<%BANNER%>

Particle Deformation during Stirred Media Milling

Permanent Link: http://ufdc.ufl.edu/UFE0022125/00001

Material Information

Title: Particle Deformation during Stirred Media Milling
Physical Description: 1 online resource (154 p.)
Language: english
Creator: Hamey, Rhye
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: design, flake, grinding, milling, obscurant, particle
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: Production of high aspect ratio metal flakes is an important part of the paint and coating industry. The United States Army also uses high aspect ratio metal flakes of a specific dimension in obscurant clouds to attenuate infrared radiation. The most common method for their production is by milling a metal powder. Ductile metal particles are initially flattened in the process increasing the aspect ratio. As the process continues, coldwelding of metal flakes can take place increasing the particle size and decreasing the aspect ratio. Extended milling times may also result in fracture leading to a further decrease in the particle size and aspect ratio. Both the coldwelding of the particles and the breakage of the particles are ultimately detrimental to the materials performance. This study utilized characterization techniques, such as, light scattering and image analysis to determine the change in particle size as a function of milling time and parameters. This study proved that a fundamental relationship between the milling parameters and particle deformation could be established by using Hertz's theory to calculate the stress acting on the aluminum particles. The study also demonstrated a method by which milling efficiency could be calculated, based on the amount of energy required to cause particle deformation. The study found that the particle deformation process could be an energy efficient process at short milling times with milling efficiency as high as 80%. Finally, statistical design of experiment was used to obtain a model that related particle deformation to milling parameters, such as, rotation rate and milling media size.
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 Rhye Hamey.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: El-Shall, Hassan E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022125:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022125/00001

Material Information

Title: Particle Deformation during Stirred Media Milling
Physical Description: 1 online resource (154 p.)
Language: english
Creator: Hamey, Rhye
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: design, flake, grinding, milling, obscurant, particle
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: Production of high aspect ratio metal flakes is an important part of the paint and coating industry. The United States Army also uses high aspect ratio metal flakes of a specific dimension in obscurant clouds to attenuate infrared radiation. The most common method for their production is by milling a metal powder. Ductile metal particles are initially flattened in the process increasing the aspect ratio. As the process continues, coldwelding of metal flakes can take place increasing the particle size and decreasing the aspect ratio. Extended milling times may also result in fracture leading to a further decrease in the particle size and aspect ratio. Both the coldwelding of the particles and the breakage of the particles are ultimately detrimental to the materials performance. This study utilized characterization techniques, such as, light scattering and image analysis to determine the change in particle size as a function of milling time and parameters. This study proved that a fundamental relationship between the milling parameters and particle deformation could be established by using Hertz's theory to calculate the stress acting on the aluminum particles. The study also demonstrated a method by which milling efficiency could be calculated, based on the amount of energy required to cause particle deformation. The study found that the particle deformation process could be an energy efficient process at short milling times with milling efficiency as high as 80%. Finally, statistical design of experiment was used to obtain a model that related particle deformation to milling parameters, such as, rotation rate and milling media size.
