanelectronicconductor,asrecentlyas2003conictingreportswerepublishedconcerningwhetherionicconductivityplaysasignicantroleinconductioninLSM.FleigreportsthatfordensepatternedLSMmicroelectrodestransportofoxideionsinthecathodeistheratedeterminingstep[ 50 ].AsproposedbyS.Adler,evenaverysmallionicconductivitysuchasthatinLSMmaycreateanactivereactionlayernearthecathode/electrolyteinterfaceinwhichionicbehaviorissignicant[ 51 ].Thethicknessofthecathodeexaminedmaysignicantlyinuencewhetherthecathodeisperceivedasapurelyelectronicconductorornot. AnecientcathodeshouldbeaporouselectronicconductororMIEC.Theprimarymethodofstudyingreactionmechanismistoexaminetheeectsofchangesinpolarization,oxygenpartialpressure,andtemperatureontheoccurringelectrochemicalprocesses.InanotherworkbyAdler,theauthorconcludesthatforLSCFat700Coxygenreductionatthegas/MIECinterfaceandsolidstatediusionintheMIECarethemajorcontributingprocesses[ 52 ].AcoupleofworksproposeanoxygenreductionmechanismthatconsistsofthreeratelimitingstepsforanLSMcathode.Thehighfrequencystepisattributedtochargetransferofoxygenionsfromthecathode/electrolyteinterfacetooxygenionvacanciesintheelectrolyte.Theintermediatestepisattributedtothedissociationofadsorbedoxygenmoleculesintoadsorbedoxygenatoms.Thelowfrequencystepisattributedtothediusionofoxideionstotheinterface[ 46 53 ].In 24

PAGE 25

54 ].TheauthorobservedthatforLSMcathodes,surfacedissociativeadsorptionanddiusion,chargetransfer,andoxygenionmigrationintotheelectrolytewerethesignicantreactionstepswithdissociativeadsorptionbeingtheratelimitingstepatlowtemperaturesandoxygenionmigrationlimitingathightemperatures.Further,theauthorfoundthatforLaxSr1xCoO3,dissociativeadsorptionandbulkdiusion(orsurfacediusion)processesaresignicantincathodicreactionwithLSCFcathodes.IncreasedperformancewithLSCFisalsoobservedandisattributedtoitshigheroxygenionconductivityandcatalyticactivity. Forpurelyelectronicconductingelectrodes,theelectrochemicalreactiondrivingfuelcelloperationisrestrictedtotheTPB[ 55 ].BecausemuchofthepowerlossinSOFCsisduetopolarizationlossatthecathode-electrolyteinterface,degradationofthisinterfacehasadeleteriouseectontheperformanceofthecell[ 55 ].IncreasingtheTPBarea,ontheotherhand,resultsinamoreecientSOFC[ 56 ]andmodernSOFCstructuresareengineeredtomaximizethisarea.M.OstergardwasoneofthersttoproposeanddevelopcompositeLSM/YSZcathodeswiththespecicintentionofincreasingtheTPBandthereforedeviceperformance[ 57 ].IncreasingtheTPBhasevenbeenshowntoincreaseelectrochemicalperformancefortheMIECLSCF[ 58 ].AqualityTPBrequireshighporosityintheelectrodeandgoodadhesionbetweentheelectrodeandelectrolyte.Breakdownofthisinterfaceisaprimarycauseofdevicedeterioration.Delaminationofthecathodeisonesourceofdegradationofthisinterfacedegradation[ 59 ].Thereactionbetweenelectrodeandelectrolyteisanothersourceofinterfacedegradationandisthefocusofoneofourstudies. 2.4.1MeasurementDetails 25

PAGE 26

Thevalueofimpedancespectroscopyasacharacterizationtechniqueisthatitproducesevidenceofthetotalpolarizationlossateachfrequencymeasured.Foragivenpolarizationprocess,lossonlyoccursiftheperturbationoccursatalowerfrequencythanthatprocessesrelaxationfrequency.Iftheperturbationoccursatahigherfrequencythantherelaxationfrequency,thesystemwon'thavetimetodissipateanypowerviathatmechanism.Conveniently,wecanlookattheentireimpedancespectrumandseetheindividualprocessescontributingandtheirrespectivesignicantfrequencyranges. 26

PAGE 27

Eachoftheindividualprocessesoccurringhasitsownrealandimaginaryimpedances,andcharacteristicfrequencyassociatedwithit.Thecapacitiveimpedance,ZC,theinductiveimpedance,ZL,andtheohmicresistance,ZR,asafunctionoffrequencyaregivenbythefollowingrelations,respectively. AsingleelectrochemicalprocesswilltraceasemicircleintheNyquistplotwiththesemicircle'sdiameterlyingonthepositivex-axis.Thisbehaviorcanbemodeledbyaseriesresistanceconnectedtoaresistorandcapacitorinparallel(Voigtelement).ThedistancefromtheorigintothebeginningofthesemicirclewillhaveamagnitudeofRS(theresistanceoftheseriesresistor).Thiscontributiontoimpedanceisduetoeitherohmicresistancesoranyprocessthatoccursatafrequencyrangemuchhigherthanthemeasuredrange.Thediameterofthesemicirclewillhaveamagnitudeequaltotheparallelresistance(Rparallel).Thepeakofthesemicirclewilloccuratthecharacteristicfrequency,!o.ThemagnitudeoftheheightofthecircleisequaltoZC,whichisrelatedtotheparallelcapacitance(CP)bytheequationabove.TheseconstantsareadditionallyrelatedviatheR-Ctimeconstant,. Thetimeconstantandthecharacteristicfrequency(!)areinverselyrelated. Thetimeconstantprovidesinformationonthekineticsofthereaction. 27

PAGE 28

Impedancedataisoftenanalyzedbyttingthespectratoamodel,whichisbasedonaprioriknowledgeofthesystem.Thistypeofmodelisrepresentedbyanequivalentcircuit[ 60 ].ForSOFCapplications,authors[ 11 12 ]haveproposedamodelbasedonthebricklayermodel[ 61 ]whichseparatesthecontributionsoftheelectrolytebulk(intragranular),electrolytegrainboundary(intergranular),andelectrodeeects(chargetransfer).Gasdiusionandionmigrationareincludedamongotherphenomenathatmaycontributetotheunresolvedspectra.JamnikandMaierderivedageneralcircuitforacellwithaMIEC[ 62 ].WhenappliedtothespecialcaseofaSOFC,theirmodelresultsinanequivalentcircuitnearlyidenticaltooneusedbyseveralauthorswhichiscomposedofadoublelayercapacitanceinparallelwithaseriesconnectionofaresistorandaVoigtelement[ 63 ].Thisequivalentcircuitisthemostcommonlyusedofthenestedcircuits. Aninherentlimitationwiththismethodisthatthetattainedisdependentonknowledgeoftheprocessescontributingtothespectrum.Twoimportantconsequences 28

PAGE 29

64 ]. Asecondcommonlyusedtechniqueminimizesambiguityattheexpenseofcondenceinthemodel.ThistechniquematchesallsemicirclespresentintheNyquistplotwithindividualVoigtelements(resistorandcapacitorplacedinparallel)connectedinseries.Themajoradvantageofthistypeofanalysisisthatcomparisonofdataamongdierentresearchgroupswithdierentassumptionsaboutthemechanismofreactionisfacilitated.Theprimaryawofthistechniqueisthatassignmentofparametersbasedonthemodelisdicultsincethemodelwasnotconstructedwithspecicparametersinmind.Inordertoassignparameterstothemodel,theEISmeasurementmustberepeatedasmeasurementconditionsand/orsamplecharacteristicsarevaried.Theresultingimpedancespectraarethenmodeledandchangesintheattainedmodelparametersarethencomparedwiththeory,leadingtoassignmentofidentitiesoftheindividualparameters. Anothermethodsometimescalledsystemidentication[ 65 ]isbasedonthereasonableassumptionthattheinput/outputbehaviorisdependentonlyonthecelltobetested[ 66 ].Insystemidentication,insteadofmodelingfromphysicallaws,amathematicalmodelisbuiltbasedsolelyontheexperimentaldata.Systemidenticationconsistsofthreesteps:pre-identication,modelestimation,andmodelvalidation.Inpre-identicationthedataismanipulatedintoaformatsuitingthechosenmodel.Modelestimationinvolvesdeterminationofthevariousparametersofthemodelandmodelvalidationteststhesuitabilityofthemodel[ 65 ].Theoutputimpedancespectrumisanalyzedwhileinput, 29

PAGE 30

Mathematicaltechniqueshavebeendevelopedtoaidindataanalysis,andinparticulartoincreasetheresolutionofthemeasureddata[ 67 ].Forverysimplesystemswithclearseparationofelectro-chemicalprocesses,modelparametersmaybeobtainedwithoutgreatdiculty.Forasfewastwoconjoinedsemicircles,mathematicaltoolshavebeendevelopedwhichaidinattainmentofmodelparameters[ 68 ].SeveralmethodsofdataanalysisbasedonFourieranalysisoftherawdatahavebeenpresented[ 67 69 ].Thesemethodsincreasethefrequencyrangeofthedata,facilitatingseparationoftheindividualactingprocessesthatshowsimilarrelaxationfrequencies.Inthismanner,processescanberesolvedwhicharenotresolvableusingconventionalmethods[ 70 ].Theadvantagesofusingthistypeoftechniqueincludethefollowing:1)timeconstantsmaybeobtainedwithlittleknowledgeaboutthesystem,2)separationofdistributionsthatarenotreadilyseparableinconventionalimpedancedataispossible,3)areducedsensitivitytorandomexperimentalerrorisgained[ 69 ].TheKramers-Kronigrelationsarealsousedtoreduceerrorassociatedwiththedataanalysis.AccordingtotheKramers-Kronigrelations,thesameinformationiscontainedintherealandimaginarycomponentsoftheimpedanceprole;therefore,indataanalysis,onlyoneofthecomponentsneedstobeconsidered[ 71 ].TheKramers-Kronigrelationshavealsobeenusedtoidentifyand/orreduceerrorintheacquireddata[ 72 ].Thistypeofanalysisisthefocusofonechapterinthiswork. 30

PAGE 31

73 ].Unfortunately,considerabledisagreementconcerningthenumberandidentityoftheelementaryelectrochemicalprocessesstepsoccurringinthecathodicreactionstillexists[ 74 ].Thisdisagreementisduetothreemainfactors:themethodofreductionisdependentonthematerialssystem,variationsinsamplepreparationandmeasurementconditionsaecttheoutput,andsubjectivityinanalysisofimpedancedataleadstoinconsistency. Cathodicreductionhasbeenstudiedonseveralmaterialssystems,eachproducingitsownresults.SomeoftheearliestworkintheareahasbeenonplatinumelectrodeSOFCsystems[ 49 ].Becauseplatinumisapurelyelectronicconductor,thecorrespondingcathodicreactionisstronglydependentonthesurfacepropertiesoftheplatinum.Morerecently,mixedionicelectronicconductors(MIECs)havebeenconsideredascathodes.Forthesesystems,thecathodicreductionreactionwouldinvolvesomeionictransportthroughthebulkoftheelectrode[ 54 ].Inshort,thereductionreactionpathwaydependsonthematerialssystemchosen. Thenextfactorleadingtoconfusionisvariationinsamplepreparationandmeasurementconditions.Evenforagivenmaterialssystemandasinglelaboratoryitisimpossibletoproduceidenticalsamplesfortestingfrombatchtobatch.Slightvariationsincathodethickness,cathodesinteringconditions,andsolidelectrolytepropertiesmaybeunavoidable.Obviously,whencomparingresultsfromlaboratorytolaboratory,thesevariationsincrease.Thebiashistoryofthesamplemayalsohavesomeeectonitsmeasuredproperties[ 75 ].Themeasurementconditionsalsoincludethetypeofmeasurementdone,i.e.2,3,and4pointprobeimpedancemeasurement,eachproducingadierentoutputforasinglematerialssystem. 31

PAGE 32

64 ].Additionally,somesignicantprocessesmaybeenvelopedbyotherprocessesanddependingonthechosenmethodofdisplayingthedataandsensitivityoftheequipmentthesehiddenprocessmaybeoverlooked.Theevaluationisfurthercomplicatedbyartifactsintroducedinthemeasurementbythemeasurementsystem.Theseartifactsaretypicallyaccountedforby88nulling00(explainedingreaterdetailbelow)orsimplytruncatingthedatatolimittheimaginaryimpedancetonegativevalues.Littleunderstandingexistsconcerningthevalidityofthedatapointsimmediatelyafterthespectrumcrossestherealimpedanceaxis. Thegoalofthischapteristoanalyzethequalityofimpedancedata,particularlyathighfrequencies.ThisisaccomplishedbydetermininghowthedatadeviatesfromaKramers-Kronigconsistentmodel.Ultimately,amethodisproposedwhichimprovestheconsistencyofthehigh-frequencydata.Thismethodwasusedinchapters6and7ofthisdissertation. 3-1 .Adryingstepwasperformedafterthescreen-printingofeachlayerinaFisherIsotempdryingovenat120Cforonehour.Afterdrying,sinteringat1100Cand1hourwasperformedinaLindberg/Bluehightemperatureboxfurnace.Theresultingsymmetricalsampleshadacathodethicknessofabout20microns. 32

PAGE 33

Photographofatypicalsample. ThesamplesweremountedinaquartztubeinsideaBarnstead/Thermolynefurnacetocontrolmeasurementtemperature.Thequartztubeconsistedofaninletandoutletforgasow,goldleadsrunningthroughaluminarodscoatedwithplatinumforshielding,andapressurecontactsampleholder.Thegoldleadswereconnectedbyplatinumwirestoaplatinummesh,whichwasusedasthecurrentcollector.Thepressurecontactholderwasdesignedinawaythatexposestheplatinummeshandadjacentcathodetotheambientgas.Airwasowedoverthesamplesat40sccm.ASolartron1260impedancegainanalyzerwasusedtomeasurethefrequencyresponseofthepreparedsamples. Electrochemicalimpedancespectroscopy(EIS)usingaSolartron1260impedancegainanalyzerwasperformedinordertomeasurethefrequencyresponseofthepreparedsamples.A50mVACvoltagewasappliedandtheinducedcurrentwasmeasuredtoproducetheimpedancespectra.Measurementwasmadeviaa2-pointconnectiontotheSolartron.Auto-integrationwasusedunder88I,long00measurementconditionswithanintegrationtimeof60seconds.I,longisaZplotoptioninwhichthecurrentismeasuredfornoiseandanattemptismadetogetconsistencyinthemeasurementswithamaximumstandarddeviationof1%whenpossible.Theactivefrequencyrangewas 33

PAGE 34

Rawimpedanceoflanthanumstrontiummanganite(LSM)measuredat900Cinair.a)Complexplaneplot.b)Imaginaryimpedancevs.frequencyplot. 1:01023:2107Hz.SMART,ZplotandZviewwereusedtoacquireanddisplaytheimpedancedata. 3.3.1High-frequencyArtifactsinImpedanceData 3-2 .Thetraditionalcomplexplaneplotshowsasinglelargearcunderthesetestingconditions.TheZjversusfrequencyplotshowninFigure 3-2 (b)isbrokenintotworegions.Inthegure,-Zjisplottedforthecapacitiveportionofthedata(lowerfrequencies)and+Zjisplottedfortheinductiveregion(higherfrequencies).DisplayingthedatainthisformatdirectlyindicatesthepeakfrequencyandthereforethetimeconstantofthecapacitivearcshowninFigure 3-2 (a).TheslopeofFigure 3-2 (b)inthelow-frequencyregimeisconstantandcloseto1,indicatingthatonlyoneprocessisdominantinthisregionandthataconstantphaseelementmaynotbenecessaryformodelinginthisregion.Asfrequencyincreases,ahigh-frequencyartifactbecomesmoreandmoresignicant.Theslopechangeinthehigherfrequenciesofthe 34

PAGE 35

76 ]. ThemeasurementmodelusesseveralRCVoigtelementsconnectedinseriestoproduceaKramers-Kronig(KK)consistentmodelforthedata.TheKKrelationsarevalidforsystemsthatsatisfyconditionsofcausality,linearity,andstability.TheKKrelationsassertthatiftheseconditionsarevalid,therealandimaginarycomponentsofimpedancedatacontainidenticalinformation,andinfact,therealpartofthedatacanbegeneratedfromtheimaginarypart,andvice-versa.TheKKrelationsareexpressedinEquations 3{1 and 3{2 forafunctionf(x)=u(x)+iv(x). FletcherappliestheKKrelationstoanRCcircuitin[ 77 ].Aninductiveartifactisaneectoftheimperfectmeasurementtechniqueandisnotspecictothepropertiesofthesampletobetested,i.e.itisnon-causal.IfthemagnitudeofthisartifactbecomessignicantthentheimaginarypartofthedatawillnolongerbelinkedtotherealpartandthedataisnolongerKKconsistent.Forthisreason,aseriesofVoigtelementswillnotproducethearcshowninFigure 3-2 (a). Whendealingwithimpedancedata,thepresenceofahigh-frequencyartifactintherawdataisnotunusual.Therearethreecommonwaystodealwiththisphenomena.Therstisignoringthehigh-frequencyportionofthedata,understandingthatitisnotuseful.Thesecondistruncatingthedataattheintersectionwiththerealaxis.Thethirdistheuseof\nulling",atechniqueinwhichanimpedancerunismadeunderopenandshortcircuitconditionsandtheresultsaresubtractedfromtherawimpedancedata.If 35

PAGE 36

Themeasurementmodelconcept[ 78 { 80 ]wasdevelopedfortheidenticationoftheerror-structureoffrequency-domainmeasurements.Orazemetal.extendedthemodeltogeneralizedidenticationofdistributionsoftimeconstantsandultimately,asystematicapproachwasdevelopedforanalysisoferrorstructureinimpedancedata[ 72 76 ].Oneoftheprimaryreasonsforthedevelopmentofthemodelwasintheidenticationofthefrequencyrangethatwasunaectedbyinstrumentalartifactsofnon-stationarybehavior.Inthiswork,themeasurementmodelisusedtodeterminetheextentofthecorruptionoftherawdatabythecommonlyobservedhigh-frequencyartifactdisplayedinFigure 3-2 (b).Becausethemeasurementmodeltechniquereliesonacomplexnonlinearleast-squares(CNLS)techniquefortheregressionofimpedanceproles,(theregressionisbasedonboththerealandimaginarypartsofthedata,simultaneously)thesolutionsattainedmustbeconsistentwiththeKramers-Kronigrelations.Inshort,theactualdataisttothefollowingrelationshipandtheparametersK,RoRk,andkarereturnedalongwithcorrespondingstandarddeviations. WhereRodescribestheohmicresistanceandKindicatesthenumberofVoigtelementsinthemodelwhileRkandkarethepolarizationresistanceandtimeconstantofthekthVoigtelement,respectively.IfthedataistotallyinconsistentwiththeKKrelations,themodelwillfailtoconvergetoasolutionandsomedatapointsmayhavetoberemoved.Becauseofthenon-causalartifactshowninFigure 3-2 (b),thehigh-frequencyportionof 36

PAGE 37

Modelt,includingthe95%condenceintervals,oftherawdataforan1100Csinteredsample.a)Realpart.(b)Imaginarypart. thedatawasnotKKconsistent,themeasurementmodelwouldnotconvergetoasolutioniftheentiredatasetwasused.Toalleviatethisproblem,weappliedthecommonlyusedtechniqueoftruncatingthedataattheZraxisandanattemptwasmadetottheremainingdatausingtheseriesVoigtelementmodel.Aconvergentsolution;however,wasnotreachedindicatingthatthedatapointsthatwerekeptstillhadsignicantcontributionfromtheartifact.Afterremovingmorehigh-frequencydatapoints,asolutionwasreached,butthemodelshowedsignicantdeviationfromtheactualdataatthehighestfrequencieskept,leadingtoincreasederrorinthemodel.Aftermorehigh-frequencydatapointswereremoved,thehigh-frequencydeviationbetweenmodelanddatadisappearedandaqualitytwasattained.ThesolutionhadsixelementswiththeconstantsgiveninTable 3-1 .Figure 3-3 showsZrandZjvsfrequencyplotsforthetting.Thegureshowsthedatawiththeusedpointsassolidcircleslocatedbetweentheverticalbars,truncatedpointsashollowcirclesoutsidetheverticalbars,themodelasasolidlineandthe95%condenceintervalsasdashedlines.TheasymmetryshowninFigure 3-3 (b)isduetotheinductiveartifactdisplayedinFigure 3-2 (b).TheinsetgureinFigure 3-2 showstheextenttowhichdatahadtoberemovedtoproduceaquality,KKconsistentsolution.DataneartheZraxisisthereforeunreliable. 37