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 Rhye Hamey.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: El-Shall, Hassan E.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022125:00001


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 E20101109_AAAABT INGEST_TIME 2010-11-09T19:34:31Z PACKAGE UFE0022125_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 24590 DFID F20101109_AABHKC ORIGIN DEPOSITOR PATH hamey_r_Page_077.QC.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
2c0e6fdcb46a33608588601cc6e7945a
SHA-1
77c529d9f3639789907a6241bf1b172f61bcf0bb
7026 F20101109_AABHJO hamey_r_Page_051thm.jpg
060d020ec4b4285a56710c9e47a1989b
e83bcee6e0acf341cd42689779b55108a975e5cf
27198 F20101109_AABHIZ hamey_r_Page_026.QC.jpg
c9ff2ce177498c6aa5097cf83037369d
1990dd51f8ab7d5d84abb7999257451fb2860f4f
1051983 F20101109_AABGFX hamey_r_Page_090.jp2
e11b7b054e83ecf322493ea676bdf858
8fce71710a9a5b75ac415155cd3acb94e0003e72
25271604 F20101109_AABGHA hamey_r_Page_034.tif
fdaeeb3d0950fecbc5c001c0ce2d6d71
9286b81180696c39a9876454d1ce1229008c3f2a
F20101109_AABGGM hamey_r_Page_083.tif
745bc68a7429e5d491f4d274790b3da1
bfa510e86733504130129d750bdd1cdf7ba09c53
6962 F20101109_AABHKD hamey_r_Page_078thm.jpg
253505490b923e6b088d65b2c120070c
709b4fc0d7e9d229715373bee87489de8b3bc8dc
7128 F20101109_AABHJP hamey_r_Page_053thm.jpg
16435bbd30885b5e6e3d691c348aa24f
fa4b229a5211d1f92f275ed0c9780147fe38276d
4815 F20101109_AABGFY hamey_r_Page_006.txt
7bf7c5ea0f6f0052ddd7a07caaaa4fcc
5c200ffc6241dbdccf702904eb774c2eca37cdc0
F20101109_AABGHB hamey_r_Page_053.tif
3dfc29596ba973e5992773193da0f774
6a59a209e4e802f835751714805955fc7cae84fd
1051940 F20101109_AABGGN hamey_r_Page_028.jp2
657337f6f2eededc691e62fb4b26cf15
3de26e0322b6f6a8289537af96c1687d46f1e157
5554 F20101109_AABHKE hamey_r_Page_083thm.jpg
af05d51fff77c704ba5ac078e4d210d4
529eb5582f1e06a117903db71e6602eb552e536b
14058 F20101109_AABGFZ hamey_r_Page_104.QC.jpg
61ed5f74073cff442af14d21148f4c14
2482343c2dfa14a49a15f4c621933398778f5fac
12507 F20101109_AABGHC hamey_r_Page_125.QC.jpg
dcf060f61d3a3ad087230128525794a5
1ac152992ecdce908f0fe9ff67c66649857fceca
26501 F20101109_AABHKF hamey_r_Page_084.QC.jpg
654a0331af58fbdcebbfbeaf8e369e0b
f21cda3a0f70f77ce47965a31eec0e3157eebb08
26166 F20101109_AABHJQ hamey_r_Page_055.QC.jpg
1e397bf9d1e4dc6f32b2e50699d958f6
062472fa26371725fc5ca75d31e01b2e4c3dc45a
52013 F20101109_AABGHD hamey_r_Page_082.pro
e79211dee66b2b6a5aae66bb4f1a5c33
f092667d41d25f7fce53cce1510f4e1f5540e59f
2279 F20101109_AABGGO hamey_r_Page_092.txt