PAGE 38

Randvalueswiththeirrespectiveerrors(standarddeviation,)forrawdatameasuredat1100C. Process# R() Element1 0.883 0.040 0.064 0.002Element2 2.02 0.19 0.24 0.01Element3 5.15 0.11 0.59 0.02Element4 1.01 0.15 1.52 0.11Element5 0.111 0.012 8.7 1.1Element6 0.040 0.004 70 11 ConstantssuchasthoseshowninTable 3-1 wereobtainedforeachofthe6repeatsoftheimpedancespectroscopymeasurement.Theconstantsobtainedwerecomparedandtherealandimaginarystandarddeviation(r;j)inthemodelwascalculatedandisdisplayedinFigure 3-4 (a).Figure 3-4 (a)showsthatr;jhasamagnitudeontheorderof103.Additionally,aconsistenttrendinboththerealandtheimaginarystandarddeviationvaluesisobserved;thestandarddeviationshavetheirlargestmagnitudevaluesinthelow-frequencyregimeanddropoasfrequencyisincreased.Thefactthattheerrorhasafrequencydependencebutstillsomerandomnessindicatesthattherearebothstochasticandnon-stochasticerrorspresentinthedata.Thefrequencydependentnon-stochasticerrorsareeitherrelatedtosystematicexperimentalerrorsorduetoimperfectionsinthemodel. Asmentionedpreviously,cuttingothehigh-frequencydataattheZraxisisacommonwayofdealingwithhigh-frequencyartifacts.WeusedtheMeasurementModeltoshowthatsimplychoosingZraxisforacutopointisinsucientforavoidingtheproblemscausedbythecommonhigh-frequencyartifact.Toeectivelyavoidcontributionsfromthehigh-frequencyartifact,datamustbecutowellabovetheZraxistoensurereliabledata. 38

PAGE 39

Standarddeviation(r;j)versusfrequencydeterminedfromsixreplicatesofdata.Thereisnosignicantincreaseintheerrorofthehigh-frequencydata,despiteitsbeingincreasedanorderofmagnitudeversustherawdata.a)Rawandb)High-frequencycorrecteddata. 39

PAGE 40

Rawandhigh-frequencycorrecteddataforLSMonyttriastabilizedzirconia(YSZ)measuredat900Cinair.a)Nyquistplot.b)Imaginaryimpedanceversusfrequencyplot. highestfrequencyportionoftheimaginaryimpedancedataisrstttotherelationshipZj=j2fLexp,wherefisthefrequencyofthemeasurement,andLexpandareconstantstobedeterminedfromthetting.Ifisequaltoone,theexpressionreducestoZj=j!Lexp,anexpressionfortheimpedanceofanidealinductor.Ifisnotequaltoone,thenthehigh-frequencyartifactisnotapureinductance.Thistisvalidinthehigh-frequencyregimebecauseasfrequencygetslarge,thecapacitiveimpedance,whichischaracteristicofthesample,approacheszeroandthemodeledartifactimpedanceapproachesalargevalue.Thehigh-frequencyportion(3105to3106Hz)oftherawdatafromFigure 3-2 (b)wasttedtoZj=2fLexpanditwasfoundthat=1:037andthatLexp=1:46106.Thefrequencyrangeischosentominimizetheinuenceofcapacitivebehavioratthelowerfrequenciesandavoidtheerrorsinthedatathatarepresentatthehighestfrequencies.ThisttingissubtractedfromtherawdatawiththeresultsshowninFigure 3-5 .Thehigh-frequencytailvisibleinFigure 3-2 (a)isdiminishedandthesymmetryoftheimaginaryimpedancedatadisplayedinFigure 3-5 (b)isincreaseddramatically.AnanalysisoftheKramers-Kronig(KK)consistencyofthedataprovidesanobjectivemethodofassessinganimprovementinthedata. 40

PAGE 41

RealtgeneratedfromimaginaryimpedancedataandKKrelationsforLSMonYSZmeasuredat900Cinair.a)Rawdata.b)High-frequencycorrecteddata. Figure 3-6 showsthetproducedbytheMeasurementModelToolboxinanimpedanceversusfrequencyformat.TheparametersforthetaredisplayedinTable 3-2 .Despitethehigh-frequencydatamanipulation,thereisstillsomeunusabledatainthehigh-frequencyregime(approaching106Hz)asshowninFigure 3-5 (b).Togetthemeasurementmodeltoconverge,wehadtothrowoutsomeofthehighest-frequencydata;however,asshowninFigure 3-6 theusabledataextendsfrom0.63Hzto50kHzwhiletheusablerawdataextendedtoonly3.9kHzasshowninFigure 3-3 .By105Hzintherawdata,thehigh-frequencyinductancehascausedtheimaginaryresistancevaluetoincreasetoapositivevalueof2,makingthedatacompletelyunusable.Thisfrequencyrangemaybecrucialwhentryingtodeconvoluteimpedancespectraassomecathodicprocessesareexpectedtohavetheirpeakfrequencyinthisregime[ 81 ].Aswiththerawdata,sixreplicateswereusedtodeterminethestandarddeviationasafunctionoffrequencyforthehigh-frequencymanipulateddatamodels.Figure 3-4 (b)showsthestandarddeviationasafunctionoffrequencyforthehigh-frequencymanipulateddata.Itshouldbenotedthattheerrorshowsthesamegeneraltrendastheerrorfromtherawdataandisofthesameorderofmagnitude.Theusabledatawasextendedanorderofmagnitudewithoutcompromisingthequalityofthedata. 41

PAGE 42

Randvalueswiththeirrespectiveerrors(R;)forhigh-frequencycorrecteddatameasuredat900C. Process# R() Element1 0.402 0.038 0.023 0.0016Element2 1.120 0.081 0.100 0.0083Element3 3.97 0.42 0.378 0.027Element4 3.54 0.45 0.831 0.058Element5 0.303 0.062 3.41 0.53Element6 0.065 0.008 33.85 6.47 Figure 3-7 showstherealvariancedividedbytheimaginaryvarianceasafunctionoffrequencyforboththerawandhigh-frequencycorrecteddata.Forthevastmajorityofthedata,2r=2jisdesirablybetween103and101.Below103Hz,thereisnosignicantdierenceinbetweenthetwocases.Forbothdatasets,theratio2r=2jdecreasesasafunctionoffrequencytosomeminimumvaluethenincreasesatthehighestfrequencies.Thereasonforthedecreaseof2r=2jasafunctionoffrequencyisthattherealpartofthedataandthushasitslargestmagnitudeatlowfrequencies,whiletheimaginarypartofthedatahasitslargestmagnitudeatabout103Hz.Theincreasein2r=2jcanbeexplainedinthattheimaginaryimpedanceapproacheszeroatthehighestfrequenciesused,therefore,jissmallcomparedtoratthehighestfrequencies. Inthepreviousanalysis,weusedVoigtelementsconnectedinseriestotthedataandanalyzetheerrorstructure.TheVoigtelementsusedwereeachcomposedofaresistorconnectedinparalleltoacapacitor.Forsomesystems,atcomposedofresistorsandconstantphaseelements(CPEs)inparallelismoreuseful.Theconstantphaseelementisamathematicaltoolusedtomodeladistributedtimeconstant.AresistorinparallelwithaCPEproducesadepressedsemicircleinthecomplexresistanceplanedescribedbythefollowingequation. Zisthecompleximpedance,Ristheparallelresistor,andQandaretheparametersoftheconstantphaseelementdescribingthefrequencydependenceandthedepression 42

PAGE 43

RealVariance/ImaginaryVariance(2r=2j)forrawandhigh-frequencycorrecteddata. ofthearc,respectively.ForasingleCPEbasedVoigtelement,theslopeinanZjversusfrequencylog-logplotatlowandhighfrequenciestendstoward+and-,respectively.If=1,theconstantphaseelementbecomesanidealcapacitor.Byexaminingthehighandlow-frequencyslopeofFigure 3-2 (b),wecandetermineifthereisconstantphaseelementbehavior.Inthelinearportionofthelow-frequencyrangeofFigure 3-2 (b),aslopeof0.878wascalculated.ThisvalueindicatesthatwehaveCPEbehavior.BecausemultiplecapacitorbasedVoigtelementsarerequiredtomodelasingleCPEbasedVoigtelement,wecannotinferthattherearesixindependentelectrochemicalprocessescorrespondingtothesixcapacitorbasedVoigtelementsusedinthemodel.Itshouldbepointedouthowever,thatthevalueoftheregressionanalysisisindeterminationofthequalityandvalidityofthedataandthefocusofthisworkisnotinidentifyingthenumberofactualphysicalprocessescontributingtothecathodicreaction. Using0.878for,theeectivecapacitance(Qeff)canbecalculatedasafunctionoffrequencyasfollowsinEquation 3{5 [ 82 ]. 43

PAGE 44

EectivecapacitancecalculatedforLSMsinteredat1100C. Intheequation,Zj(f)istheimaginaryimpedanceasafunctionoffrequency.Figure 3-8 showstheeectivecapacitanceasafunctionoffrequencycalculatedforLSMsinteredat1100C.TheeectivecapacitancefortheLSMsamplestabilizesataround1000Hzindicatingadoublelayercapacitanceofaround100F. Tofurtherillustratethesignicanceofdatacorrection,wehaveevaluatedactualparametersfortheelectrochemicalprocessesoccurringinthecathodicreactionofLSMonYSZinairathightemperatures.WeusedaseriesresistanceandtwoR-CPEVoigtelementsconnectedinseriestomodeltheindividualprocessesoccurringinthecathodicreaction.ForeachofthetwoVoigtelements,thereturnedparameterswereR,Q,and.TheseparameterscanbeusedtocalculatethetimeconstantforeachoftheR-CPEVoigtelementsaccordingtoEquation 3{6 Ourpreviouswork[ 83 ]hasshownthatthismodelisacceptableandthatthehigh-frequencyprocesscanbeattributedtochargetransferwhiletheintermediatefrequencyprocessisrelatedtoadsorptionofmolecularoxygen.Becauseofthehigh-frequencyartifact,theinitialttingforthechargetransferprocessintherawdatareturnedaQvaluegreater 44

PAGE 45

Changeinpolarizationresistance,RP,constantphaseelementcoecientsQand,andtimeconstant,fromdeconvolutionofrawandcorrecteddataforadsorption(1)andchargetransfer(2)processes. () (mF) (ms) () (mF) (ms) Raw 8.58 0.120 0.913 0.531 0.779 0.0908 1 0.0707Corrected 8.59 0.120 0.912 0.531 0.849 0.0677 1 0.0575%change 8.24 34.0 n/a 23.0 thanone,sowexedthisvalueatone,eectivelymodelingthechargetransferprocesswithanRCVoigtelement.Table 3-3 givestheparametersreturnedfromthemodelandFigure 3-9 displaystherawandcorrecteddataalongwithcorrespondingindividualprocessesfromthemodel.Boththetableandgureshowthatthelow-frequency,higherimpedanceprocess(adsorption,labeled88100inthetable)ispracticallyunchangedbytheperformanceofthehigh-frequencycorrectionprocess;however,evaluationofthehigh-frequencyprocess(chargetransfer,labeled88200)issignicantlyalteredbythecorrection.AsseeninTable 3-3 thedeterminedvalueforthechargetransferpolarizationresistanceandtimeconstantchangeby8.3and23%,respectively.Itisclearthatthehigh-frequencyinductivefeaturecansignicantlydistortdeterminedelectrochemicalparamatersforhigh-frequency,lowimpedancemagnitudeprocesses.Inotherwords,inclusionofKramers-Kroniginconsistentdataintheevaluationcanleadtosignicantdeviationofdeterminedparametersfromtheiractualvalues.Becausetheinductivefeaturehaslargevaluesatveryhighfrequencies,performanceofdatacorrectionisofparticularimportanceforthosewhohaveoptimizedtheirsystemtominimizeimpedanceandimproverateofreaction. 45

PAGE 46

ModeledelectrochemicalprocessoccurringinLSMonYSZmeasuredat900Cinair. 46

PAGE 47

3-10 .Thegureshowsthatforsample1,the Figure3-10. Repetitionsofanimpedancemeasurementtakenat800Cinair. totalresistanceincreasedfromtherstmeasurementtothethird.Itisunclearwhethertheinstabilityobservedindicatesthattheprocessofimpedancemeasurementaltersthesample,orwhetherthesampleisunchangedandtheresistancedierenceisduetovariationsinsampleorientation.Theprolesofsamples2and3alsoshowvariationintotalresistance.Changesofthismagnitudearegreaterthananyerrorsassociatedwiththemodelingofthedataoranycorrectableerrorassociatedwithhigh-frequencyartifacts. Theseries(ohmic)resistanceandtotalcathodicresistanceofthemeasurementsshowninFigure 3-10 aredisplayedinFigure 3-11 .They-axisusedinFigure 3-11 isthetimeconstantofthepeak,measuredastheinverseoftheangularfrequencyofthepeakoftheprolesinFigure 3-10 .AsseeninFigure 3-11 (a),theseriesresistanceofsample1decreasesfromalmost6toabout5.2fromthersttothethirdrun.Samples2and3hadseriesresistancesof5.1and5.7,respectivelygivingameanseriesresistanceof5.6andastandarddeviationof0.45.Thisvariationiseitherduetothemicrostructural 47

PAGE 48

Resistancevaluesforrepetitionsofanimpedancemeasurementtakenat800Cinair.a)Seriesresistance.b)Cathodictotalresistance. variationsinthesamples(despitethefactthattheycomefromthesamebatchandweresinteredtogether)ordierencesinthequalityofthepressurecontacts.Figure 3-11 (b)displaysthetotalcathodicresistancemeasuredforeachsample.Thisvalueincreasesfrom84toabout112fromthersttothethirdmeasurementofsample1.Samples2and3hadcathodicresistancevaluesof96and86,respectively.Samples1,2,and3hadameancathodicresistanceof88.5andastandarddeviationof6.8.Since=RC,theslopeofthedataisequaltothecapacitancewhichwascalculatedtobe14.3F.Theseriesresistancedidnothaveaconstantcapacitance. 48

PAGE 49

83 ].Additionally,wehaveshownthatthedataexhibitsCPEbehaviorandamodelintendedtodescribethephysicalmechanismsofthecathodicreactionshouldthereforebebasedonR-CPEtypeVoigtelements.Whenmodelingthedataforthepurposeofdeterminingelectrochemicalparameters,high-frequency,lowimpedanceelectrochemicalprocessareparticularlyvulnerabletotheinductiveartifact.Alimitingdouble-layercapacitanceofaround100Fwasfoundathighfrequencies.Arepeatabilitystudywasperformedanditwasfoundthatbothseriesresistanceandtotalcathodicresistancehadstandarddeviationsofaround8%,whichislargerthaterrorsassociatedwiththemodelingwhicharegenerallylessthan1%. 49

PAGE 50

2.3 .Asmentioned,theoverallcathodicreactioniscomposedofmanyindividualsteps.Foranelectronicconductingcathode,suchasLSM,thegeneralsequenceofstepsdeningthecathodicreactionisfairlywellagreedon;however,thesignicanceofeachoftheindividualstepsisstilldiscussed.ThesequencebeginswiththearrivalofO2moleculesthathavediusedthroughtheporouscathodetothevicinityoftheelectrolyte.Atsomepointafterthisdiusion,themoleculesareadsorbed(dissociativelyorascompletemolecules)onthesurfaceoftheLSM.Theseadsorbedspeciesnowdiuseonthesurfaceoftheelectrodetowardsthetriplephaseboundary,theareawherethecathode,electrolyteandgasphasemeet.Atthesametime,electrons(orholes)travelthroughthecathode,whileoxygenvacanciestravelthroughtheelectrolyte.Forapurelyelectronicconductingcathode,theoxygenspeciesarerestrictedfromthebulkoftheelectrode,theoxygenvacanciesarelimitedtotheelectrolyte,andtheelectronicspecies(electronsandholes)arelimitedtothecathode.Forthecompletecathodicreactiontooccur,allofthesespeciesmustcometogether;therefore,thecathodicreactionislimitedtotheTPB. Modelingofthecathodicreactionisoftenperformedusinganelectricalcircuitdesigncontaininginductors,capacitorsandresistors.AsingleelectrochemicalstepwithanassociatedcapacitanceandaresistancecanbemodeledbyaVoigtelement(asingleresistorandcapacitorinparallel)[ 84 85 ].AnassumptionthatthevariouselectrochemicalstepsofthecathodicreactionoccursequentiallyleadstoamodelconsistingofaseriesconnectionofVoigtelements.Thisisareasonableassumptionforapurelyelectronicconductingcathodeinwhichonlyonereactionpathwayislikely.Anotherpossibilityisthatoneormoreofthestepsaremutuallydependent.Inthisscenario,asimpleseriesconnectionofVoigtelementsmaynotbesuitable. 50

PAGE 51

Impedancespectrafor1100Csampleatvariousmeasurementtemperaturesinair.a)Nyquistplot.b)Imaginaryimpedanceversusfrequencyplot. InthissectionimpedancespectroscopyisusedtoexaminethecontributionofeachofthesignicantelectrochemicalprocessestotheoverallcathodicreactioninsymmetricLSMonYSZ.AseriesVoigtelementmodel(withconstantphaseelementsinplaceofcapacitors)isusedtoexaminethepO2dependence,activationenergies,relativepolarizationresistancemagnitudes,andtimeconstantstheindividualreactions.TheactivationenergiesandpO2dependenceswerecomparedwiththevaluesdeterminedfromotherauthorstoaidinidenticationoftheindividualelectrochemicalprocesses. 3.2 .Inthischapter,pO2wasoneofthevariablesexamined.Oxygen,air,andargongaseswereusedtoproducethedesiredatmosphericconditions.ForpO2satorabove0.01atm,massowcontrollerswereusedtoregulatetheowofargonandairontothesample,producinggasowsofknowncompositionat40cc/min.ForpO2sof0.001atmandless,aZIROXSGM5-ELelectrolysisdevicewasusedtogeneratethedesiredconcentrationataowrateof100cc/min. 51

PAGE 52

Deconvolutionofanimaginaryimpedanceversusfrequencyproleintovariousindividualcontributingprocesses. 4-1 (a)and(b)areNyquistand-Z"vs.frequencyplotsofsymmetricLSMsamplessinteredat1100Cfor1hatvariousmeasurementtemperaturesinair.Thepeaksandchangesinslopeofthe-Z"vs.frequencyplotindicatethecharacteristicfrequenciesofthesignicantstepsofthecathodicreaction.The900CarcinFigure 4-1 (a)isbasicallysemicircular,butnotquitesymmetrical.Thisgeometryindicatesthatthereisonelargeresistanceprocessenvelopinganothersmallerresistancehighfrequencyprocess.Atthelowertemperaturesdisplayedin 4-1 (b),twootherprocessesareapparentwhichareattributedtooxygenvacancydiusionthroughthebulkandgrainboundariesoftheelectrolyte[ 10 ].Theseprocessesaresmallerinmagnitudethanthecathodeprocesses,butatlowertemperatures,thecathodeprocessesarenotseenintheprole.Athighertemperatures,theelectrolyteprocesseshaveverysmallresistanceandareoverwhelmedbythecathodicprocessesandinductiveartifacts. Eachoftheimpedanceprolesincludedin 4-1 wasseparatedbyasubtractiontechniqueintothevariouscontributions.Becausetheprimaryfocusofthisworkisidenticationoftheindividualreactionstepsseparatedbytheirrespectivetimeconstants,useofconstantphaseelementsintheplaceofthecapacitorsintheVoigtelementsofthe 52