a328f24dbfc3bb6cbf642e8918327cdb
f7bbd095daa351d6fc8bb36747a74f3bebba022a
27532 F20101109_AABHKG hamey_r_Page_087.QC.jpg
15cad5fff7fa92dd3807e08e73a3e5b6
cf863feab04715a7396eeada2c35f2494f2a37aa
27920 F20101109_AABHJR hamey_r_Page_057.QC.jpg
56f453f6f79db04c9d69ea66288b7678
eb491ecffc0a56e61c5d3da1ba32986772e4686e
19906 F20101109_AABGHE hamey_r_Page_016.QC.jpg
f0d56c119e0c5005f2c823217348c513
82620e03e2ec4bfcaabfb295ab241be4083649ff
440 F20101109_AABGGP hamey_r_Page_142.txt
737c2b75508f4c2374d1e9b6cd3c5901
1344fa09f8fd6897bf67962c45d0adb1993979c7
6867 F20101109_AABHKH hamey_r_Page_089thm.jpg
b254a93767acd683e1f6b5fa9aa233f2
f6c56bf5d1956e1d4cd31eeadd8eb49a7984f918
7408 F20101109_AABHJS hamey_r_Page_057thm.jpg
d5a2da43657bf6b2fd127c94a3eed492
67ccd06fb1704a0f397f3781384511a9edcfc619
3800 F20101109_AABGHF hamey_r_Page_115thm.jpg
d20c494be03ba50e013b0b3e1b44be2f
b7715562a577bdd15458aba34e920cfda0f3d615
62716 F20101109_AABGGQ hamey_r_Page_064.jpg
8e5bb07fbf5b317d51e3e62834791c37
e03cf4d44a4f7b1ba691b6ac92162e7360e1fe42
26143 F20101109_AABHKI hamey_r_Page_091.QC.jpg
306200eacb0f7878019e858bfb9b31be
17348c363a695bfc4934d360476994c68f6dd0e8
26561 F20101109_AABHJT hamey_r_Page_058.QC.jpg
c7261a7eddf79a25c4162849e21e1687
00574a76535452cf4506406915db5da39a54a1a6
172 F20101109_AABGHG hamey_r_Page_148.txt
b633f3db206ca9ad03f4cc7044dbc711
1125ea62a62286fdc39156689671aeaffe5f4648
54712 F20101109_AABGGR hamey_r_Page_103.pro
937fa6d64587e8f7ac294f8e6fb7a989
0b355c87cb5a1a514fceed9971e779f80233e4c3
7588 F20101109_AABHKJ hamey_r_Page_092thm.jpg
a6ed0483f60cbdc71202a7c8cb85a652
4d423686a9f362dfd2cf5349b9efef9dbda2b499
6008 F20101109_AABHJU hamey_r_Page_064thm.jpg
30d5fc28d25cbb0c753a0444be2ebb61
1659ce7d6e8f694f8c4006381ad30eca9bf186a7
1051946 F20101109_AABGHH hamey_r_Page_134.jp2
87dd087a8949e3f0d9cef63d1161b825
c29028584a8a51b5f440d66f595ba9fdd55206b4
7738 F20101109_AABGGS hamey_r_Page_138.pro
629adfe918644e87cb11bfc9490a18a9
bc2ec7f60c13aceda5d6772647e8e9dcce8f9584
27874 F20101109_AABHKK hamey_r_Page_093.QC.jpg
6428237f2f06d073a1edc036e6b495f1
732bdb928b58853582c7309e129cd4934b539416
4573 F20101109_AABHJV hamey_r_Page_065thm.jpg
effb321c1fac8b8e4efe2188cf08bc42
1bb24bf3af724f3571aed8c7509be085918e86d4
7654 F20101109_AABGHI hamey_r_Page_050thm.jpg
15d3d0f59c8c74fcdd2b2d26d2850ef2
1ce68a703e713c68e5cddabd7b086778b50fd78f
F20101109_AABGGT hamey_r_Page_133.tif
3ea3283222c70153aeb83347828e9795
2220ca22cbcceb99682e873ebe1677156c3ddd71
3823 F20101109_AABHKL hamey_r_Page_094thm.jpg
5c2e48de275f4ec7fdab255edd70c201
0b8dbee8dfdf522b3d83a2493083439f6cc56e04
4339 F20101109_AABHJW hamey_r_Page_067thm.jpg
76d2c8027504cc0c0fd93d5490ec5c3d
e170d8e06b6261b6af0af9026c66c776a7d37a76
84360 F20101109_AABGHJ hamey_r_Page_047.jpg
6841d8459bbb7f33f4ed0f6b90293777
aaae85796c0f983b44baf5e8ef1781329dc4cba2
2171 F20101109_AABGGU hamey_r_Page_079.txt
88ce0704a5dcf58f2f23c5f0b0671be9
b1af35346de1875333c6b01eb9ffc7eb0ce50e7c
7775 F20101109_AABHLA hamey_r_Page_131thm.jpg
142104c6fa15f1205d0c8cf1957a9233
7481c728922ac336e4b9ee5a251e3d709e4321c8
11146 F20101109_AABHKM hamey_r_Page_098.QC.jpg
2d8acf4d0dc0aca59cba92bc969dd6f6
520b252340fbc723fd675c0a1a0ab0f4df8c1732
2784 F20101109_AABHJX hamey_r_Page_070thm.jpg
7d57dd7705b309acc14af64bf583df37
9a9a7dd7452919e3c6a41ea5586e75f70571d0f7
676036 F20101109_AABGHK hamey_r_Page_085.jp2
de639bcd72023fc72cd460228f25f883
e54ca6bfb3a5872e388437553449e55ba41c1fd1
1051969 F20101109_AABGGV hamey_r_Page_046.jp2
49543f8aadcaf6ddb4ec22c032c85910
63313e7b3e8d25cd57904760ced53a73ad46ea47
7181 F20101109_AABHLB hamey_r_Page_132thm.jpg
79eae40a0c792c8adadde2f92b41aad6
d91c4060400d2317805d8cf6a6c30e5f64210eb3
11639 F20101109_AABHKN hamey_r_Page_099.QC.jpg
b849d8c1079162ad3e90f5e8e4efb5cc
f1ff3d79733262cad74c2e658a3d6c2cf277764f
4343 F20101109_AABHJY hamey_r_Page_071thm.jpg
8736a157bb673f8d7d8b2bd8627ba8e5
9bde4a6d828bce16efe9964716ef7a5df19a853d
57936 F20101109_AABGHL hamey_r_Page_136.jpg
deb78b0672a661edb7fe9a723deb7a4c
b351aa4730e8a5733c632b9b0da9a98abdf70692
33563 F20101109_AABGGW hamey_r_Page_139.jpg
3094e8ed727742d1eae887de1d12d418
49ac3267de782dd948d865a58305964ed70763d5
26784 F20101109_AABHLC hamey_r_Page_133.QC.jpg
362aeed07ecacae5ad21ead9d3a2748d
0843425f3c6fd0a6c5df5e6e82f6cd4ca79cc0c3
3926 F20101109_AABHKO hamey_r_Page_100thm.jpg
d1a4d919267cfe9a5ea17e2446008f08
01ca9e6335a7cb6603a52bb851c4eb183307c4cc
26397 F20101109_AABHJZ hamey_r_Page_072.QC.jpg
9d65e792bbb3d22f16509f68e78dbe05
629bb483483d638c4cf199530f1149e157dcd68e
5103 F20101109_AABGIA hamey_r_Page_061thm.jpg
56d9672f6216ab2d70ade32e236d1450
aa01c9bac06cf31fbf78bc96c7b60eedbd702c8d
77920 F20101109_AABGHM hamey_r_Page_035.jpg
699b864fc7023a024d63e9a50d93cbd8
9996c3cd4cd060c2cb3976078fa829ca15ee532d
14716 F20101109_AABGGX hamey_r_Page_065.QC.jpg
8e181cec7e033eb86003cb3d0a28c05d
757150875829d87f501c33d6c55144aa1c45602d
7299 F20101109_AABHLD hamey_r_Page_133thm.jpg
31d3ac8d4247a373b89d6b80f17717e8
0b3f548bb1e28b4150896e568cf1486ca30f4eca
7917 F20101109_AABHKP hamey_r_Page_102.QC.jpg
66f12c1c70a2e5fff9e8c9f187e4131d