PAGE 53

4-2 illustratestheseparationofthe500CmeasurementofsymmetricLSMonYSZcellssinteredat1100Cintothreecontributingelectrochemicalprocesses.EachoftheindicatedelectrochemicalprocessesismodeledbyasingleR-CPEVoigtelement.Thisseparationwasrepeatedatallmeasurementtemperaturesforthesample. Figure 4-3 (a)and(b)displaytheArrheniusdependenceofpolarizationresistanceandtimeconstantfortheLSM1100C1hsample.Thetimeconstant()iscalculatedasdescribedinEquation 4{1 InEquation 4{1 Risthepolarizationresistance,Qandarethenon-exponentialandexponentialtermsoftheconstantphaseelement,respectively.Theactivationenergy(Ea)forthepolarizationresistancesoftheindividualreactionscanbecalculatedasfollows. InEquation 4{2 ,Risthepolarizationresistance,Roisaconstant,kistheBoltzmanconstant,andTisthetemperature. ExaminationofFigure 4-3 (a)revealsthatthereissomechangeinslopeinthehightemperatureregionofprocess4.Thisisanindicationthatthechangesinslopeinthehightemperatureregimewereduetoeitherchangesinthehomogeneityofthedominantelectrochemicalprocessstep,oracontributionfromadierentelectrochemicalprocesswithadierentactivationenergy.Figure 4-3 (b)indicatesasomewhatsmootherproleinthesameregionforthesameprocess.Thissupportstheformerofthetwopossibilities 53

PAGE 54

TemperaturedependenceoftheseparatedcontributionsinLSMonYSZsinteredat1100Cfor1hinair.ThenumbersindicatetheprocessstepnumbergiveninTable 4-1 .a)Polarizationresistance.b)Timeconstant. Table4-1. Selectelementarystepsofthecathodicreactioninsamplessinteredat1100C. ProcessIdentity xinEquation 4{3 Ea(eV) 1 Ionicdiusion(bulk) 0.0 1.1 -2 Ionicdiusion(grainboundary) 0.0 1.04 -3 Chargetransfer 0.0 0.97 8.51054 Dissociationandsurf.di. -0.15 1.17 1.81015 Gasdiusionthroughcathode -1.1 0.0 6.0100 4-1 Figures 4-4 (a)and(b)displaytheinuenceofpO2onthevariouselectrochemicalprocessesoccurringat900C.pO2srangingfrom0.21to1106atmwerecreatedbymixingairwithargon.TheFigureindicatesthatasecondprocessinthelowfrequencyregimegainssignicantmagnitudeatlowpO2values.Theprocessisrstresolvableat0.001atmandgrowsaspO2isfurtherreduced.ThedependenceofpolarizationresistanceonpO2isdisplayedinFigure 4-5 .ThehighfrequencysectionofFigure 4-4 (b)showsan 54

PAGE 55

ImpedanceresponseofanLSMcathodeonelectrolytesamplemeasuredat900CwithpO2(atm)asaparameter.a)Nyquistplot.b)Imaginaryimpedanceversusfrequency. increaseinimaginaryimpedanceforthelowerpartialpressuremeasurements.Webelievethattheincreaseinimaginaryimpedanceabove105HzisanexperimentalartifactduetoinductionandnottheresultofsomenewhighfrequencyprocessthatisnotpresentatthehigherpO2s.ComparisonofFigures 4-5 (a)and(b)showsthatthelowfrequencyprocesshasasignicantlystrongerdependenceonpO2thantheotherprocesses.ThedependenceofpolarizationresistancecanbeexpressedasafunctionofpO2accordingtothefollowingexpression. Theexponentialvalue,x,isequaltotheslopeofthelineinFigure 4-5 anditsvaluesareindicatedin 4-1 .ThestrongdependenceofthisprocessonpO2indicatesthatitisverycloselyrelatedtooxygendiusion.Becauseimpedancespectroscopyisanelectrochemicalmeasurementtechnique,thediusionofgasmoleculescannotberegistereduntiltheyareconvertedintoaspeciesthatcanparticipateintheelectrochemicalreaction[ 74 ].Arapidlyadsorbingsurface,whichpermitsanoxygenuxequaltotheowofgasthroughtheopenpores,wouldproducesuchaneect.Thelowestfrequencyarc(process5)inthelowestpartialpressureregimeisduetoaprocesslimitedbythebulkdiusionof 55

PAGE 56

Dependenceofcathodicpolarizationresistancesinan1100CsinteredsampleonpO2.Thenumbersindicatetheprocessstepnumbergivenin 4-1 .a)Measuredat800C.b)Measuredat900C. oxygenmoleculestotheadsorbingsurface.Thetwoprocesses(process1and2)whicharepresentinthehighfrequencyregimeandatlowtemperatures,becomehiddenathighertemperatures.Theseprocessesarelikelyduetotheelectrolytebulkandgrainboundary,aconclusionsupportedbytheirdeterminedactivationenergies.Thisleavestwoasyetunidentiedprocesses,whichcanbeattributedtothecathode.Oneoftheseprocesseshasmuchsmallerimpedanceandislocatedinthehighfrequencyregime,andtheotherislocatedinthelowerfrequencyregimeandhasmuchlargerimpedance,greatlyinuencingtheoverallshapeoftheNyquistplotathightemperatures.AchargetransferlocatedattheTPB,wouldberatherfastprocesswithweakdependenceonpO2.Figure 4-5 (a)indicatesthatprocess3isnearlyindependentofpO2,whileprocess4ishasanexponentialdependenceof-0.15.Chargetransferistwostepsremovedfrommolecularoxygen,whileadsorptiondirectlyinvolvesmolecularoxygen;therefore,process3islikelyduetochargetransfer.Adsorption,dissociation,andsurfacediusionareallpossibilitiesfortheidentityofprocess4,whichhasastrongerdependenceontemperatureandpO2thanthechargetransferreaction.TheidentitiesofthecontributingprocessesaresummarizedinTable 4-1 56

PAGE 57

4-6 .Inthegure,processes1and2areduetoionictransferthroughthebulkandgrainboundaryoftheelectrolyte,respectively.(Theprocessindicatedby0isanartifactrelatedtoovercorrectionfrominductanceinthesystem.)Numbers3and4indicatecathodicprocesswhicharerelatedtochargetransferanddissociativeadsorption,respectively.Process3isofmuchsmallermagnitudethan4andcloseinrelaxationfrequencyandisthereforeenvelopedinthegure.Process5isonlyresolvableatlowpO2sandisrelatedtobulkdiusionofoxygengastothereactionsite.Polarization Figure4-6. Impedancedatameasuredatvarioustemperaturesforan1100C,1hsinteredsampleat0.002%O2.ThenumbersindicatetheprocessidentityfromTable 4-1 57

PAGE 58

4-1 .Sincetheprocessindicatedbya0isnotaresultofanyphysicalphenomenaassociatedwiththecathodicreductionreaction,itisnotincludedinthetable.ItwasfoundthatchargetransferanddissociativeadsorptionprocesseswerenotstronglydependentonpO2,whereasthebulkdiusionrelatedprocesswasstronglydependentonpO2. 58

PAGE 59

55 ].AuthorshaveusedcompositecathodestoincreasethetriplephaseboundarylengthandshownthatincreasingtheTPBarearesultsinreducedelectroderesistanceforLSM/YSZsystems[ 56 57 ].Indealingwithsingle-phasecathodes,triplephaseboundarylengthcanbemaximizedbyensuringhighporosityintheelectrodeandgoodadhesionbetweentheelectrodeandelectrolyte.Despitebeingusefulforhigh-temperatureSOFCs,theperformancestabilityofLSMonYSZfuelcellsisanissue[ 2 73 86 ].MuchofthepowerlossinSOFCsisduetopolarizationlossatthecathode/electrolyteinterface,onesourceofwhichisdegradationofthisinterface.Thisdegradationcanbemicrostructural,suchasseverecoarseninganddelamination,orcompositionalaswiththeformationoftertiaryphases.Tertiaryphases,whichmayformduringfabricationorlongtermhightemperatureoperation,areofteninsulatingandthereforedetrimentaltoSOFCperformance[ 87 ].Theseandothereectscanbeinducedinatimelyfashionthroughtheuseofhightemperatureanneals.Theimpactofmicrostructuralandinterfacialchangescausedbyharshannealsontheoverallcathodicreactionisstudiedinthischapter. 68 88 89 ].At1000C,diusionofMnintoYSZisnegligible,butfrom1200-1400CMndiusionbecomesconsiderable[ 90 ].AconsequenceofthisfactisthatregionsformneartheinterfacethataredecientinMnandtherefore,highinLa,Sr,Zr,andO.Ifconcentrationoftheseelementsbecomessucientlyhigh,formation 59

PAGE 60

91 ].Ofthesetwophases,theonethatmaterializesisdeterminedbytheSrdopantconcentrationintheLSMandthetemperatureofanneal[ 2 89 92 93 ].BothofthesephaseshavehigherresistivitiesthanYSZsotheirpresenceisdeleterioustodeviceperformance[ 94 { 97 ].ChiodelliandScagliottihavemeasuredtheconductivityoftheLa2Zr2O7layerviaimpedancespectroscopy[ 98 ].TheauthorsecientlyformedanLZlayerbysolidstatereactionoflanthanumoxideandasinglecrystalYSZsubstrate.Theworkreportsconductivitiesof2104and8:6102Scm1forLZand9.5mol%yttriaYSZ,respectivelyat1000C.Additionally,itisreportedthattheconductivitydierenceincreasesastemperaturesarelowered.YandZrwillalsodiusefromYSZintotheLSM,however,thisdiusionoccurstoalesserextentasshownbyYangetal.[ 86 ].ThepresenceofSrintheLSMhasbeenshowntosuppressthediusionofMnintotheYSZ[ 99 ].ThisleadstoaphasecompositionandreactionlayerthicknessthatisdependentonSrcontent[ 68 ].KenjoandvanRoosmalenhaveindependentlyreportedthatLZformationcanbesuppressedbyusingnon-stoichiometricLSM[ 92 96 ]. Taimatsuetal.foundthatthereisalimitedtemperaturerange,inwhichsecondaryphasesformfromsolidstatereactions,suitableforstudy.Theworkfoundthatabove1450C,liquidphaseswerealwaysformedatLaMnO3/YSZinterfacesandnear1300C,reactionsweretooslowfortheirprocessestobeexaminedwithinafewweeks.Below1425Creactionsproceededinthesolidstate,andmorphologiesofreactionzonesresembledoneanother.Therefore,annealingnear1400CwasusedtoexaminethereactivityofLaMnO3withYSZbytheauthors[ 100 ].Otherworkshavealsoused1400Cannealstoproducereactionlayers[ 86 88 ].Yangetal.formed3-4micronthickreactionlayersconsistingofSZandLZwitha1400C48hrringofascreenprintedLSM(0.3Sr)onYSZinterface. 60

PAGE 61

68 ]. 3.2 .Inthischapter;however,LSMpowderprovidedbyNextechwasused.Thepowderwasmixedwithbonders,plasticizers,andsolventstoproduceaninkofthedesiredviscosity.TheLSMpowderwasstoichiometricwitha1:4SrtoLaratio.Aftersinteringat1100C,samplesweresubjectedtohigh-temperaturepost-annealsintendedtosimulatethepossibleeectsoflong-termoperationinatimelyfashion. AJEOL1400SEM(scanningelectronmicroscope)equippedwithenergy-dispersiveX-rayspectroscopy(EDS)wasusedforsampleimaging.X-raydiraction(XRD)wasperformedonaPhilipsAPD3720XRD.TEManalysiswasperformedusingaJEOL200CXTEM(tunnelingelectronmicroscope)byMarkClarkandisincludedinAppendixA.AllmicrostructuralcharacterizationwasperformedattheMajorAnalyticalInstrumentCenter(MAIC)attheUniversityofFlorida. 5.4.1ElectrochemicalandMicrostructuralCharacterization 5-1 .Inthisgurea600Cmeasurementtemperaturewasusedinthefrequencyresponseanalysis. 61

PAGE 62

Complexplaneplotsmeasuredat600CofsymmetricalLSMonYSZsamplesassinteredandaftera1400C,48hanneal. Uponcomparisonoftheprolebeforeandafterthe1400Canneal,itisapparentthatoneprocesswithacharacteristicfrequencyof3.2Hzispresentinbothspectra.Examiningthelow-frequencyregimerevealsdrasticchangeasthesystemexhibitslinearbehaviorwithananglecloseto45.Suchbehaviorhasbeenreportedin[ 64 ]andisdescribedasaWarburgimpedancecausedbyadiusionlimitation.Inthehigh-frequencyregime,aprocessappearsinthepost-annealedsamplethatwasnotapparentintheassinteredsample.Thecausesandsignicanceofthesechangesarediscussedbelow. Figures 5-2 and 5-3 displayEISprolesforthesamplebeforeandaftertheapplicationofasubsequent1250C48hanneal,respectively.TheEISdatawastakenatlowmeasurementtemperatureswherethehigh-frequencybehaviorisemphasized.ComparisonofFigure 5-3 (a)to 5-2 (a)and 5-3 (c)to 5-2 (c)revealslittlechangeinthehighandintermediate-frequencyprocesses(electrolytebulkandgrain-boundarypolarizationresistance,respectively)afterthe1250Canneal.Inthelow-frequencyregimethe1250Cannealseemstohaveacleareect.Thelow-frequencyprocess(relatedtomolecularadsorption),whichissemicircularinFigure 5-2 ,isreplacedbyWarburgbehaviorinFigure 5-3 (especiallyapparentat500C).AsexplainedbyMacdonaldinreference[ 64 ],Warburg 62

PAGE 63

Impedancespectraforassintered(1100C)sampleatvariousmeasurementtemperatures.a)Complexplaneplot.b)High-frequencyviewofthecomplexplaneplot.c)Imaginaryimpedancevs.frequencyplot. 63

PAGE 64

Impedancespectraforsampleafterasubsequent1250C,12hannealmeasuredatvariousmeasurementtemperatures.a)Complexplaneplot.(b)Highfrequencyviewofthecomplexplaneplot.c)Imaginaryimpedancevs.frequencylog-logplot. 64

PAGE 65

Scanningelectronmicroscopy(SEM)imagesofthecathode/electrolyteinterfaceassinteredat1100C. behaviorisgenerallyaconsequenceofadiusionlimitation,thusthe1250C12hanneallikelyhindersthediusionofambientoxygentothereactionsite,signicantlyimpedingthecathodicreaction.ThisisduetocoarseningoftheporousLSMtosuchanextentthatthediusionofoxygenthroughthecathodeisreducedoreliminated.Additionally,coarseningoftheLSMcouldgreatlyreduceordestroytheTPB,whichwillalsoinhibitthecathodicreaction. Figures 5-4 through 5-6 displaythemicrostructureofthesymmetricsamplesattheLSM/YSZinterface.Figure 5-4 showstheassinteredmicrostructureafteran1100C1hanneal,whileFigure 5-5 (a)showsthemicrostructureaftera1250C12hanneal.ComparisonoftheguresillustratesthecoarseningoftheLSMmicrostructurethatoccurswithanannealofthisseverity.ThismicrostructuralchangeforecastedbythechangesintheimpedanceprolewiththeharshannealisveriedbytheSEMimages. Figure 5-5 (a)showscoarseningoftheLSMmicrostructureaftera1250Cannealof12h.After48hthecoarseningbecomesmorepronounced(Figure 5-5 (b)).ThecoarseningoftheLSMafter1400Cannealismorecomplete(Figure 5-6 ).Infact,thecoarseningafteronly1hrat1400Coccurstoagreaterdegreethanafter48hat1250C.FocusingontheLSM/YSZinterface,itisnoticedthatforthe1400Cannealed 65

PAGE 66

ScanningelectronmicroscopyimagesoftheLSM/YSZinterfaceforsamplessinteredat1250C.a)Sinteredfor12h.b)Sinteredfor48h. Figure5-6. Scanningelectronmicroscopyimagesofthecathode/electrolyteinterfaceforsamplessinteredat1400C.a)Sinteredfor1h.b)Sinteredfor12h. 66

PAGE 67

Impedancespectrafor1400C,12hannealedsampleatvariousmeasurementtemperatures.a)Complexplaneplot.b)High-frequencyviesofthecomplexplaneplot.c)Imaginaryimpedancevs.frequencyplot. samples,acoalescenceofvacanciesintoextendedinterfacialporesoccursthatisnotpresentinthe1250Cannealedsamples.Theeectofthismicrostructuralchangeisinhibitionofthecathodicreaction,byblockingelectronsfromreachingtheTPB. Thelowtemperaturefrequencyresponseofatestsamplesubjectedtoa1400C12hannealisdisplayedinFigure 5-7 (a-c).ComparingFigure 5-7 withFigure 5-3 illuminatesthedierencesbetweenannealat1250and1400Contheimpedancespectra.Theelectrolytebulkandgrain-boundaryprocessespresentinFigure 5-3 (a)arereplacedbyonehigh-frequencyprocessinFigure 5-7 (a).Theimpedancemagnitudeofthehigh-frequencyprocessvisibleinFigure 5-7 (a)ismorethananordermagnitude 67

PAGE 68

5-3 (a).Thisincreaseisduetothemicrostructuralandinterfacialchangesthatoccurwiththe1400Cannealincludingtheformationofinsulatingtertiaryphases,eradicationoftheTPB,andtheformationofextendedinterfacialpores.Signicantcoarseningoccursat1250CasdisplayedinFigure 5-5 ;howeveritisclearfromtheimpedanceprolesthat1400Csinteredsampleissignicantlymoredegraded.FocusingonFigures 5-3 (c)and 5-7 (c),weseethatthelow-frequencybehaviorofbothsystemsissimilarwiththeexceptionofthe500Cmeasurement,yetdierentfromFigure 5-2 (c).Althoughthe500CimpedanceproleinFigure 5-6 (c)appearstobegintocloseoatlow-frequencies,itislikelythatwerethelow-frequencylimitoftheEISdecreased,adiusionlimitationtailwouldappear.ThisisanexpectedresultofthecoarseningoftheLSMmicrostructure,whichpreventsfreeowofmolecularoxygentotheelectrolyte. Figure5-8. High-frequencyarcresistanceversusannealtemperatureforvariousanneal(temperature,time)pairsmeasuredat400C. Figure 5-8 illustratesthedependenceofthehigh-frequencyEIScontributiononannealtemperatureandtime.Theguredisplaysagradualincreaseofhigh-frequencyarcresistancewithannealattemperaturesbelow1400C.Inthistemperaturerangethereisnocleardependenceonannealtime.At1400C,however,theimpedance 68

PAGE 69

5-6 whichshowsthatby1hthemicrostructurehasbecomenearlydense,leavinglittlepossibilityofincreasedcoarseningwithlongertimeanneals.Factorsotherthanmicrostructuremustbeconsideredwhenexplainingtheobservedincreaseinhigh-frequencyimpedance. Energy-dispersiveX-RaySpectroscopy(EDS)linescanofMnKintensityatLSM/YSZinterface.a)Assintered.b)After1400C,48hanneal. Figure 5-9 showsEDSlinescansofthemanganeseKintensityattheLSM/YSZinterfaceofassinteredsampleswithandwithouta1400C48hanneal.DespitethefactthatMnisknowntobeafastdiuserathightemperatures[ 88 90 ],theproleismoreabruptintheharshlyannealedsample.Thisapparentcontradictioncanbeexplainedby 69