3b4abcecfd0290916ef01ecb95ba238b34f58ff8
F20101109_AABGIB hamey_r_Page_146.tif
2066ea81363d440d5d201d67b442c4da
836f3bf9079380982539c59265cce6a867ab07e3
4364 F20101109_AABGHN hamey_r_Page_116.pro
4e786f00588c47bf56625197e0a838d8
8db6d429657c649f8672e16751b526bfc30802ad
175680 F20101109_AABGGY hamey_r_Page_126.jp2
82055d09f669e0e8383a680c9c999687
46389543b79ac7e8c45a66379f66c6c2588fc36c
16471 F20101109_AABHLE hamey_r_Page_136.QC.jpg
4e90a431ef3fb88ea4bec9a3e79308b1
ab3fff6a873bd1744e4ab284bd7437da32e46cd4
20610 F20101109_AABHKQ hamey_r_Page_103.QC.jpg
d7c24fc7504f3e9adb67d56cde53ee8b
bd3659213b48419b4e68974be29cb0dc29ffb856
6898 F20101109_AABGIC hamey_r_Page_025thm.jpg
b3e54802a9dcacd70ebebbbb657541fb
4467079bb30efcb1591f4f2979b545f5b52793de
4072 F20101109_AABGHO hamey_r_Page_138thm.jpg
ca75f8444a6b70af59eaa64255e91564
2bc30c8b3a9f69e4aad2dbdb7ce739fa6ec8777f
1051956 F20101109_AABGGZ hamey_r_Page_145.jp2
e80f7cb4c4df1bbb39ff0ab2b86f82e4
72f77c789b1759901d034be604f7852c48792eaa
11418 F20101109_AABHLF hamey_r_Page_139.QC.jpg
f5071eb3abf0eb245d71629cfe3835f1
0e90a026cb1c14607c32bd5d7d744be74ff562d8
F20101109_AABGID hamey_r_Page_006.tif
655ebfe508b8373001fc2c9f3b45a413
e848d50ceda90936441f0a7d3f073fc0cec72829
11284 F20101109_AABHLG hamey_r_Page_140.QC.jpg
5b9d539113dc1083de519cb061133ffe
61f14d414216ea75adf89abbdb4a33d8f5ad6c28
4292 F20101109_AABHKR hamey_r_Page_104thm.jpg
97d74f9d54049623beb89415b052cabe
9e5d6a4df034ef8dcac05f06c099a729558e46db
305961 F20101109_AABGIE hamey_r_Page_112.jp2
7eef00df56ddd2495188a642227db530
32897c4a0ba578d3bf95b22e2b3c5b4628267e04
1051897 F20101109_AABGHP hamey_r_Page_144.jp2
c386c809f253066a446433aafdd6b334
62c14c9494f4734e4742c1c784ddb2d2c54604d7
5720 F20101109_AABHLH hamey_r_Page_141thm.jpg
f6758d144ae01342b8e8d18266c2c4d7
607b813904e6ce301091eae41ae99f1dede2842d
3866 F20101109_AABHKS hamey_r_Page_105thm.jpg
e9f491743e239d694e6bbb8121f936d3
2e44b59896a25a537fae20f2d1805717ddb9f2c6
45471 F20101109_AABGIF hamey_r_Page_069.jpg
57c39cbd05bfc795d7b3af9cecf1144e
4f9e430ce4185d13bf4e08c4155e95971131d457
F20101109_AABGHQ hamey_r_Page_030.tif
a16753e1396da7ceaf417855765df4f2
a8568346b0c60c419d99e7f9c0db78b4c045728d
3181 F20101109_AABHLI hamey_r_Page_142thm.jpg
6b1a7cd9be71a52a0ba00d47431116ca
4316a453fafd297710cc055881e8bb7826130914
3588 F20101109_AABHKT hamey_r_Page_106thm.jpg
b596a531eafe4defa7abe9bdf8678040
041873c709cfbff6d5f17d29fe431543592f5f48
6895 F20101109_AABGIG hamey_r_Page_136.pro
ee82ceb31ae3e66abcdb56564f1fed2a
7b5d4cd45ccfbdf3962d8d75a88e61a8b4550533
F20101109_AABGHR hamey_r_Page_049.tif
6ad020bbc56e5cf1cd93a1648a772202
28af4577c53d7c0ceabf25ac53c6519904e6da3c
25957 F20101109_AABHLJ hamey_r_Page_144.QC.jpg
0ba328c09b61dcfec40d15cfd0f0063f
da184c6b28533f1ad632dd56bc6a88c1de916fb8
9764 F20101109_AABHKU hamey_r_Page_110.QC.jpg
628ce8f29d4f779fd4a81dbb985cdb95
13d8390ee37636af4b785d5665f23d744843d0b8
2157 F20101109_AABGIH hamey_r_Page_031.txt
1417f028b27854124c858debf38c4e33
768e9028b3b2bdc3c69cf092bf0aa44fb6fb05c1
358861 F20101109_AABGHS hamey_r_Page_070.jp2
0971d816aadf702ec5d3f25238984079
b2c6a19b79448a1135b273d73b853373162727bc
7172 F20101109_AABHLK hamey_r_Page_147thm.jpg
efaaff7f318a13653722f89baab3ab8c
1fa5d94177a045c001a148b25579b3c4bf92f1a9
8211 F20101109_AABHKV hamey_r_Page_111.QC.jpg
8f29d21c761ac1e96d166073195fbd6f
f1a9ed78fef080ff543b0fb2eeb4e2167f7ccf55
81004 F20101109_AABGII hamey_r_Page_149.jpg
090da01fe450505930c170cb8fdcf5a1
a45eb3963f49b366fcd1d15e60ad531b9f263f50
10247 F20101109_AABGHT hamey_r_Page_108.QC.jpg
d5187afbe87881ada50979dd1cdcd904
8a4d08a99bba007b3b7630c0ae3688a3c4651720
4821 F20101109_AABHLL hamey_r_Page_148.QC.jpg
ef84c6d37780053b0da36efed63ce57d
ba86946ef2aaf29deb547544e68483597468cf71
3841 F20101109_AABHKW hamey_r_Page_118thm.jpg
5959cd1a0049754b67c550417c850fe2
3e7c8c48c38b2b1850c024d971353780f7e83813
82251 F20101109_AABGIJ hamey_r_Page_091.jpg
d7c270cb58ed91e2198a30112e44364a
fc5b5919cae57517ac4d33d2ac894ffc059aabac
6706 F20101109_AABGHU hamey_r_Page_063thm.jpg
1c9ce279cf9bdd0799bd8f63b1db070b
df0c58d79841c98297e493d50a84f93b36573258
24557 F20101109_AABHLM hamey_r_Page_149.QC.jpg
d3a63ca0e7fd2a92ebb23db59e8a87dc
aa4f76e6c36fa275ecdad3d55fb43f1dac472241
10600 F20101109_AABHKX hamey_r_Page_120.QC.jpg
13f40d1245f0a7bf3090c1326e6ff356
8819ea72dc885877ca4d44101c22feff40bae3bd
F20101109_AABGIK hamey_r_Page_137.tif
cd1eb0b7099f29b5d6d01fe935575fc9
c7c1e6acf639f14cdeba6377f88db6fb0e81f3e1
1051972 F20101109_AABGHV hamey_r_Page_093.jp2
c0934b1f50abf2676ef86a145f5416d0
6108d319b5d82cca87843d17e80c1bf57cb1a5e8
6913 F20101109_AABHLN hamey_r_Page_149thm.jpg
6c39f56381e927943a77caf5ac3ccf5a
19b76f07293fc0ed1b98005a87c93d5f3bb34c98
3830 F20101109_AABHKY hamey_r_Page_123thm.jpg
dcc715e8034af6443a38b1c7da1f2692
acdad6a681abfd59655266f370f55710d07a0c95
1344 F20101109_AABGIL hamey_r_Page_104.txt
24baa62916c4135083081d06a9d42921
86ce3162fa213cd797222c6ecb280aaf7402532f
161 F20101109_AABGHW hamey_r_Page_064.txt
37082ac305cdf567548986b95360650d
bc04413b65e074dba3e8d354888a3885b11036e8
24952 F20101109_AABHLO hamey_r_Page_152.QC.jpg