PAGE 70

The1250C,1325C,and1400C,12hannealedsampleswereexaminedbyXRDasdisplayedinFigures 5-10 (a)-(c),respectively.Inallsamples,energypeakscharacteristicofLSMandYSZwereobserved.Inadditiontothesepeaks,thereisevidenceofatertiaryphasepresentattheinterfaceofthe1325Cand1400Cannealedsamples.TheproximityofsomeofthetertiaryphasepeakstothepeaksofLSMhindertheanalysis.ToreducethecomplexityoftheXRDspectra,thesampleswerebathedinhighlyconcentratedhydrochloricacidtoremovetheLSMlayer.XRDoftheLSMstrippedsamplesclearlyshowasetofpeakscorrespondingtoYSZandanothersetofpeaksforthe1325Cand1400Cannealedsamples.Thesecondsetofpeaksmatchesthosebelongingtoaknownlanthanumzirconate(La2Zr2O7)sample.AdditionalcompositionalanalysisusingtunnelingelectronmicroscopywasperformedbyDr.MarkClarkandisincludedinAppendixA.Theseresultsshowthata0.2micronthicktransitionalregionrichinlanthanumandzirconiumformsattheLSM/YSZinterfaceafterannealingat1400for48hours.TheinterfacialregionwasshowntohaveasimilarcrystalstructureasYSZ,indicatingthatmanganesediusesintotheelectrolyteasthetertiaryphaseforms. 70

PAGE 71

X-raydiractionofsamplessubjectedtopost-annealsintering.NoLZpeaksareobservedforthe1250Csinteredsample.a)Sinteredat1250C.(b)Sinteredat1325C.c)Sinteredat1400C. 71

PAGE 72

72

PAGE 73

81 ].Becausehightemperaturesarerequiredforfabrication,tosomeextentthesechangescannotbeavoided.Theoverallresultisthatthecathodicreaction,whichisdependentonoxygengasowingthroughthepores,diusingtowardsthereactingsite,andbeingtransferredtotheelectrolyteisaectedbythealteredmicrostructure.Theimpactofmicrostructuralandinterfacialchangesontheelectrochemicalstepscontributingtotheoverallcathodicreactionisinvestigatedinthischapter. Inapreviouschapter,theimpactofveryhigh-temperatureannealsonthecathodicreactionwasexamined.Bothdramaticchangesinmicrostructurewereproducedandtertiaryphaseswereformedatthecathode/electrolyteinterface,alteringthecathodicreaction.Inthischapter,microstructuralchangesareproducedbysinteringatlowertemperaturesinanattempttodecouplemicrostructuralchangesandtertiaryphaseformation.Additionally,non-stoichiometricLSM((La0:8Sr0:2)0:98MnO3)wasusedwhichhasbeenshowntodecreaseformationoftertiaryphases[ 92 96 ].Establishingadirectrelationshipbetweencathodemicrostructureandelectrochemicalperformancewillclarifythecathodicreductionreactionpathwayandaidinidenticationoftherate-limitingstep,whichisnotknownconclusively[ 73 74 ]. SeveralworkshavebeenperformedwiththegoalofestablishingarelationshipbetweenLTPBandelectrochemicalproperties.Typically,theDCresistance,ortheentirecathodicresistancedeterminedfromimpedancespectroscopyarerelatedtoelectrochemicalperformance.OneofthemostfrequentlycitedworksintheareawasperformedbyMizusakietal.whofoundthattotalcathodeconductivity,measuredbyimpedancespectroscopyat1000C,isessentiallyproportionalto(LTPB)1fordrip 73

PAGE 74

101 ].Theresultsfromthisworkworkaresomewhatquestionablesinceonlytwodatapointsareplottedtogivetheobserveddependence.Additionally,whenadierentfabricationtechniquewasused,anon-linearpowerdependenceoftotalcathodicpolarizationresistanceonLTPBwasreported. AnotherfrequentlycitedworkwasperformedbyKuznecovetal.[ 102 ],whoderivedarelationshipsuccessfullyexplainingthelineardependencereportedbyMizusakii.e.Rcathode/(LTPB)1.Inthemodel,theauthorassumesthatsurfacediusionofadsorbedspeciestowardstheTPBdominatesRPandthattheDCresistancecanbemodeledbyconsideringthisowofadsorbedspeciestotheTPB.Themodelbasicallyrelatesthecathodicresistancetotheuxofadsorbedspeciesonthesurfaceofthecathode.AsreportedbyMacdonaldetal.asurfacediusionlimitationisoftenmanifestedbyWarburgbehaviorintheimpedanceprole[ 110 ];however,wedidnotseeWarburgbehaviorathigh-frequenciesandsothedevelopmentdescribedintheworkmaynotbeappropriateforourdata. InalaterworkbyKuznecovetal.itisreportedthatforkineticscontrolledbybulkdiusionthroughthecathode,Rcathode/(MIEC/YSZcontactareaLTPB)0:5[ 103 ]ThisdevelopmentisbasedontheworkofAdleretal.[ 52 ].ThemodelofAdleretal.wasintendedtodescribecathodeswithsignicantionicconductivityandisreportedbyAdleretal.tobeinconsistentwithLSMbehavior.Thismodelconsidersowofionicspeciesthroughthecathodebulktobethelimitingfactorandcalculatestheconductivityfromthisow.FromthedataofKuznecovetal.,apowerdependenceof-0.39canbecalulated(fromonlytwodatapoints)whenrelatingtotalRcathodetoLTPBforLSMmeasuredat950C[ 103 ]. Inanotherwork,Fleigshowedthatforwelldened,densepatternedLSMmicroelectrodesofcirculargeometry,thetotalcathodicresistanceisproportionaltoelectrodediameter(D)tothe-2.1powerwhenmeasuredat800C.Inthisgeometry,LTPBisequaltothe 74

PAGE 75

50 104 ].Fleigconcludedthatsincetotalcathodicresistancescalesinverselywithelectrodecontactarea(area=0.25D2)andlinearlywiththickness,abulkpaththroughtheelectrodedeterminestheoxygenreductionrate,withtransportofoxideionsinLSMbeingtherate-determiningstep.ItshouldbepointedoutthatthedensecircularelectrodesusedbyFleighadthicknessesofonly0.1to0.25manddiameterontheorderof60m,soionictransportthroughtheelectrodecouldbeappreciable.Incontrast,inthepresentworkweusedporouselectrodeswiththicknessontheorderof20mandindividualparticlesizesofaround1m.Therefore,dierentreactionpathwaysarelikely.Forexample,ifanoxygenmoleculeisadsorbedatthecenterofthetopofoneofthecirculardisks,theadsorbedspeciesmusttravelover30mtoreachtheelectrolyteviaasurfacepath,butonly0.1mtoreachtheelectrolytethroughthecathodebulk.Ontheotherhand,forrelativelysphericalparticles,thepathtotheelectrolyteviathesurfacewillbearoundthesamedistanceforbothsurfaceandbulkdiusion.Obviously,boththeshapeofthecathodeandtherelativeeaseoftransportthroughthebulkversusonthesurfacewilldeterminewhichpathisfavorableforanadsorbedspecies.Additionally,surfacearea(notvolumenormalizedsurfacearea)alsoscaleslinearlywithcathodethicknessandsoadditionalevidenceisneededtosupportthebulktransportconclusion. OneoftherstauthorstoutilizeknowledgeofarelationshipbetweenpolarizationresistanceandLTPBwasOstergardet.al.whodecreasedpolarizationresistancebyformingcompositecathodeswhichhaveincreasedLTPB[ 57 ].Ithasbeenreportedthatthedependenceofpolarizationresistanceoncompositecathodesthicknessdependsonthemeasurementtemperatureandthatascompositecathodethicknessincreases,polarizationresistancedecreasesuntilgasdiusioneectsbecomeimportant.[ 105 106 ] 75

PAGE 76

75 ].Incontrast,thisworkmakesuseofadualbeamFIB/SEM(Focusedionbeam/scanningelectronmicroscope)formicrostructuralcharacterizationthatallows3-Dreconstruction.Ofthemicrostructuralparameterscommonlystudied,fourareconsideredtobethemostcriticaltoelectrochemicaleciency.Theseincludeporesurfacearea,triplephaseboundarylength(LTPB),porosity,andtortuosity. Anoxygenmolecule,whichhasdiusedthroughthegasphasetothecathode,mustrstbeadsorbedbeforeitcanparticipateinthereductionreaction.ThisadsorptioncanoccurveryclosetotheTPBorfurtheraway,dependingonthediusivityoftheadsorbedspecies.Ithasbeenproposedinfact,thatoxygenreductioninanelectronicconductorcanbeco-limitedbybothadsorptionandsurfacediusion[ 107 ].Bothadsorptionandsurfacediusionaredependentontheporesurfacearea;therefore,poresurfaceareaisoneofthekeymicrostructuralparametersforourinvestigation. Ascoarseningoccurs,smallcathodeparticlesattheinterfacecoalesceintolargerones,thusreducingthetotalTPBlengthpersurfacearea.OtherauthorshaveshownthatincreasingLTPBresultsinreducedelectroderesistanceforLSM/YSZsystems[ 57 108 ].Thisreductionisadirectconsequenceofthefactthatinpureelectronicconductingelectrodes,theelectrochemicalreactiondrivingfuelcelloperationisrestrictedtotheTPB[ 55 ]duetotheexclusionofionsfromthebulkoftheelectronicallyconductingcathodeandofelectronsfromthebulkoftheelectrolyte.Wethereforecananticipateanincreaseinchargetransferresistanceassinteringtemperatureisincreased. 76

PAGE 77

Tortuosityisapropertythatquantiesthecomplexityofthepaththroughwhichadiusingparticlemusttravelinordertoreachadesireddestination.IntermsofSOFCs,tortuosityisaunitlessparameterdenedasthedistancetraveledbyamoleculeexitinganimpinginggasowasittravelsthroughtheporouscathodetoreachthesolidelectrolyte,dividedbythestraight-linedistance.AlargetortuositycorrespondstoaconvolutedpathforagivengasmoleculetotraverseinordertogofromthegasstreamtotheTPB.Becausedatainthreedimensionsisnecessaryforatruetortuosityanalysis,verylittleworkispublishedforactualsystems.ThedualbeamFIBgivesusthethree-dimensionaldatanecessaryforthemathematicalevaluation.Weexpectthatcathodemicrostructureswithalargetortuositywillshowanincreaseingasdiusionpolarizationresistanceandrelatedelectrochemicalproperties. 3.2 .Inthischapter,samplesweresinteredattemperaturesrangingfrom950to1325Cforonehour.Thesinteringtemperaturerangewaschosentoproducemicrostructuralchangeswhichcouldbequantiedandcomparedtochangesintheelectrochemicalbehaviorofthesamples.Microstructuralimageswereattainedusingadualbeamfocusedionbeam/scanningelectronmicroscope(FIB/SEM,FEIStrataDB235)byAijieChen.TheFIB/SEMsetupisdescribedinAppendixB. 77

PAGE 78

Scanningelectronmicroscopy(SEM)images,createdusingafocusedionbeam/SEM(FIB/SEM),ofLSMonYSZsinteredfor1hatvarioustemperatures.a)Sinteredat1100C.b)Sinteredat1200C.c)Sinteredat1300C. 6.3.1EectofSinteringonMicrostructure 6-1 (a)6-1 (c)showinterfacialcrosssectionsofLSMonYSZsinteredat1100,1200,and1300C,respectively.Itiseasytoseethatby1300C,themicrostructurechangesdrastically.Comparisonofgures 6-1 (a)and(b)revealsmoresubtledierences.The1100Cannealedsampleappearstobeslightlymoreporousthanthe1200Cannealedsample.Thisapparentdierenceissoslightthattoconclusivelysaythereisanychangeinporosityrequiresquantitativecalculations.High-frequencyartifactsinthedatawereaccountedforasdescribedinsection 3.3.2 Visually,themostnoticeabledierencebetweenthe1100and1200Csinteredsamplesisintheconnectivity.The1100Csinteredsampleshowsmanyroundshaped 78

PAGE 79

BecausetheTPBisofparticularimportancetothecathodicreaction,wecloselyexaminedthecathode/electrolyteinterface.Oninitialinspection,itappearsthatthecathodetoelectrolytecontactsurfaceislargerforthe1200Csinteredsamplethanthe1100Csinteredsample.Severaloftheparticlesclosetotheelectrolyteforthe1100Csinteredsampledonotappeartocontacttheelectrolyte.Incontrast,forthe1200Csinteredsampleface-to-facecontactshavebeenformedbetweenthecathodeandtheelectrolyte.Thenatureofinterfacialvoidsalsochangessignicantlybetween1200and1300C.Figure 6-1 (c)showstheformationoflargeinterfacialvoidsatthecathode/electrolyteinterface.Theselargeinterfacialvoidsformassmallervoidscoalescewhilebeingrestrictedfromthedenseelectrolyte.FormationoftheseinterfacialvoidswillgreatlyreducethemeasurableTPBlength. FIB/SEMwasperformedonsamplessinteredatthevarioustemperaturesandmicrostructuralfeatureswerequantiedasdescribedinAppendixB.Porosity(p),volumenormalizedporesurfacearea(SV),andTPBlengthvalueswerecalculatedateachtemperature,whiletortuosity()valueswerecalculatedfromthe3-Ddataattainedatselectedtemperatures.Porositywascalculatedfromtheporearea/totalareaineachSEMimage.Thecalculationwasrepeatedforallslicesinthesampleandanaverageporositywasattained.TheseresultsareplottedinFigures 6-2 and 6-3 79

PAGE 80

PoresurfaceareaandLTPBasafunctionofsinteringtemperature.a)Poresurfacearea.b)LTPB. FromFigure 6-1 ,wecanseethatasthesinteringtemperatureincreasesfrom1100to1300C,themicrostructurechangesfromonewithsmallporestoamicrostructurewithlargepores.Fromelementarygeometry,wewouldexpectamicrostructurewithmanysmallporestohavealargersurfaceareathanonewithlargepores.Figure 6-2 (a)conrmsthisndingandshowsthatthevolumenormalizedporesurfaceareadecreasesassinteringtemperatureincreasesfrom1150to1325C. Figure 6-2 (b),showsthatLTPBdecreaseslinearlyassinteringtemperatureisincreasedinthetemperaturerangedisplayed.Thereareoutliersat1225and1250C.Thesedeviationscouldbecausedbylocalizedinterfacialvoids,anunusuallyneinterfacialmicrostructure,oralowerthananticipatedsinter.Adeterminationastowhetherthecauseoftheoutlyingpointsisalocalanomalyoracharacteristicofthebulksamplecanbemadebystudyingtheelectrochemicalbehaviorofthesamplesinquestion,whichisperformedlater. Figure 6-3 (a)showsthattheporosity(calculatedbyAijieChen)startsatabout30%fora950Csinteredsampleandincreasesslightlywithincreasingsinteringtemperatureto1200Candthenbeginstodropo.By1400C(notshown),theporosityhasdroppedtolessthan5%indicatinganalmostdensecathodelayer.Thistrendissupportedbyour 80

PAGE 81

Porosityandtortuosityasafunctionofsinteringtemperature.a)Porosity.b)Tortuosity.(BothparamaterscourtesyofAijieChen.) qualitativeanalysisofFigure 6-1 andthefactthatthemeltingtemperatureofLSMonYSZisontheorderof1450C[ 100 ]. Thetortuositywascalculated(calculatedbyAijieChen)forselecttemperatures.Tortuosityvaluesof3.23,2.18,and4.27werecalculatedforthe1100,1200,and1300Cannealedsamples,respectively.Theminimumtortuosityoccursatabout1200Cindicatingthatthegasmoleculeshavethemostdirectpathtotheinterface.Opposingtrendsaccountsfortheminimumthatisobserved.Atlowsinteringtemperatures,particlesizeremainssmall,andgasmoleculesareredirectedmanytimesastheytraversethepathtotheLSM/YSZinterface.Athighersinteringtemperaturestheporesarelarge;however,someofthepathsmaybecomeclosedo,limitingthenumberofavailablepathways. 6-4 and 6-5 show800CimpedancemeasurementsofLSMonYSZsinteredatvarioustemperaturesinair.Figures 6-4 (a)and(b)areNyquistplotscoveringtheentirefrequencyrangeandhigh-frequenciesonly,respectively.TheprolesshowninFigure 6-4 (a)aregenerallyasymmetricalinthehigh-frequencyregime.TheeectismorepronouncedinFigure 6-4 (b),whichonlyshowsthehighestfrequencyportionofthedata.Thecauseofthisasymmetryisthepresenceofmultipleprocessesoccurring 81

PAGE 82

Nyquistplotsmeasuredat800CforLSMsinteredatvarioustemperaturesinair.a)Allfrequenciesincluded.b)Highfrequenciesonly. 82

PAGE 83

Imaginaryimpedancevs.frequencyplotmeasuredat800CforLSMsinteredatvarioustemperaturesinair. overthefrequencyrangeexamined.Assinteringtemperatureisincreased,thepresenceofthehigh-frequencyprocessbecomesmorepronouncedasseeninFigure 6-4 (b).Figure 6-5 displaysthefrequencydependenceoftheimaginaryimpedance.Displayingthedatainthisformatmakesapparentthedecreaseinpeakfrequencyoftheoverallreactionassinteringtemperatureisincreased.The1325Csinteredsampleshowsachangeinslopeatabout1kHz.Aninectionisonlyobservedwhentwoormoreelectrochemicalprocessesaresignicant. ImpedanceSpectroscopyofLSMcathodesonYSZsubstrateshasbeenthesubjectofamultitudeofworks[ 53 109 ].MostauthorsagreethattwonoticeableprocessesoccurinoptimallysinteredLSMonYSZathighmeasurementtemperaturesinoxygenrichatmospheres.Atlowoxygenpartialpressuresathirdprocessrelatedtothediusionofoxygengasmoleculesthroughtheopenporesofthecathodetotheactiveregionisobserved. Unfortunately,agreementontheisolationandidenticationofthehighandintermediate-frequencyprocesseshasnotbeenascomplete.Reasonsfordisagreementinclude1)themechanismofreactionisdependentonmeasurementconditions,2)themechanismofreactionisdependentonthesamplepreparationandsamplehistory, 83

PAGE 84

Nestedelementequivalentcircuitusedfortting.Zhfrepresentsfeaturesoccurringattoohighafrequencytobeanalyzed.CPE1isassociatedwiththedoublelayercapacitance,R1isthechargetransferresistance,andR2andCPE2arerelatedtoadsorption. and3)noconsensusisreachedforevaluationofimpedancedata.Toovercomethersttwoproblemsitisimportantforauthorstospecifyascompletelyaspossibleallexperimentaldetails,particularlywhenmicrostructureisnotanalyzed.Inthiswork,wehavecharacterizedthemicrostructureandwillrelateelectrochemicalpropertiesdirectlytothemicrostructureofeachsample.Thethirdproblemisnoteasilysolved. Typically,impedancedataisanalyzedbyttingthedatatoanequivalentcircuit.Oneschoolofthoughtproposesdevelopingamodelwhichisbasedonaprioriknowledgeofthesystem.SeveralauthorshaveanalyzedLSMonYSZusingthismethod.Themostoftenusedcircuitcontainsadoublelayercapacitanceinparallelwithaseriesconnectionofachargetransferresistanceandamasstransferrelatedelement.Forelectronicconductors,themasstransferinterpretationisreplacedbyadsorptionand/orsurfacediusion.ThemasstransferrelatedelementiseitheraVoigtelement,anite-lengthWarburgelement,orsomegeneraldiusionelementthatisnoteasilydenedintermsofcircuitelements.Additionally,allcapacitorsmaybereplacedbyconstantphaseelementstoaccountforinhomogeneitiesinthesystem.ThistypeofcircuitwithslightvariationshasbeenusedbyseveralauthorsandisdepictedinFigure 6-6 [ 46 48 63 112 113 ].Themajordrawbackofthismodelisthateachauthortypicallyhastheirownvariationofthemodelmakingcomparisonofparametersattainedbetweengroupsdicult.ThecommonlyusednestedcircuitshowninFigure 6-6 (withcapacitorsinsteadofconstantphaseelements)wasproducedfromamoregeneralmodelinaworkbyJamnikand 84

PAGE 85

SeriesVoigtelementequivalentcircuitusedfortting.Zhfrepresentsfeaturesoccurringattoohighafrequencytobeanalyzed.EachVoigtelementiscomposedofaresistorandaconstantphaseelement. Meyer[ 62 ].Inthemodel,CPE1representsthedoublelayercapacitance,R1representsthechargetransferresistance,andR2andCPE2representamasstransferphenomenon.Henceforth,wewilltreatLSMasanelectronicconductorandthereforereplacethemasstransferprocessbyadsorptionand/orsurfacediusion.MacdonaldexplainsthatwhensurfacediusionissignicanttheRandlesequivalentcircuitisexpected;however,ifnosignicantlydiusingintermediatesarepresentthediusionalimpedanceisreplacedbyaresistorandcapacitorinparallel[ 110 ].Inthiswork,aWarburgtypeslopewasnotseenathigh-frequencies;therefore,itislikelythatadsorptionismoresignicantthansurfacediusion. AnalternativeequivalentcircuitbasedonaseriesconnectionofVoigtelementsisalsousedbymanyauthorsanddisplayedinFigure 6-7 [ 54 84 114 115 ].Inthistypeofmodel,assignmentofidentitiestotheindividualprocessesisaccomplishedbyidenticationofactivationenergies,pO2dependences,biasvoltagedependences,andothercircumstantialevidence.Themajoradvantageofmodelinginthisfashionisthatcomparisonofeortsbetweendierentgroupsisfacilitated;however,becausethemodelisnotderivedspecicallyforthesystem,condenceisdiminished.Jiangetal.hasusederrorstructureanalysistoshowthatbothofthesemodelscanaccuratelyproducethedesiredresponse[ 116 ].Inourpreviouswork,weexaminedactivationenergies,pO2dependencesandotherevidenceandconcludedinagreementwithothersthatchargetransferwasthehighanddissociativeadsorptionwastheintermediatefrequencyprocesses[ 111 ].Inbothmodels,Zhfrepresentsthetotalimpedanceofallprocessesoccurringat 85

PAGE 86

LookingbackatFigures 6-4 and 6-5 weseethattheintermediatefrequencyprocesshasalargerpolarizationresistancemagnitudebutthatthehighfrequencyprocessincreasesinrelativemagnitudeassinteringtemperatureisincreased.Itshouldbepointedoutthatalargermagnitudemeansalargerpowerconsumptionduetothatmechanism,butdoesnotnecessarilymeanthatthatmechanismistherate-limitingstep.Anincreaseinchargetransferresistanceisevidenceofadecreaseinthequalityofthecathode/electrolyteinterface,wherechargetransferoccurs.Chargetransferpolarizationresistancebecomesmoresignicantathighersinteringtemperaturesduetothedeteriorationofthetriplephaseboundary.Assinteringtemperatureisdecreased,thishigh-frequencyprocessbecomeslesspronouncedandinductiveartifactsbecomesignicantinthehigh-frequencyportionofthedata. BothequivalentcircuitmodelswereusedtotimpedanceprolessuchastheonesshowninFigure 6-4 .Figure 6-8 isincludedasanexampleillustratingthedeconvolutionofthedata.Theimpedanceproleshownisforthe1200Csinteredsamplemeasuredat800Cair.Figure 6-8 (a)showsthedataandthettingobtainedusingthenestedmodel,whileFigure 6-8 (b)showsthedataalongwiththetting(solidline)fromtheseriesmodel.Additionally,Figure 6-8 (b)showstheinidividualcomponentswhicharesummedtoproducetheseriesmodeltting.Becausebothmodelsaccuratelytthedata,moreanalysisisnecessarytodeterminewhichofthetwoismoreappropriate. Sincethemeasurementwasdoneinair,thepolarizationresistanceduetobulkgasdiusionisnegligibleandonlytwoVoigtelementsarenecessaryintheseriestting,oneforadsorption(dashedline)andoneforchargetransfer(dottedline).AscanbeseeninFigure 6-8 (b),asingleprocesswithrelativelylargemagnitude(adsorption)providesthemajorcontributiontotheprole.Above104Hz,chargetransferbecomessignicantandcausestheoverallproletodeviatefromthesymmetriccontributionduetoadsorption. 86

PAGE 87

Deconvolutionofimpedanceprolefrom1200Csinteredsample,measuredat800Cinair,usingbothequivalentcircuitmodels.a)Nestedmodel.b)Seriesmodel. 87

PAGE 88

Theprocesswasrepeatedatvarioussinteringtemperaturesrangingbetween1150and1325C.ThesinteringtemperaturerangewaschosentobeginabovetemperatureswheresinteringisincompleteandendbelowthemeltingtemperatureofLSMonYSZ,whichisaround1450C[ 100 ].Previousresearchshowsthatby1400C,thechargetransferresistancehasincreaseddramaticallybecausetheLSMlayerisfullydense,eectivelydestroyinganytriplephaseboundaries[ 81 ].Theimpedancewasperformedat800Cinair.Infuturework,wewillexaminethecathodicreactioninlowoxygenpartialpressureregimeandrelatethebulkgasdiusionpolarizationresistancetoporosityandtortuosity. Figure 6-9 (a)showsthesinteringtemperaturedependenceofchargetransferRPdeterminedfrombothmodels.Forbothmodels,thechargetransferpolarizationresistanceincreasesexponentiallyassinteringtemperatureisincreased.TheindividualnatureoftheVoigtelementsintheseriesmodelmaycontributetotheimprovedtfortheseriesmodelascomparedtothenestedmodelforchargetransferresistance.TherelativelylargescatterinthechargetransferdataforthenestedmodelisrelatedtothefactthatchargetransferRPisanorderofmagnitudesmallerthantheadsorptionRP,andthetwoprocessesaresolvedforsimultaneously.Incontrast,asubtractiontechniquewhichremovedprocessesindividuallywasusedinthedeconvolutionfortheseriesmodel.Figure 6-9 (b)showsthedependenceofadsorptionRPonsinteringtemperature. 6.3.3.1Seriesmodelevaluation 6-10 relatesthechangeinelectrochemicalperformancecausedbyvaryingsinteringtemperaturetothecorrespondingmicrostructuralchangesbyshowingtheinuenceofTPBlengthonchargetransferRPandtheinuenceofporesurfacearea 88

PAGE 89

Temperaturedependenceofpolarizationresistance(RP)inairdeterminedusingbothseriesandnestedequivalentcircuitsmeasuredat800C.a)ChargetransferRP.b)AdsorptionRP. 89

PAGE 90

RelationofchargetransferandadsorptionpolarizationresistancedeterminedfromtheseriesVoigtelementequivalentcircuittomicrostructuralquantities(measuredinairat800C). onadsorptionRPwithelectrochemicalparametersdeterminedfromtheseriesVoigtelementmodel.Focusingrstonchargetransfer,weseethatthechargetransferresistanceincreasesastriplephaseboundarylengthdecreases.IntheLTPBvs.sinteringtemperatureplotshowninFigure 6-2 (b),wenotedthatthereweretwooutlierpointslocatedat1225and1250Candproposedreasonsfortheirpresence.WhenrelatingchargetransferRPtotheactualmicrostructure,LTPB,weobserveonlyoneoutlierlocatedat1225C.Thecauseoftheoutlierat1250CmustbeabulkcharacteristicbecausetheriseinLTPBwasaccompaniedbyacorrespondingdropinchargetransferRP. Turningourattentiontothedatapointforthe1225Csinteredsample,weseethattheloweredLTPBvalueisnotaccompaniedbyacorrespondingincreaseinchargetransferRP.Infact,thechargetransferRPforthe1225Csinteredsampleissimilarinmagnitudetothe1200and1250Csinteredsamples.SincethelowLTPBvalueseenat1225CisnotaccompaniedbyanincreaseinchargetransferRP,wecanconcludethatthelowLTPBvalueiscausedbyalocalizedphenomenasuchasaninterfacialgapandis 90

PAGE 91

Curvettingofthedataindicatesthatthereisapower-lawdependenceofchargetransferRPonLTPBatan800CmeasurementtemperaturegivenbyEquation 6{1 OtherauthorshaveshownadependenceofoverallpolarizationresistanceontheinverseofLTPBatameasurementtemperatureof1000C.Mizusakietal.assumedthatchargetransferinnottherate-determiningreactioninthedevelopmentoftheirmodel[ 34 ].Kuznecovetal.developedamodelwhichassumesthatoxygenreductiontakesplaceeverywhereontheLSMsurface,i.e.chargetransferisnotlimitedtothetriplephaseboundary[ 102 103 ].Inthisscenario,polarizationresistanceispredictedtobeproportionalto(LTPB)1.Thisdependencewasexplainedusingmodelswhichwerebasedonsurfacediusionlimitation.Atloweroperatingtemperatures,reactionkineticsassociatedwiththechargetransferreactionmaybecometheratelimitingstep.Apowerlawdependencecanalsobepredictedifweconsiderthereactionkineticsassociatedwiththechargetransferreaction.Foragivenchemicalreaction,therateofthereaction,,isdependentontheconcentrationofthereactingspeciescaAandcbB. Afteradsorptionofgasmoleculesonthesurfaceofthecathode,achargetransferreactionoccursatthecathode/electrolyteinterface.J.Nowotnyetal.outlinedthevariouspossibleadsorption-chargetransferreactioncombinationsinreference[ 49 ](seeFigure 2-1 ).Letusassume,fornow,thatadsorptionandchargetransferoccuraccordingtothefollowing,respective,reactions. 91

PAGE 92

2O2;ads+VoOxo+s(6{4) InEquation 6{4 ,sisasurfacesite.Thepreviousequationsdescribethemolecularadsorptionofoxygenfollowedbyaseparatechargetransferstep.Therateofthechargetransferreactionindicateshowquicklychargedspeciesarebeingtransferredacrossthecathode/electrolyteinterface.Thisexchangeofchargedspeciesdeterminestheexchangecurrent,Io.Itisconvenienttoconsidertheexchangecurrentdensity(io),whichistheexchangecurrentperunitareaandisgivenbythefollowingrelationship. InEquation 6{5 ,Aintistheplanarareaofthecathode/electrolyteinterface,i.e.64mm2inthiswork. Iftheindividualspeciesofthechargetransferreactionaretreatedasreactantsandproducts,thentheexchangecurrentdensitycanbeexpressedinEquation 6{6 InEquation 6{6 ,theforwardandreverserateconstantsaregivenbykfandkr,andQaccountsforbalanceofunits.InEquation 6{6 Theinterfacialreactionisimpededbyachargetransferresistance,Rct,whichunderequilibriumconditionshasbeenshowntobeinverselyproportionaltoIo. InEquation 6{7 ,Risthegasconstant,T,n,andFhavetheirusualmeaning. 92

PAGE 93

6{5 intothechargetransferequation,Equation 6{7 ,givesthefollowingexpressionforRct. UsingiofromEquation 6{6 ,wecanexpressRCTasdescribedinEquation 6{9 InEquation 6{9 ,Q0=QAintnF=RT. Theamountofspecies88i00availabletoreactperunitarea((ci)TPB)islimitedbyLTPBperunitarea.InordertorelatethenumberofspeciesinthevicinityoftheTPBtothebulkconcentrationofspeciesintheirrespectivephases,weneedtomultiplytheconcentrationofspeciesbytheTPBvolumeofeachrespectivephase. InEquation 6{10 ,ATPB;iisthecross-sectionalareaoftheTPBforeachoftherespective88i00phases.Substitutingtheeectiveconcentrationsofactivespecies[i]TPBintoEquation 6{9 givesthefollowingexpressionforthedependenceofRctonLTPB. Theexponentialquantity(n+m+p)givesthereactionorderdependenceonLTPB.Inchemicalreactionsthereactionordercanbegivenbythecoecientsinthebalancedchemicalequation.IfweassumethatthechargetransferreactionisoftheformofEquation( 6{4 )andthattheexponentialterms,m,n,andparegivenbythecoecientsinEquation 6{4 ,thenareactionorderof-3.5ispredictedandRct/(LTPB)3:5whichisexactlywhatwasobserved. Thusfarinthiswork,wehavereferredtotheintermediatefrequencyprocessasadsorptionratherthan88dissociativeadsorption00asoftenreportedintheeld.The 93

PAGE 94

6-10 .Thedatawasttoapower-lawrelationshipresultinginEquation 6{12 Becauseimpedancespectroscopyisanelectrochemicaltechnique,itcannotdetectthepresenceofadsorbedspeciesunlesstheyparticipateintheelectrochemicalreaction.Ifweassumethatthechargetransferreaction,Equation 6{4 ,istheratelimitingstepthenthespeciesgeneratedintheadsorptioncanonlybedetectedafterthechargetransferreactionoccurs,i.e.theratewedetectgenerationofadsorbedspeciesislimitedbytherateofthechargetransferreaction.Everytimeadsorption(Equation 6{3 )occurs,anO2;adsisgenerated.ForeachO2;adsgenerated,however,thechargetransferreaction(Equation 6{4 )canoccurtwice.Wecanconcludethatifthereactionischargetransferlimited,andadsorptiondoesnotoccurdissociatively,therateoftheadsorptionreactionisonehalfthatofthechargetransferreaction.WeexpectthereactionorderdependenceofadsorptionRponLTPBtobesmallerthan-3.5andinfactwendittobe-1.76.Itshouldbepointedoutthattheprocesswearereferringtoasadsorptionisdiculttodistinguishfromarrivalofadsorbedmolecularoxygentothetriplephaseboundarybyothermeans.Ifsurfacediusionissignicantatthismeasurementtemperature,oxygenmoleculescannotonlyarriveatthereactionsitebyadsorption,butalsobysurfacediusionafteradsorptionelsewhereonthecathodesurface,couplingadsorptionandsurfacediusion.Otherauthorshavereportedadependenceofoverallpolarizationresistanceon(LTPB)1whenchargetransferisnottheratelimitingstep.Ourdependenceof1.76indicatesthatthereactionofadsorbedmoleculesisreducedwhenchargetransferisratelimiting.Ourndingsarenotinconsistentwithothersinthatourmeasurementswerecarriedoutat800C 94

PAGE 95

RelationofchargetransferandadsorptionpolarizationresistancedeterminedfromthenestedequivalentcircuittoLTPB(measuredinairat800C). whilemuchofthepreviousworkhasbeencarriedoutatnear1000C.Atthelowermeasurementtemperature,theadditionalpolarizationcomponentsbecomelargeraidingindeconvolutionandthechargetransferreactionmaybeslowed,eectivelychangingtheratelimitingstep.Varyingoxygenpartialpressureandtemperaturewillhaveaneectofchangingtheratelimitingstepandfutureworkwillinvestigatetheinuenceofoxygenpartialpressureandtemperatureonthedeterminedreactionorder. 6-11 showsthedependenceofR1(chargetransfer)andR2(adsorptionrelated)fromFigure 6-6 onLTPB.Curvettingofthedatarevealedpowerdependencies. 95

PAGE 96

50 ]forthedependenceoftotalresistanceat800C.ThisresultisnotunexpectedsinceR2hasthelargermagnitudeofthetwoprocessesandmakesupthemajorityofthecathodicimpedance. Previously,weassumedthatmolecularadsorptionledtoachargelessadsorbedspeciesparticipatinginthechargetransferreactionoccurringattheTPB.WenowconsiderthepossibilitythatoxygenadsorptionleadstoanegativelychargedintermediatewhichisoneofthemanypossiblereactionsproposedbyNowotnyetal.[ 49 ].ThecorrespondingadsorptionandchargetransferreactionsareexpressedinEquations 6{15 and 6{16 ,respectively. 1 2O02;ads+Vo+e0Oxo+s(6{16) Inthesereactions,theadsorbedspeciespossessesanegativecharge.Becauseofthis(andanylatticedistortionsassociatedwithadsorption),theindividualadsorbedspeciesarerepelledfromoneanother.Forthisreason,theamountoflowenergysitesavailablemaybereducedascomparedtoanunchargedadsorbedspecies.Theconcentrationofadsorptionsitesmaydirectlyinuencetherateofthereaction.InaworkbyMizusakietal.sites(s)wereusedinamodelwhichrelatestherateofthedissociativeadsorptionreactiontothecurrentdensityoftheelectrode[ 117 ].Inthework,theauthorswereinterestedinthetotalconductivity.Inthiswork,bothchargetransferandadsorptionareofinterestandsowemustconsidertheindividualreactionscorrespondingtochargetransferandadsorptionindependently. Intheprevioussection,anexpressionforexchangecurrentdensitywasderivedbyassumingthattherateofreactionisdirectlyrelatedtotheconcentrationofalloftheindividualspeciesinvolvedinthereaction,includinge0andh.Analternativeapproachisusedinelectrochemistrytolinktheexchangecurrenttotherateofproductionof 96

PAGE 97

6{16 ,ifanO02;adsandaVocombinetoproduceanOxo,ane0isconsumed(oralternatively,ahisproduced).ThisapproachisbasedontheButler-VolmerEquation(Equation 6{17 )whichdependsonboththeforwardandreversereactions. InEquation 6{17 ,R,T,andFhavetheirusualmeaning,sdescribesthesurfaceoverpotential,andaandcaretheanodicandcathodicapparenttransfercoecients,respectively.TheButler-Volmerequationdescribesthedependenceofthethecurrentdensity(in)onanappliedpotential.InEIS,smallACpotentialsareappliedatvariousfrequencies.Inthiswork,theappliedpotentialoscillatesaround0Vandat0V,inapproachesio,theexchangecurrentdensity.Forthisreason,itismostappropriatetoconsideriointhemodelingofthesystem.ThefollowingapproachistypicallyusedinaqueouselectrochemistrywhereoxygenionsareO2insteadofOxoandvacanciesandreactionsitesarenotusuallyconsidered.Inthefollowingdevelopment,wewillusethisformulism,butwilltryandincorporatethesignicanceofvacanciesandreactionssites,whichareimportantinsolidstatesystems. Anexpressionfortheexchangecurrentdensity(io),usingtheformalismofNewman,canbeexpressedbythefollowingrelationafterrearrangingterms[ 118 ]. InEquation 6{18 ,nrepresentsthenumberofchargestransferredinthestep,FisFaraday0sconstant,kcandkaarecathodicandanodicrateconstants,ci;anodicandci;cathodicaretheconcentrationsoftheanodicandcathodicspecies,respectivelyandpiandqiarethecoecientsoftheanodicandcathodicspeciesinthechargetransferreaction(Equation 6{16 ),respectively.Whetheranelectronisconsumedintheforwardreaction(holeisproduced),oranelectronisproducedinthereversereaction,netchargeisowing 97

PAGE 98

Thereisanactivationenergyassociatedwiththisreactionandadierentactivationenergyisassociatedwiththereversereaction,i.e.theproductionofO02;adsandVofromthedissolutionofanOxo.Evenwithnoappliedpotential,thesereactionsmayoccurduetotheinternalenergyofthesystem.Ifapotentialisapplied;however,itwillaectthetwoactivationenergiesindierentmanners,dependingonthedirectionofthebiasandtheparticularsofthesystem.Forinstance,in(La0:8Sr0:2)0:98MnO3,thereareaboutfourtimesasmanyMnxMnasthereareMnMn.BecauseoftheavailabilityofMnxMntochangevalencies,abiaswhichfavorstheproductionofholeswillmoreecientlycreatecurrentthanthereverse.Suchaninuencecanbeaccountedforbytheutilizationofthesymmetryfactor,.UponexaminationofEquation 6{18 ,weseethatif=0,thentheanodicreactantsaredisregardedinthecalculationoftheexchangecurrentdensity.ThissituationcorrespondstoEquation 6{16 proceedingonlyintheforwarddirection.Avalueof0.5representsboththeforwardandreversereactionsoccurringinunison,anequilibriumcondition. Andso,fromEquations 6{18 and 6{16 wehavethefollowingrelations. If=0:io=nFkc[(cO02;ads)0:5cVo]=0:5:io=nFk0:5ck0:5a[(cO02;ads)0:25(cVo)0:5(cOo)0:5(cs)0:5](6{19) Combiningequation 6{19 andequation 6{8 ,wehaveexpressionsforRct. If=0:Rct=1 Thequantitycidescribestheconcentrationofthei0thspecieswhichisavailabletoparticipateintheinterfacialreaction.Theconcentrationofthei0thspeciesperunitarea 98

PAGE 99

6{10 ApplicationofEquation 6{10 toEquation 6{20 givesadirectrelationshipbetweenRCTandLTPB.Sowehave:if=0, if=0.5, Thus,dependingonhowfarreaction 6{16 isdisplacedfromequilibrium,theexponentialdependenceofRCTonLTPBwillbebetween-1.75and-1.5.Thisisconsistentwiththeobservedtrendofthenesteddata,RCT/(LTPB)1:6. TheadsorptionreactionwhichproducestheintermediatespeciesO02;ads,isexpressedinEquation 6{15 .Applicationoftheexchangecurrentequation(Equation 6{18 )totheadsorptionreactiongivesthefollowingrelations. If=0:io=nFkc[(cO2;g)(cs)]=0:5:io=nF(kcka)0:5[(cO2;g)0:5(cs)0:5(cO02;ads)0:5](6{23) Applyingtheseexchangecurrentdensitiesandtheconcentrationequation(Equation 6{10 )tothechargetransferequation(Equation 6{8 )givesthefollowingrelationshipsbetweenRpandLTPBfortheadsorptionreactiongiveninEquation 6{15 99

PAGE 100

if=0.5, Fromdatadeconvolutionusingthenestedequivalentcircuit,apowerdependenceofadsorptionpolarizationresitanceonLTPB)2:1wasobserved.ThisdependencewasidenticaltothepowerdependencereportedbyFleigfortotalcathodicresistance[ 50 ].Theobservedpowerdependence,-2.1,mostcloselymatchesthe=0casewhichpredictsapowerdependenceof-2.Formanychemicalreactions,hasavaluecloseto0.5.Forthis,apowerdependenceof-1.5ispredictedfortheadsorptionrelatedprocess.If=0,thedissociativeadsorptionreactiondescribedinEquation 6{15 isnotinequilibrium,whichisconsistentwiththeideathatdissociativeadsorptionistheratelimitingstepasreportedbyothers[ 119 ]. 50 ].ThistendencywasexplainedbyFleigbyconsideringbulktransportofionicspeciesthroughthecathode.Forthisbulktransporttooccur,adsorptionmustoccuroneithertheentireporesurfaceareaoranactiveregionoftheporesurfaceareawhichiswithinsomecriticaldistance,ofthecathode/electrolyteinterface.Ineithercase,theareaonwhichadsorptioncanoccurisdirectlyproportionaltothevolumenormalizedporesurfacearea,SV.Alternatively,if 100

PAGE 101

Forsimplicity,wewilltreatonlyoneofthepossibilitieslistedinthepreviousparagraphhere;however,adsorptionforeachofthescenariosshouldbelimitedbySV.Forthemoment,assumeadsorptionoccursaccordingtoEquation 6{15 overtheentiresurfaceareaofthecathode.Inaddition,assumethatsurfacediusionofadsorbedintermediatesisnotalimitingfactor.Thisassumptionisreasonable,asWarburgbehaviorwasnotseeninthevariousimpedanceproles.AnexchangecurrentexistsbasedonthegenerationofthechargedintermediatesO02;ads,whichwillparticipateinthechargetransferreactionattheTPBafterdiusingtoanactivelocation.ThereisaresistancetothiselectrochemicalreactionwhichcanbeexpressedasRads.Theexchangecurrentdensity(io;ads)isnotdirectlydependentonLTPBsinceadsorptionisnotlimitedtotheTPB,likechargetransfer,butmayoccurovertheentiresurfacearea.Thetotalexchangecurrent,however,islimitedtothetotalporesurfaceareaperunitvolume.TheresistancetotheadsorptionreactiondescribedinEquation 6{15 ,canbeexpressedaccordingtoEquation 6{8 .Previously,forchargetransfer,Aintwastheplanargeometricareaofthecathode.Foradsorption,AintmustbereplacedbyAint;pore,whichrepresentsthepore/cathodeinterfacialareaasdescribedinEquation 6{26 6{26 ,SVrepresentsthesurfaceareaperunitvolume,Aintisequalto64mm2,andthecathodethickness(tcathode)isequaltoabout20m.SubstitutionofEquation 6{26 intoEquation 6{8 givesthefollowingrelation. 101

PAGE 102

Relationofadsorptionpolarizationresistancedeterminedfromanestedmodeltosurfaceareaperunitvolume(measuredinairat800C).Redlinerepresentstheactualtanddashedlinerepresentapowerdependenceof-1. Theexchangecurrentdensity,io;adsnowrepresentstheformationofthechargedintermediate,O02;adsasadsorptionoccurs.Figure 6-12 displaysthedependenceofRadsonSV.ThettinginthegurerevealsRads/(SV)1:3,whichisclosetoapowerdependenceof-1,asillustratedbythedashedlineinthegure. ObservationofthetrendlinesinFigure 6-12 revealsthatallofthedatapointsexcepttwoliealongthedashedline.Ifthedependenceofadsorptionpolarizationresistancedoesindeedshowadependenceonporesurfaceareatothe-1power,thenthesameprocesswouldshowadependenceonLTPBtothe-2power,ifthesurfaceareaisproportionaltothetriplephaseboundarylengthsquared.Thisisareasonableassumptioniftheparticlesarerelativelyuniforminsizeandsphericalinshape.ThepowerdependenceobservedinEquation 6{14 andFigure 6-11 canbeexplainedusingbothmodels.Furtherresearchisrequiredtodeterminewhichisvalid. 102

PAGE 103

UseofaseriesVoigtelementmodelledtoapowerdependenceof-3.5forchargetransferresistanceonLTPBandadependenceof-1.75foradsorptionpolarizationresistanceonvolumenormalizedsurfacearea.Theexponentialdependenceof-3.5waspredictedbyapplicationofprinciplesofreactionkineticstoachargetransferstepinvolvingunchargedadsorbedmolecularoxygen(O2;ads).Inthismodel,weassumedthatthecoecientsofthespeciesinthechargetransferreactiondeterminesthepowerdependenceoftherespectivespeciesinthecurrentexchangereaction,theconcentrationofthespeciesabletoparticipateinthecathodicreactionarelinearlydependentontheamountoftriplephaseboundarylengthperunitarea,andthatthereversereactionsarenegligible. Comparisonofelectrochemicalparametersfromthenestedmodeltothemicrostructuraldatarevealedadependenceof-1.6and-2.1forchargetransferandadsorptiononLTPB,respectively.Fortheseprocesses,powerdependencesof-1.5and-2,respectively,werepredictedbyassumingthattheadsorbedintermediateisoftheformO02;ads,theexchangecurrentcanbeexpressedbyEquation 6{18 ,theconcentrationofthespeciesabletoparticipateinthecathodicreactionarelinearlydependentontheamountoftriplephaseboundarylengthperunitarea,andthatthevalueofisequaltozero. Sinceadsorptionpolarizationresistancemakesupthemajorityofthetotalcathodicresistance,thisindividualprocesscanbecomparedtotheresultsofotherswhoreported 103

PAGE 104

50 ]whoseanalysiswasperformedonthesamematerialatthesamemeasurementtemperature.Fleigconcludedthatsincetotalcathodicresistanceisproportionalto(LTPB)2,andscaleslinearlywithcathodethickness,bulkconductivitythroughtheLSMissignicant.WehaveproposedanalternateexplanationforLSMwhichdoesnotdependonbulkionicdiusionthroughLSMwhichhasalowionicconductivityat800C.Otherauthorshaveusedhighermeasurementtemperaturesandproducedresultsinconsistentwithours;however,fewdatapointswereusedtodemonstratearelationshipbetweenresistanceandLTPB.Additionally,itisreportedthatachangefromWarbugbehaviortonon-Warburgbehavioroccursataround800Cindicatingthattheratelimitingstepmayundergoatransitioninthistemperatureregime[ 113 ]TheworksofKuznecovetal.(LSM,950C)andMizusakietal.(LCM,1000C)wereperformedathighertemperatureswerefasterreactionkineticsattheTPBandhigherionicconductivityinLSMareexpected[ 103 ]. Polarizationresistanceoftheadsorptionrelatedprocesswasobservedtohaveapowerdependenceof-1.3onvolumenormalizedporesurfacearea.Adependenceof-1ispredictedbybothmodels,assuminguniformgeometryofparticlesaspreviouslydiscussed.ItwasdemonstratedthatbothinterpretationsoftheadsorptiondatafromthenestedmodelwillpredicttheobserveddependenceofRPonLTPB.Forthisreason,futureworkisnecessarytodeterminethecorrectmodel. 104

PAGE 105

AlthoughtheionicconductivityofLSCFislessthanitselectronicconductivity(0.03vs2.9Scm1at800C)theionicconductivityissignicantlyhigherthanthatofLSM(107Scm1)[ 38 { 40 ].TheactiveregionforcathodicreductionisnolongerrestrictedtotheTPB.Ineect,thecathodicreactioncanoccuratthecathode/gas/electrolyteTPB,thecathode/gasinterface,orthecurrentcollector/gas/cathodeinterface.Thepreferredsiteofthecathodicreactiondependsnotonlyontheionicandelectronicconductivityofthecathode,butalsoonthecatalyticnatureoftheMIECandthecurrentcollector,thethicknessofthecathode,theresistanceofspeciestransferredacrossthevariousinterfaces,andthelocaloxygenconcentrationattheprospectivereactionsite[ 48 ].Thesereactionpathwaysactinparallelandtherefore,thereactionwillproceedinwhatevermannerminimizestotalresistance.BecausetheelectronicconductivityofLSCFissignicantlygreaterthantheionicconductivity,thepreferredreactionpathwayiswithchargetransferoccurringatornearthecathode/gas/electrolyteTPBwhereplentyofoxygen(enoughoxygentoecientlyconverttheelectroniccurrentinthecathodetoioniccurrentintheelectrolyte)ispresent.Iftherequiredioniccurrentisgreaterthanthatwhichthesupplyofoxygenallows(concentrationpolarization),thentheionicconductingpropertiesofLSCFbecomesignicant.Theactivereactionareawillexpandupthesurfaceofthe 105

PAGE 106

51 ]. Forthisreason,inoxygenrichatmospheresanimpedanceprolesimilartothecaseofourelectronicconductingcathode(LSM)isanticipated.OurpreviousresultsshowedthatforLSM,twoelectronicprocesses(chargetransferandadsorption)dominatethecathodicreactioninair.Inlowpartialpressuresofoxygenathirdelectrochemicalprocess,relatedtothebulkdiusionofgaseousoxygenappearsatverylowfrequencies.ForLSCF,inadditiontothesethreeprocesses,anadditionalarcshouldbepresentduetoionictransportthroughtheMIECwheneverthethirdarc,associatedwithconcentrationpolarizationispresent. 3.2 .Inthischapter,however,theLSMwasreplacedwithLSCFsuppliedbyNextechMaterials,Ltd.Sinteringwasperformedforonehourattemperaturesrangingfrom800to1150C.Sinteringattemperaturesaboveandbelow1000CwasperformedinLindberg/Bluehighandlowtemperatureboxfurnaces,respectively.Argonandairwerecombinedtoproduceaowrateof100sccmforpO2slessthan0.21%.ForpO2sgreaterthan0.21%,oxygenmixedwithargonwasusedratherthanair.High-frequencyartifactsinthedatawereaccountedforasdescribedinsection 3.3.2 7-1 (a)and(b)showNyquistplotsmeasuredinairat700CforsymmetricLSCFonYSZsamplessinteredathighandlowtemperatures,respectively.Thecorrespondingimaginaryimpedancevs.frequencyplotisshowninFigure 7-2 .Forsinteringtemperaturesabove900Cthemagnitudeoftheimpedanceproleincreasesassinteringtemperatureisincreased,asshowninFigure 7-1 (a).Fortemperaturesbelow900C,thepolarizationresistancemagnitudedoesnotdecreasesignicantlyassinteringtemperatureisreduced.Assinteringtemperatureisreduced,theform 106

PAGE 107

Impedanceresponseoflanthanumstrontiumcobaltironoxide(LSCF)onYSZinairatvarioussinteringtemperatures.a)Highsinteringtemperatures.b)Lowsinteringtemperatures. Figure7-2. ImaginaryimpedanceversusfrequencyforLSCFmeasuredat700Cinairatvarioussinteringtemperatures. 107

PAGE 108

ApplicationofaseriesVoigtelementbasedequivalentcircuittoLSCFsamplesinteredat1000Candmeasuredat700Cinair. oftheproledegradesandtheshapebecomesmoredepressedasshowninFigure 7-1 (b).Thisdepressionistypicallyviewedasadistributionoftimeconstantsamongthesignicantelectrochemicalprocessesoccurring.Atthelowestsinteringtemperature,800C,theimpedanceprolemaynotbestableduetotheproximityofsinteringandtestingtemperatures.Therstwell-denedarcoccursat950Candabove950Cthepolarizationresistancemagnitudeincreasesrapidlywithtemperature.Thisprovidesanindicationthat950CistheoptimumsinteringtemperatureforLSCFonYSZ.FromFigure 7-2 ,itisclearthatthe1000Csamplehastwofrequencypeaks,oneataround30Hz,andoneataround104Hz.Thistrendcontinuesassinteringtemperatureisincreasedasindicatedbytheasymmetricnatureoftheproles.By1150Cbothprocessesaresimilarinmagnitudeandfrequencyandthusdiculttodistinguish. AseriesVoigtelementmodelwasusedtottheimpedancedataofthe1000Csinteredsample.Figure 7-3 displaysthepolarizationcontributionswhichmakeuptheimpedanceproleofthe1000Csinteredsample,measuredat700Cinair.LikeinLSM,thereisasmallmagnitudehigh-frequencyprocessandalargermagnitude 108

PAGE 109

ParametersdeterminedfromequivalentcircuitttingforLSCFinair,measuredat700C. intermediate-frequencyprocess.AsimilarttingwasperformedforeachoftheprolesshowninFigure 7-2 .Atthelowestsinteringtemperatures,thettingwascomplicatedbythedepressednatureoftheprole,whileatthehighestsinteringtemperaturesdeconvolutionwasdicultduetothelowrelaxationfrequenciesofsomeoftheprocesses.Theshapeoftheprolesandthequalityofthetting,leadsustoassumethattwoprocessescontributetotheoverallimpedanceprolewhenmeasuredinairat700C,particularlyatintermediatesinteringtemperaturessuchas1000C. Figure 7-4 showsthepolarizationresistancesobtainedfromttingeachoftheimpedanceprolesshowninFigure 7-2 withaseriesVoigtelementmodel.Thegureshowsaminimuminchargetransferpolarizationresistanceat900Candaminimuminadsorptionpolarizationresistanceat850C.Atsinteringtemperaturesabove900C,adsorptionpolarizationresistancebecomeslargerinmagnitudethanchargetransfer,whilebelow900Cthetrendislesspronounced.Itislikelythatby950CgoodinterparticularadhesionbetweentheindividualLSCFparticlesandadhesionbetweentheLSCFandtheYSZhasbeenachieved.Sinteringabove950Conlydegradesthecathodeasevidenced 109

PAGE 110

The950CsinteredsamplewasimpedancetestedinavarietyofpO2sat700C.Figures 7-5 (a)and(b)displaytheresultsoftheimpedancetestingbrokendownintohighandlowerpO2regimes,respectively.FromFigure 7-5 (a),weseethatthereislittlechangeintheimaginaryimpedancevs.frequencyplotfrom58to10%oxygen.Evidently,themechanismofthecathodicreactionisunchangedbyvariationsinpO2concentrationifoxygenisstillquiteabundant.ThereisaslightincreaseinpolarizationresistanceaspO2isdroppedfromthehighvalueof58to10%oxygen.At3%oxygen,webegintoseeadeviationfromthesimpletwoprocessproleoccurringathigherpO2s.By0.67%oxygen,wecanclearlyseetheformationoftwolowfrequencyprocesswhicharenotpresentabove10%.Deconvolutionofthesamplemeasuredat0.09%O2isshowninFigure 7-6 .Inthegure,fourcathodicprocessesareapparent.ForLSM,wesawonenewprocessatlowpartialpressuresofoxygenattributedtoconcentrationpolarizationandrelatedtobulkgasdiusiontothereactionsite.ForLSCF,twolowfrequencyprocessesappearsimultaneously.ItislikelythatoneprocessisduetoconcentrationpolarizationcreatedbythelackofoxygenmoleculesavailableatthereactionzoneandthesecondprocessisrelatedtoLSCFcompensatingforthisinavailablilityofoxygenmolecules.Thetwohigh-frequencycathodicprocessesdonotappeartobesignicantlyaectedbythechangeinoxygenconcentrationfrom3%to0.32%.Below0.32%,thelow-frequencyprocessescontinuestobecomemorepronunced,particularlythelowestfrequencyonewhichdisplaysasharpdependenceonpO2.Surprisingly,inthisregime,theoverallmagnitudeofthehighest-frequencypolarizationresistance(chargetransfer)decreasesaspO2decreases.Thistrendisoppositetheeectseenwhenoxygenisabundant.Asconcentrationpolarizationbecomesmoreprominent,thereactionmechanismshiftsinsuch 110

PAGE 111

Impedanceresponseatvariousoxygenpartialpressuresof950CsinteredLSCFonYSZmeasuredat700C.a)Highoxygenpartialpressures.b)Lowoxygenpartialpressures. 111

PAGE 112

ApplicationofaseriesVoigtelementbasedequivalentcircuittoLSCFsamplesinteredat950Candmeasuredat700Cat0.09%O2. awaythatminimizesthecontributionoftheoriginalchargetransferreaction,possiblyduetoVoformation. TheimpedanceprolesofFigures 7-5 (a)and(b)werettedusingafourVoigtelementbasedseriesmodelwiththeresultsdisplayedinTable 7-1 andFigures 7-7 through 7-9 .Asmentionedpreviously,thebulkdiusionprocessisnotapparentathigherpO2sandthereforemodelparametersforthebulkdiusionprocessbeginatlowerpO2s.Figure 7-7 showstheseriesresistancecontribution.UnlikeLSM,LSCFshowsanincreasingohmicresistance(Rs)aspO2isdecreased.FittingthedatapointstoEquation 4{3 givesadependenceofohmicresistanceonpO2tothe0.054power.Asmentionedinthebackgroundsection,theelectronicconductivityofLSCFiscreatedbyformationofholeswhenSrisincorporatedonaLasiteinthelattice.AspO2isreduced,anincreasingnumberofvacanciesareformedreducingtheconcentrationofholes;therefore,theionicconductivitygoesupwhiletheelectronicconductivitygoesdown.ThisdecreaseinelectricconductivityoftheMIECcausesthetrendseeninFigure 7-7 112

PAGE 113

Polarizationresistancevaluesinsforvariouselementarystepsofthecathodicreactioninlanthanumstrontiumcobaltironoxidesamplessinteredat950Candmeasuredat700Catvariousoxygenpartialpressures. pO2(%) 9.16 1.87 5.52 -41.0 9.50 1.86 5.80 -21.0 9.76 1.83 6.27 -12.5 10.03 2.06 6.85 -3.0 10.74 1.97 7.71 0.71 -1.0 10.73 1.06 7.87 1.51 0.530.67 11.88 2.12 8.49 2.07 0.730.09 12.02 0.92 7.75 2.44 6.240.003 11.95 0.66 5.62 3.67 69.69 Figure7-7. OhmicseriespolarizationresistancefrommodelttingforLSCFsinteredat950Candmeasuredat700CasafunctionofpO2. 113

PAGE 114

43 ].TheelectroneutralityconditioninLSCFisgiveninEquation 7{1 ValencychangesamongtheFeandCoionsaccountforn=[M0M]andp=[MM],while[Sr0La]isequalto0.2.Alsoin[ 43 ],aplotofconductivity(measuredby4pointprobe)onpO2revealsadependenceofabout0.094inthehighpartialpressureregimeat800C.FromthedataofBucheretal.reportedaconductivitypowerdependenceof0.097at700Ccanbecalculated[ 120 ].Becausetheholeisthemajorcarrier,thepowerdependenceofhonpO2shouldalsobeabout0.097.Thepowerdependenceobserved,0.054,closetothisvalue. Forchargetransfer,thepowerdependenceobserved,0.077,isalsoclosetothepowerdependenceofionicconductivityofLSCFreportedbyWang.ThisvaluehasgreatererrorduetotheverylowRvaluereported.AchargetransferresistanceindependentofpO2wouldindicatethatthisprocessoccursinasimilarfashionasforpurelyelectronicallyconductingLSM.However,VoformationwouldreduceRctasisobserved.ThedecreaseinRctatlowpO2salsomayindicatethatthisprocessbecomesincreasinglyinsignicantastheionicpathwaythroughtheMIECbulkbecomesfavorable. Theintermediate-frequencyarcwaspreviouslyattributedtoanadsorptionrelatedprocess.ForLSCF,thisprocessshowstwodistinctregimes.Atpartialpressuresabove0.32%,intermediate-frequencypolarizationresistancedecreasesaspO2increases,whilebelow0.32%(approachingconcentrationpolarization),thepolarizationresistancedecreasesaspO2decreases.Thecorrespondingpowerdependenciesare-0.086and0.099forthehighandlowpO2regimes,respectively.ForhighpO2,thereactionsislikelyconnedtothevicinityoftheTPBandtherefore,weobserveanegativepO2powerdependencelikethatovservedinLSM(-0.15foradsorption/surfacediusion).Inthe 114

PAGE 115

High-frequencypolarizationresistancesasafunctionofpO2forLSCFonYSZsinteredat950Candmeasuredat700C.a)ChargetransferRP.b)AdsorptionRP. 115

PAGE 116

Figure 7-9 (a)showsthepartialpressuredependenceoftheionicprocess.ThisprocesswasnotseeninLSMatlowpO2s,andsosothisprocessmustbedirectlyrelatedtotheionicconductivityoftheMIEC.Therearethreeregimes,twoofwhicharevisibleinFigure 7-9 :1)pO2>3%oxygen,2)3%oxygen>pO2>0.67%oxygen,and3)pO2<0.67%oxygen.Intherstregime,anyresistanceassociatedwiththisprocessistwosmalltoanalyze(thereforenodatapointsabove3%oxygen).InthisregimeactivationpolarizationdominatesandthecathodicreactionisconnedtotheTPB,likeinLSM.Inregimestwoandthree,thevacancyconcentrationintheMIEChasincreasedenoughtomakeionicconductivitythroughtheMIECbulkacompetitivepathway.AsmentionedneartheendofSection 2.2.3 ,ionicconductivityinLSCFdisplaysamaximumataround102atm[ 42 ].Belowthisconcentration,asharpdropoinionicconductivityexistsasreportedbyWangetal.[ 43 ].Inregime2,theionicconductivityintheMIECisatamaximumandsobulkprocessesintheMIECarefavorable.Inregime3,theionicconductivityintheMIECdecreasessignicantly,possiblyduetodefectassociation.Inaddition,thevacancyconcentrationattheMIEC/gasinterfaceishigh,soreactioncanoccurassoonastheholesarriveatapossiblereactionsite.ThisprocessislimitedbytheowofhthroughtheMIEC.SincetheholeistheprimarycarrierinLSCF,theconductivitydependenceofLSCFonpO2shouldmatchthepolarizationresistanceofthisprocess.TheobserveddependencyofRPonpO2,0.095,isinverycloseagreementwiththendingsofWangetal.andBucheretal.fromwhosedatapowerdependenciesofconductivityonpO2of0.094,and0.097canbemeasured,respectively. Inregimetwo,averystrongdependenceonpO2isobserved.Thisisexplainedbyconsideringthatthisprocessstepeectivelycompeteswiththebulkgasdiusion 116

PAGE 117

Low-frequencyolarizationresistanceasafunctionofpO2forLSCFonYSZsinteredat950Candmeasuredat700C.a)MIECspecicprocessRP.b)oxygengasdiusionRP. 117

PAGE 118

7-9 (b)).AtallpO2s,theresistanceofthisprocessislessthanthepolarizationresistanceassociatedwithbulkgasdiusionasseeninFigures 7-9 (a)and(b).AtlowpO2s,aconcentrationgradientisformedinthevicinityoftheTPBandtocompensate,thebulkpaththroughtheMIECbecomesactive,reducingthetotalresistance.AspO2increases,lessofaconcentrationgradientisformedandthebulkgasdiusionresistancedecreases;reactionviatheTPBbecomesfavorable.Ifnoconcentrationgradientexists,thereactionshouldtakeplacecompletelyinthevicinityoftheTPBandsincetheelectronicconductivityismuchhigherthanionicconductivityinLSCFandthealternatereactionpathway(throughthebulkoftheMIEC)isnotfavorable.Themeasuredvalueof-0.86forthedependenceofbulkgaspolarizationresistanceonpO2isclosetothevalueofunityseenforthebulkdiusionprocessintheelectronicconductingcathode.ThereasonforthediminishedvaluecouldberelatedtothefactthattrueconcentrationpolarizationmaynotbeachievedbecauseoxygenmoleculesarenotstrictlyprovidedfromregionsintheimmediateTPBvicinityduetothecontributionoftheionicprocess. ThedependencesofthevariousprocessesonmeasurementtemperatureandoxygenpartialpressureissummarizedinTable 7-2 .TrendingofthetotalresistancemeasuredbyimpedancewasreportedbyMurrayetal.andisincludedinthetable[ 58 ].ThetrendofthetotalresistancefollowsthethatofthelargestRPoftheindividualprocesses.AtlowpO2s,MurrayreportedadiscontinuityintotalresistancedependenceonpO2andspeculatedthatdierentmechanismslimitthecathodeperformanceindierentpO2regimes.Ourresultssupportthathypothesis. Figure 7-10 displaystheeectsofsinteringtemperatureontheelectrochemicalprocessesoccurringat700Cand0.09%oxygen.Thispartialpressurewaschosentoputthecathodicreductionreactioninthetheconcentrationpolarizationregime(lessthan0.67%oxygen)discussedabove.LikeinFigure 7-4 thereisahighandalowsinteringtemperatureregimeandtheyseparateataround950C.Focusingrstonsintering 118

PAGE 119

Propertiesofthevariouscathodicprocessesinlanthanumstrontiumcobaltironoxide.TheresultsofMurrayetal.[ 58 ]fortotalresistanceareincluded.ThesmallpO2dependenciesobservedareclosetothepO2dependencyoftheholeconcentrationinLSCF(0.094)reportedbyWangetal.[ 43 ]andtheionicconductivityofLSCFathigherpO2s(0.097)reportedbyBucheretal.[ 120 ]. pO2<0.001atm Ea(air) Ea(0.09%O2) Ohmic -0.054 -0.054 -0.40 -0.38Chargetransfer 0.077 0.077 -1.72 -1.63Adsorption -0.086 0.098 -1.41 -1.56MIECprocess -0.71 -0.095 n/a -1.50Bulkgasdiusion -0.85 n/a n/a 0.50TotalRP-Murray -0.91 -0.038 -1.63 ParametersfrommodelttingforLSCFat0.09%oxygen,measuredat700Casafunctionofsinteringtemperature. 119

PAGE 120

7-4 Figures 7-11 (a)and(b)showtheactivationenergiesdeterminedbyvaryingthemeasurementtemperatureofthe950Csinteredsamplemeasuredinairandat0.09%oxygen,respectively.Theactivationenergyforchargetransferis-1.72and-1.63eVinairandat0.09%oxygen,respectively,whiletheactivationenergyforadsorptionis-1.56and-1.41eVinairandat0.09%oxygen,respectively.ActivationenergiesfortheMIECspecic(-1.50eV)andbulkdiusion(0.50eV)werealsodeterminedforthelow 120

PAGE 121

Activationenergiesofvariouselectrochemicalprocessesfor950CsinteredLSCFonYSZ.a)Measuredinair.b)Measuredat0.09%oxygen. 121

PAGE 122

4.3 .ThepositivecorrelationwithmeasurementtemperatureseenforbulkgasdiusioninFigure 7-11 (b)maybecausedbyanincreasingoxygengradientformedinthevicinityoftheTPBsincetheotherprocessesoccurmoreecientlyathighertemperatures. AspO2isreduced,analternatereactionpathwayinvolvingionictransportthroughthebulkoftheMIECbecomesmoresignicant.Thischangeismanifestedbytheformationoftwoadditionallow-frequencyarcsintheimpedanceprole.LSM,ontheotherhand,presentsonlyonenewarc,relatedtothebulkgasdiusionofoxygen,intheconcentrationpolarizationregime.ThesecondarcisamostlikelyaresultofasurfaceexchangeprocessatthegasLSCFinterfacewhichleadstoapathwayforionicconductivitythroughtheMIECbulk.AtlowpO2s,thispathwaybecomesmorefavorablebecause1)aconcentrationgradientformsdepletingtheregionneartheTPBofmolecularoxygenand2)anincreasednumberofoxygenvacancies(leadingtohigherionicconductivity)formintheMIECduetothelowpO2oftheambientgas.AtlowpO2s,theactivationenergiesofchargetransferandadsorptionare-1.63and-1.56eV,respectively,whilefortheprocessspecictotheMIECandbulkgasdiusion,theactivationenergiesare-1.50and0.50eV,respectively. 122

PAGE 123

123

PAGE 124

Theelectrochemicalpropertiesofthecathodicreactionwereinvestigatedbyimpedancespectroscopyandothercharacterizationtechniqueswithagoalofidentifyingthesignicantindividualprocesses.Inordertoaccomplishthistask,high-frequencyimpedancedatahadtobeanalyzed.Unfortunately,thequalityofthisdatawasdiminishedbyhigh-frequencyinductiveartifacts.TheinuenceoftheseartifactsonimpedancedatawasanalyzedanditwasfoundthatseveraldatapointsatfrequencieslowerthantheZr-axishadtoberemovedfromtherawdataiftherawdatawastobeused.Thequalityofthedatawasimprovedbyttingthehigh-frequencyportionofthedata(whichwasshowntobeinductiveinnature)toZj=j!Landsubtractingtheresultfromtherawdataovertheentirefrequencyrange.Performingthisoperationincreasedtheamountofusabledatainthehigh-frequencyregimebyanorderofmagnitudeallowinganalysisoffastoccuringelectrochemicalprocesses. Itwasfoundthattheinairathightemperaturestwoprocessesaremostsignicant,chargetransferandadsorption.Thedependenciesofthesetwoprocessesonmeasurementtemperatureandoxygenpartialpressurewereinvestigatedinchapterthree.Itwasfoundthatthechargetransferresistanceissmallerinmagnitudethantheadsorptionrelatedprocess.Inadditiontothesetwoprocesses,abulkdiusionrelatedprocessbecomessignicantatlowpartialpressuresofoxygen.PolarizationresistanceactivationenergiesandtimeconstantsaregeneratedfromthemodelparametersandgiveninTable 4-1 Thecathodicreactionwasfoundtobedramaticallyinuencedbythesinterapplied.Inadditiontoalteringthemicrostructure,over-sinteringcanalsocausetheformationoftertiaryphases.Theelectrochemicalreactionwasdrasticallyinhibitedatthehighestsinteringtemperatures.Atthesesinteringtemperatures,itwasshownthatlanthanumzirconate,aninsulatinglayerhadformedatthecathode/electrolyteinterface. Toseparatetheeectsofmicrostructurefromtertiaryphaseformation,sinteringatlower(butstillhigh)temperatureofnon-stoichiometricLSMwasperformed.SEM/FIB 124

PAGE 125

ThestudywasextendedtoLSCFinthenalchapter.ItwasfoundthatthecathodicreactiononLSCFbehavessimilarlytoLSMathighpartialpressuresofoxygen.IncomparisontoLSM,LSCFhadlargermagnitudeactivationenergiesforchargetransferandtheadsorptionrelatedprocess.Atlowerpartialpressuresofoxygen,theionicreactionpathwaybecomessignicantandanarcintheimpedanceprolenotseeninLSMappears.LSCFsamplesweresinteredatvarioustemperaturesandcorrespondingevolutionoftheimpedanceprolewasexamined.Thedatatakeninthischapteristobecomparedtomicrostructuralinformationfromthesampleswhichisyettobeextracted. 125

PAGE 126

FigureA-1. TunnellingelectronmicroscopyimageofLSMonYSZinterface,withEnergydispersivespectrometry(EDS)prolesinset.ForEDSproles,Blue=Zirconium,Red=Manganese,Green=Yttrium,Purple=Lanthanum,Yellow=Strontium.(CourtesyofMarkClark) TheinterfaceofanLSM/YSZsamplesinteredat1400Cfor48hourswasanalyzedusingenergydispersivespectroscopy(EDS)andtunnelingelectronmicroscopy(TEM)byDr.MarkClark.ATEMimagewithsuperimposedEDSlinescansoftheLSM/YSZinterfaceisdisplayedinFigure A-1 .Thelinescansindicatethattheamountofdiusioniselementspecic;interfacialporesisnotlikelythecauseoftheabruptproleseeninFigure 5-9 (b).Theabruptnessofthemanganeseprolecomparedtothelanthanumproleindicatingthatlanthanum,butnotmanganesefromtheelectrodeisdiusingintotheYSZ.Likethelanthanumprole,theconcentrationproleofzirconiumalsoexhibitsagradualdecrease,thoughthistimetheconcentrationgradientisdecreasing 126

PAGE 127

Selected-areadiractionpatterns(SADPs)oftheelectrode,Figure A-2 (a),andelectrolyte,Figure A-2 (b)wereattainedusinga200kVacceleratingvoltage,acameralengthof20cm.Comparisonoftheguresallowsacleardistinctionbetweentheseregions.Theelectrolytedisplayedadiractionpattrnindicativeofanfcccrystalwithabeamdirectionalongthe[1 1 2]zoneaxis.Thepatternoftheelectrodewasnotastrivialduetotheelectrode'scomplexcrystalstructure.Thepatternofthetransitionregion,Figure A-2 (c),wasidenticaltothatoftheelectrolyte,suggestingthatthetransitionalregionisformedbythediusionoflanthanumfromtheLSMregionintotheYSZregion,preservingthecrystalstructureoftheYSZ.Anotherpossibilityisthatthetransitionregionexhibitedaphasechange,butthenewphasewasalsoofcubicbodystructurewithasimilarlatticeconstant. 127

PAGE 128

Selected-areadiractionpatternsforLSM,YSZ,andthetransitionalregion.Patternsattainedforthe[1 1 2]beamdirectionwitha200kVacceleratingvoltage.a)DiractionpatternforLSM.b)DiractionpatternforYSZ.c)Diractionpatternofthetransitionalregion. 128

PAGE 129

FigureB-1. SetupofFIB/SEMindicatingalignmentofionandelectronbeamwithsample. AdualbeamFIB/SEM(FEIStrataDB235)wasusedtocreateathree-dimensionalimageofthemicrostructureofthesymmetricsamplesasdescribedbelow.Thesymmetricsamplesweremountedona45aluminummount,whichwastiltedanother7sothatthefaceofthesymmetriccathodewasparalleltotheionbeamasshowninFigure B-1 Beforeexposureofthesampletotheionandelectronbeams,thesymmetricsamplesweresputtercoatedwithplatinumtominimizechargingandprotectthesample.ThedualbeamFIB/SEMwasusedtoablatesuccessivelayersofspeciedthicknesswithSEMimagingaftereachablation.Theseuniformlyspaced2-Dimageswerethenaligned,producinga3-Dimage.Serialmillinginthez-directionwasconductedwithstepsofabout50nmpersliceusingaGa+ionbeam.About30imagesweretakenpersample.AftereachFIBslice,SEMimagingwasperformedatamagnicationof12,000X.Thismill-image-millprocedurewasrepeatedfromtheplatinumprotectivesurfacecoating,throughthecathodeanddowntothedenseYSZelectrolyte.TheSEMimagesweretakenat38withrespecttothesamplefacenormalresultinginelongationoftherawimagesin 129

PAGE 130

Porosity(p),volumenormalizedporesurfacearea(SV),andTPBlengthvalueswerecalculatedateachtemperature,whiletortuosity()valueswerecalculatedfromthe3-Ddataattainedatselectedtemperatures.Porositywascalculatedfromtheporearea/totalareaineachSEMimage.Thecalculationwasrepeatedforallslicesinthesampleandanaverageporositywasattained.TheseresultsareplottedinFigures 6-2 and 6-3 Theporesurfaceareareportedisnormalizedperunitvolume.From[ 121 122 ],theporesurfacearea(SV)perunitvolumecanbecalculatedaccordingtoEquation B{1 L3=2PL(B{1) InEquation B{1 ,dSisthesurfaceareaelement,L3istheunitvolume,andPListhenumberofphasechanges(gastosolid)perunitlengthandwascountedmanuallyfromeachoftheSEMimagesthroughthebulkofthecathode. Thetriplephaseboundarylength(LTPB)wascalculatedbyapplicationofthefollowingequation[ 121 122 ]. ForLTPB,PLwascountedfromthepore/LSMphasechangesperunitlengthintheSEMimagesattheLSM/YSZinterface.TheunitsforthecalculationofLTPBandSVarem1,whichisdimensionallyaccurateforalengthnormalizedperunitsurfaceareaandanareanormalizedperunitvolume. ThetortuositywastheonlymicrostructuralparameterthatwasnotattainedforeachSEMandaveragedmakinguseoftheuniformlyspacedFIBslices.Tortuositywascalculatedbyestimatingthelengthagasparticlemusttravelasitdepartstheimpinginggasowandtravelstotheelectrolytedividedbythestraight-linethickness.Themethod 130

PAGE 131

B{3 canbeusedtoestimatethetotaldistancetraveledbyaparticle(L)usingthePythagoreanTheorem. InEquation B{3 (xn;yn;zn)arethecoordinatesofthenthpointusedandthereareNtotalpointsdetermined.Onceattained,Lcanbedividedbythestraight-linedistancetogivethetortuosity. ThevolumenormalizedporesurfaceareawascalculatedasdescribedinEquation B{1 usingthreeSEMimagesforeachsinteringtemperature.Foreachimagealinewastakeninthreedirectionsforatotalofninemeasurementspersinteringtemperature.ThetemperaturedependenceoftheaverageporesurfaceareaandcorrespondingstandarddeviationasafunctionoftemperatureisplottedinFigure 6-2 (a).Fortheareanormalizedtriplephaseboundarylength,asinglelineneartheinterfacewastakenfromeachofthreeSEMimagespersinteringtemperature.TheresultswereaveragedandplottedinFigure 6-2 (b). 131

PAGE 132

[1] M.C.Williams,J.Strakey,W.Sudoval,J.ofPowerSources,159(2006)1241. [2] N.Q.Minh,J.oftheAmericanCeramicSociety76(1993)563. [3] H.-J.Ziock,E.J.Anthony,E.L.Brosha,F.H.Garzon,G.D.Guthrie,A.A.Johnson,A.Kramer,K.S.Lackner,F.Lau,R.Mukundan,N.Nawaz,T.W.Robinson,B.Roop,J.Ruby,B.F.Smith,J.Wang,TechnicalLosAlamosNationalLaboratoryPublicationLA-UR-02-5969,Proceedingsofthe28thInternationalTechnicalConferenceonCoalUtilization&FuelSystems,Clearwater,FL,2003. [4] N.Q.Minh,T.R.Armstrong,J.R.Esopa,J.V.Guiheen,C.Horne,J.J.VanAckeren,in:S.C.Singhal,H.Iwahara(Eds.),Proceedingsofthe3rdInternationalSymposiumonSolidOxideFuelCells,TheElectrochemicalSociety,Inc.,Pennington,NJ,1993,p.801. [5] M.Yashima,M.Kakihana,M.Yoshimura,SolidStateIonics86/88(1996)1131. [6] F.A.Kroger,H.J.Vink,in:SolidStatePhysics,3rdEd.,AcademicPress,NewYork,NY,1956,p1. [7] J.F.BaumardP.Aberlard,in:N.Claussen,M.Ruhle,A.H.Heuer(Eds.),AdvancesinCeramics12,ScienceandTechnologyofZirconiaII,AmericanCeramicSociety,Columbus,OH,1984,p.555. [8] C.B.Choudhary,H.S.Maiti,E.C.Subbarao,in:E.C.Subbarao(Ed.),SolidElectrolytesandTheirApplications,PlenumPress,NY,1980,p.1. [9] E.C.SubbaraoH.S.Maiti,SolidStateIonics11(1984)317. [10] X.Guo,M.Maier,J.Electrochem.Soc.148(2001)E121. [11] J.E.Baurle,J.ofPhys.Chem.Solids30(1969)2657. [12] M.Kleitz,H.Bernard,F.Fernandez,E.Schouler,in:A.H.Heuer,L.W.Hobbs(Eds.),AdvancesinCeramics,vol.3,ScienceandTechnologyofZirconia.AmericanCeramicSociety,Columbus,OH,1981,p.310. [13] M.J.Verkerk,B.J.Middelhuis,A.J.Buggarf,SolidStateIonics6(1982)159. [14] E.P.Butler,R.K.Slotwinski,N.Bonanos,J.Drennan,B.C.H.Steele,in:N.Claussen,M.Ruhle,A.H.Heuer(Eds.),AdvancesinCeramics12,ScienceandTechnologyofZirconiaII,AmericanCeramicSociety,Columbus,OH(1984)p.572. [15] J.H.Kuo,H.U.Anderson,D.M.Sparlin,J.ofSolidStateChemistry87(1990)55. [16] M.Kertesz.,I.Riess,D.S.Tanhauser,R.Langpape,F.J.Rohr,J.ofSolidStateChemistry42(1982)125. 132

PAGE 133

T.Hashimoto,N.Ishizawa,NMizutani,M.Kato,J.ofMaterialsScience23(1988)1102. [18] K.Katayama,T.Ishihara,H.Ohta,S.Takeuchi,Y.EsakiE.Inukai,J.oftheCeramicSocietyofJapan97(1989)1324. [19] A.Hammouche,E.L.Schouler,M.Henault,SolidStateIonics28-30(1988)1205. [20] G.V.SubbaRao,B.M.Wanklyn,C.N.R.Rao,J.ofPhysicalChemistry32(1971)345. [21] N.Q.Minh,T.Takahashi,in:ScienceandTechnologyofCeramicFuelCells,Amsterdam;NewYork:ElsevierScience(1995). [22] A.Hammouche,E.Siebbert,A.Jammou,MaterialsResearchBulletin24(1989)367. [23] H.Taguchi,D.Matsuda,M.Nagao,K.Tanihata,YU.Miyamoto,J.AmericanCeramicSociety75(1992)201. [24] H.Taguchi,D.Matsuda,J.ofMaterialsScienceLetters14(1995)12. [25] A.Endo,M.Ihara,SolidStateIonics86-88(1996)1195. [26] H.S.Maiti,A.Chakraborty,M.K.Paria,in:S.C.Singhal,H.Iwahara(Eds.),Proceedingsofthe3rdInternationalSymposiumonSolidOxideFuelCells,TheElectrochemicalSociety,Inc.Pennington,NJ,1990p.190. [27] A.Chakraborty,P.S.Devi,H.S.Maiti,MaterialsLetters20(1994)63. [28] A.Chakraborty,P.S.Devi,S.Roy,H.S.Maiti,J.ofMaterialsResearch9(1994)986. [29] R.Basu,S.Pratihar,MaterialsLetters32(1997)217. [30] L.W.Tai,P.A.Lessing,J.oftheAmericanCeramicSociety74(1991)5. [31] T.Ishihara,T.Kudo,H.Matsuda,Y.Takita,J.oftheAmericanCeramicSociety77(1994)1682. [32] B.Gharbage,M.Henault,T.Pagnier,A.Hammon,MaterialsResearchBulletin26(1991)1001. [33] L.G.J.deHaart,K.J.deVries,A.P.M.Carvalho,J.R.Frade,F.M.B.Marques,MaterialsResearchBulletin26(1991)507. [34] J.Mizusaki,H.Tagawa,K.Tsuneyoshi,A.Sawata,J.oftheElectrochemicalSociety138(1991)1867. [35] L.-W.Tai,M.M.Nasrallah,H.U.Anderson,D.M.Sparlin,S.R.Sehlin,SolidStateIonics,76(1995)259. 133

PAGE 134

L.-W.Tai,M.M.Nasrallah,H.U.Anderson,D.M.Sparlin,S.R.Sehlin,SolidStateIonics,76(1995)273. [37] Z.Li,M.Behruzi,L.Fuerst,D.Stover,in:S.C.Singhal,H.Iwahara(Eds.),SOFC-III,PV93-4,TheElectrochemicalSociety,Pennington,NJ,1993,p.171. [38] Yasuda,K.Ogasawara,M.Hishinuma,T.Kawada,M.Dokiya,SolidStateIonics86-88(1996)1197. [39] G.Ch.Kostogloudis,Ch.Ftikos,SolidStateIonics126(1999)143. [40] Y.Teraoka,T.Nobunga,K.Okamoto,M.Miura,N.Yamazoe,SolidStateIonics48(1991)207. [41] S.Carter,A.Selcuk,R.J.Chater,J.Kajada,J.A.Kilner,B.C.H.Steele,SolidStateIonics53-56(1992)597. [42] M.Katsuki,S.Wang,M.Dokiya,T.Hashimoto,SolidStateIonics156(2003)453. [43] S.Wang,M.Katsuki,M.Dokiya,T.Hashimoto,SolidStateIonics159(2003)71. [44] K.Tsuneyoshi,K.Mori,A.Sawata,J.Mizusaki,H.Tagawa,SolidStateIonics35(1989)263. [45] A.Hammouche,E.Siebert,A.Hammou,M.Kleitz,A.Caneiro,J.oftheElectrochemistrySociety138(1991)1212. [46] M.Ostergard,M.Mogensen,ElectrochimicaActa38(1993)2015. [47] Y.Takeda,R.Kanno,M.Noda,Y.Tomita,O.Yamamoto,J.oftheElectrochemicalSociety134(1987)2656. [48] M.Liu,J.Electrochem.Soc.145(1998)142. [49] J.Nowotny,T.Bak,M.K.Nowotny,ancC.C.Sorrell,AdvancesinAppliedCeramics104(2005)154. [50] J.Fleig,Annu.Rev.Mater.Res.33(2003)361. [51] S.B.Adler,SolidStateIonics111(1998)125. [52] S.B.Adler,J.ElectrochemSoc.143(1996)3554. [53] X.J.Chen,K.A.Khor,S.H.Chan,J.ofPowerSources123(2003)17. [54] S.P.Jiang,SolidStateIonics,146(2002)1. [55] A.J.Appleby,F.R.Foulkes,FuelCellHandbook,VanNostrandReinhold(1989). 134

PAGE 135

D.Herbstritt,A.Weber,E.Ivers-Tiee,in:U.Stimming,S.C.Singhal,(Eds.),SolidOxideFuelCellsVI,PV99-19,TheElectrochemicalSocietyProceedingsSeries,Penington,NJ,1999,p.972. [57] M.Ostergard,C.Clausen,C.Bagger,M.Mogensen,ElectrochimicaActa40(1995)1971. [58] E.P.Murray,M.J.Server,S.A.Barnett,SolidStateIonics148(2002)27. [59] D.Herbstritt,A.Krugel,A.Weber,E.Ivers-Tiee,ElectrochemicalSocietyProceedings2001-16(2001)943. [60] B.Boukamp,SolidStateIonics20(1986)31. [61] VanDijk,A.J.Burgraaf,Phys.StatusSolidi(a)63(1981)229. [62] J.Jamnik,J.Maier,Phys.Chem.Chem.Phys.3(2001)1668. [63] F.Baumann,J.Fleig,H.Habermeier,J.Maier,SolidStateIonics177(2006)177. [64] J.R.Macdonald,ImpedanceSpectroscopy,JohnWileyandSons,NY(1987). [65] L.Ljung,SystemIdentication,Prentice-Hall,EnglewoodClis(1987). [66] H.Schichlein,M.Feuerstein,ECSproceedings99-19(1999)1069. [67] H.Schichlein,A.Muller,AppliedElectrochemistry32(2002)875. [68] K.Kleveland,M.Einarsrud,C.S.Schmidt,S.Shamsili,S.Faaland,K.Wiik,T.Grande,J.oftheAmericanCeramicSociety82(1999)729. [69] A.Franklin,H.Bruin,PhysicalStatisticalSolids75(1983)647. [70] A.Muller,H.Schichlein,ECSproceedings99-19(1999)925. [71] D.L.Misell,R.J.Sheppard,J.ofPhysicsD:AppliedPhysics6(1973)379. [72] M.Orazem,P.Shukla,M.Membrino,ElectrochimicaActa47(2002)2027. [73] S.Singhal,K.Kendall,HighTemperatureSolidOxideFuelCells:Fundamentals,DesignandApplications,ElsevierAdvancedTechnology,Oxford,UK,2003p.243. [74] S.Adler,ChemicalReviews,104(2004)4791. [75] F.H.vanHueveln,H.J.Bouwmeester,J.Electrochem.Soc.,144(1997)134. [76] M.E.Orazem,J.Electroanalyt.Chem.,572(2004)317. [77] G.Fletcherin:J.Ducham(Ed.),MathematicalMethodsinPhysics,Wm.C.BrownCommunications,Dubuque,IA,1994p.448. 135

PAGE 136

P.Agarwal,M.E.Orazem,L.H.Garcia-Rubio,J.ElectrochemSoc.,139(1992)1917. [79] P.Agarwal,O.D.Crisalle,M.E.Orazem,L.H.Garcia-Rubio,J.ElectrochemSoc.,142(1995)4149. [80] P.Agarwal,M.E.Orazem,L.H.Garcia-Rubio,J.ElectrochemSoc.,139(1992)4159. [81] J.R.Smith,E.D.Wachsman,ElectrochimicaActa51(2006)1585. [82] M.E.Orazem,B.Tribollet,booktobepublished.,2007,Chapter19. [83] J.R.Smith,A.Chen,D.Gostovic,D.Hickey,D.Kundinger,K.L.Duncan,R.T.Deho,K.S.Jones,E.D.Wachsman,SolidStateIonics,tobepublished,(2007). [84] X.J.Chen,K.A.Khor,S.H.Chan,SolidStateIonics,167(2004)379. [85] J.-D.Kimet.al.,SolidStateIonics,143(2001)379. [86] C.Yang,W.Wei,CeramicEngineeringandSciencePro.23(2002)733. [87] C.Brugoni,U.Ducati,M.Scagliotti,SolidStateIonics76(1995)177. [88] H.Taimatsu,K.Wada,H.Kaneko,J.oftheAmericanCeramicSociety75(1992)40. [89] K.Wiik,C.R.Schmidt,SFaaland.S.Shamsili,M.-A.Einarsrud,T.Grande,J.oftheAmericanCeramicSociety82(1999)721. [90] S.K.Lau,S.C.Singhal,1985FuelCellSeminarAbstracts,(1985)p.107. [91] H.Tagawa,N.Sakai,T.Kawada,M.Dokiya,SolidStateIonics40/41(1990)398. [92] J.A.M.vanRoosmalen,E.J.P.Cordfunke,SolidStateIonics52(1992)303. [93] G.Stochinol,E.Syskakis,A.Naoumidis,J.oftheAmericanCeramicSociety78(1995)929. [94] H.Y.Lee,S.M.Oh,SolidStateIonics90(1996)133. [95] Y.C.Hsiao,J.R.Selman,SolidStateIonics98(1997)33. [96] T.Kenjo,M.Nishiya,SolidStateIonics57(1992)295. [97] J.A.Labrincha,J.R.Frade,F.R.M.Marques,J.ofMaterialsScience28(1993)3809. [98] G.Chiodelli,M.Scagliotti,SolidStateIonics73(1994)265. [99] H.Yokokawa,N.Sakai,T.Kawada,M.Dokiya,DenkiKagaku57(1989)821. 136

PAGE 137

H.Taimatsu,H.Kaneko,K.Wada,J.F,in:B.V.R.Chowdari,Q.liu,L.Chen(Eds.),RecentAdvancesinFastIonConductingMaterialsandDevices,WorldScientic,Singapore,RepublicofSingapore,1990,p.417. [101] J.Mizusaki,A.Tagawa,K.Tsuneyoshi,A.Sawata,J.Electrochem.Soc.138(1991)1867. [102] M.Kuznecov,P.Ostchik,K.Eichler,W.Scharath,Ber.Bunsenges.Phys.Chem.102(1998)1410. [103] M.Kuznecov,P.Ostchik,P.Obenaus,K.Eichler,W.Scharath,SolidStateIonics157(2003)371. [104] V.Brichzin,J.Fleig,H.-U.Habermeier,G.Cristiani,J.Maier,SolidStateIonics152-153(2002)499. [105] M.Juhl,S.Primdahl,C.Manon,M.Mogensen,J.ofPowerSources611996173. [106] T.KenjoM.Nishiya,SolidStateIonics57(1992)295. [107] N.L.Robertson,J.N.MichaelsJ.Electrochem.Soc.137(1990)129. [108] D.Herbstritt,A.Weber,E.Ivers-Tiee,ElectrochemicalSocietyProceedings,99-19(1999)972. [109] S.P.Jiang,J.G.Love,Y.Ramprakash,JournalofPowerSources147(2000)3195. [110] J.R.Macdonald,ImpedanceSpectroscopy,JohnWileyandSons,NewYork,NY,1987,p.74. [111] J.R.Smith,A.Chen,K.L.Duncan,M.E.Orazem,E.D.Wachsman,ElectrochemicalSocietyTransactions1(2006)243. [112] A.Mitterdorfer,L.J.Gauckler,SolidStateIonics111(1998)185. [113] E.P.Murray,T.Tsai,S.Barnett,SolidStateIonics110285. [114] F.h.VanHueveln,H.J.M.Bouwmeester,F.P.F.VanBerkel,J.oftheElectrochemicalSociety144(1997)126. [115] Y.-K.Leeet.al.,J.ofPowerSources115(2003)219. [116] S.P.Jiang,J.G.Love,Y.Ramprakash,JournalofPowerSources110(2002)201. [117] J.Mizusaki,K.Amano,S.Yamauchi,K.Fueki,SolidStateIonics22(1987)313. [118] J.Newman,K.E.Thomas-Alyea,ElectrochemicalSystems,JohnWiley&Sons,Inc,Hoboken,NJ,2004,p.213. [119] J.Mizusaki,K.Amano,S.Yamauchi,K.Fueki,SolidStateIonics22(1987)323. 137

PAGE 138

E.Bucher,W.Sitte,G.B.Caraman,V.A.Cherepanov,T.V.Aksenova,M.V.Anayev,SolidStateIonics,177(2006)3109. [121] S.A.Saltykov,StereometricMetallography,2nded.,Metallurgizdat,Moscow,1958,p.446. [122] C.S.Smith,L.Guttman,Trans.AIME19(1953)81. 138

PAGE 139

JeremiahRobinsonSmithwasbornatanearlyageinLongBeach,CaliforniatoPaulandPeggySmith.Attheageofthree,Jeremiah,hissisterDionne,andhisbrotherPaulJr.movedalongwithMr.andMrs.SmithtoSanAntonioTexas.Whenhewaseight,theSmithsmovedtoMaryland.Itwasn0tlongbeforeJeremiahmadenewfriendsinMaryland,butstilllongedforTexas.Mrs.Powell0sfourthgradeclasswasoneofthetoughestacademicyearsofJeremiah0slife.(IfhismotherhadnotdecidedtodosomeofJeremiah0sendlesshomeworkJeremiahwouldneverhavemadeitpastthefourthgrade.)Astheyearswentby,JeremiahbegantoidentifywithMarylandandeventuallynolongerconsideredhimselfaTexanalthoughhisloyaltytotheDallasCowboysneverwaivered.WhiletakingclassesatGlenallanElementarySchool(GoGators!),Jeremiahdiscoveredthathewasagoodstudent.Receiving$2foranAand$1foraBwasalltheextramotivationJeremiahneededtobecomeaconsistentmemberofthehonorroll.AsJeremiahmatured,herealizedthatabigpartofhisacademicsuccesswasthathegrewupwithtwobrightoldersiblingswhoexposedhimtonewideasandwerealwayshappytohelphimwithhishomework.JeremiahwashappytoattendJohnF.Kennedy,thesamehighschoolattendedbyhisbrotherandsister.AsJeremiah0shighschoolyearsbegantocometoanend,Jeremiahwassadtoseehisbrotherandsisterspendlessandlesstimeathomeandeventuallymoveoutofthehouse.Jeremiahdecidedthathetoowouldmoveoutandplacedapremiumonscholarshipoersthatincludedroomandboard.ThiswasoneofthekeyfactorsthatledhimtochooseUniversityofMaryland,BaltimoreCounty(UMBC)forschoolingoverHowardUniversity,wherehisbrotherandsisterattendedcollege.Roomandboardwasn0ttheonlyreasonJeremiahattendedUMBC.OneofJeremiah0sregretsfromhighschoolwasthatveryfewofhishighschoolfriendswereblack,despitethefactthathishighschoolwasextremelydiverse.JeremiahwasrecruitedtoUMBCtojointheMeyerhoScholarshipprogram,aprogramdevotedtodevelopingblackPh. 139

PAGE 140

D.sintheeldsofscienceandengineering.AlthoughJeremiahdidn'tunderstandwhataPh.D.wasatthetime,herelatedtotheotherrecruitedstudentsanddecidedtobecomeaMeyerhoScholar.Tothisday,someofJeremiah0sbestfriendsareotherMeyerhoScholars.AtUMBC,Jeremiahmajoredinphysicspartlybecausehehadalwayshadanintrinsiccuriosityabouttheworldandpartlybecausephysicswasviewedasthetoughestmajor.Asgraduationapproached,thecourseworkbecamemoreandmoreabstractandthereforelessandlessinterestingtoJeremiahandherealizedhedidn0twanttopursuegraduatestudiesinphysics.JeremiahdiscoveredtheeldofmaterialsscienceandengineeringbychanceandelectedtoattendgraduateschoolattheUniversityofFlorida.UponarrivalattheUniversityofFlorida,JeremiahjoinedtheresearchgroupofDr.KevinJones.JeremiahwashappytojoinDr.Jones0groupbecausethefocusoftheresearchwasonsilicontechnology,anareathathaddirectindustrialimportance,unlikemanyoftheeldsingraduatelevelphysics.Afteraboutthreeyearsofstudy,Dr.JoneswaspromotedtochairoftheDepartmentofMaterialsScienceandEngineeringandwasforcedtocutbackonhisresearch.Jeremiahwassaddenedtolearnthattheprojectwithwhichhewasinvolvedwasnotbeingrenewed.JeremiahwasinvitedtojointheresearchgroupofDr.EricWachsman.Hewasgratefulandexcitedtoaccepttheinvititationashewouldbeinvolvedinaprojectfocusingonfuelcellresearch.ThisresearchwasappealingtoJeremiahbecausenotonlydidithaveimmediatetechnologicalsignicance(unliketheabstractphysicswhichboredhim),butalsohadenvironmentalimpact.Jeremiahwasundeterredbythosewhoinformedhimthathisprojectwasextremelydicultandquestionedifitwasevenpossible.Atseveralpoints,Jeremiahbecamediscouragedincludingwhenherealizedhisgoalofgraduatingintwoadditionalyearswasunrealistic.Fortunately,manypeopleencouragedandprayedforJeremiahduringthesetimesandheneverquittrying.JeremiahgraduatedwithhisPh.D.in2007.