737f8a0cb460d821bdae282de86735ee
c5be237abac7859433d5b7a0b96852329b0b0d36
27908 F20101109_AABHKZ hamey_r_Page_131.QC.jpg
40507be446358a8e1e3689bc04a7cfcb
dd2e7a181a922e78a49203dcab6d0310f67f18ee
F20101109_AABGIM hamey_r_Page_005.tif
c5f903a479e1def12b4dd28594834500
fb4c96e7d1bddbf26ad76f60f9cbc65a15f9d3d0
50923 F20101109_AABGHX hamey_r_Page_077.pro
f82833fc343571aa7ffdd85a809d3fdb
255f852655b80e8f61a1f467fa74bdc7008051fb
8423998 F20101109_AABGJA hamey_r_Page_065.tif
e9d90f135c564df0df6cdf030263451f
3cfd380ad87835e8dc66370c765040f3a60854af
8903 F20101109_AABHLP hamey_r_Page_153.QC.jpg
1ab850a70a8fdad4dd8e2da908d59d30
67de5c3877075f223bb6749c95533ee686989973
2053 F20101109_AABGIN hamey_r_Page_128.txt
32fce7844f3afb2908524d10af96eaee
93602f274405049f81207328d521d044400e419f
2163 F20101109_AABGHY hamey_r_Page_131.txt
78b49dfabf36330d701db4f68f3e0481
bacc630cd8b052dd2b51a0a6294c28bcdf92ca0e
F20101109_AABGJB hamey_r_Page_150.tif
378eb78dc27a28257f4fe596d6d5bb4d
077009d1dcdeba40de939324d565cd2b461bb2e2
3013 F20101109_AABHLQ hamey_r_Page_153thm.jpg
18d218c3965f5e2ee69b7645e9ce6aad
ba3c822536f938a3aefecdb833e977ca62665064
F20101109_AABGIO hamey_r_Page_079.tif
f71b8fee7cec11909f3e813603171937
49baa646f486c8157ae473b3f488bc480bffc719
3960 F20101109_AABGHZ hamey_r_Page_125thm.jpg
6439c98b4ed2aac8564ac3603883aa4f
c4854346eaf0e2357d5a3f2ef6d18eda3b914958
56021 F20101109_AABGJC hamey_r_Page_026.pro
27de571a286a8a53a82624d7595870d3
c84c22a085acb2801a89a5ab13dd8b3fe9f082e7
764571 F20101109_AABGIP hamey_r_Page_083.jp2
9448bdffe4cd1b52899935369bb441e1
a2673ee69df5fcf906eb78101825b50757b0f468
1783 F20101109_AABGJD hamey_r_Page_114.txt
c090a68ef5cf98e3dc14d5ed3fa0d4b2
aac6987a4004fba989998d94b2473357370ab1a7
F20101109_AABGJE hamey_r_Page_145.tif
ddcd645012c1a8df4434dbfbfd9a3e35
905624515520130d6b6257c561df0425f20b6f69
14525 F20101109_AABGJF hamey_r_Page_098.pro
cd920e80240f8056e0a8147be7979bff
8600522b69e5518e1e002271fcb8052a16052206
38093 F20101109_AABGIQ hamey_r_Page_037.jpg
472bdc2d4b2421ceccbcdc68c1f9d30f
084748a177c27dd6a3416f7bdfceef5586b8f9a7
18160 F20101109_AABGJG hamey_r_Page_004.QC.jpg
86d68f414f21b4bfb1a391597cffd9a1
803085d12c078a882b8adfa11a4e861e75e5a0cd
81039 F20101109_AABGIR hamey_r_Page_005.jpg
09b8a05c7eaeb6474208a4f11c2bfe24
0cfb63347eda57262a88c2bff1b06f8e4d8983ee
1348 F20101109_AABGJH hamey_r_Page_002thm.jpg
0144917ae6c810fa189cbe25c64ac6e9
2fd6722fb98764b92bbc9cf3cc12db17f2458be0
23663 F20101109_AABGIS hamey_r_Page_006.QC.jpg
e785497096aa4fc60399b06b60cdeadc
ba69953c15e08926b973c9cd99fc98e3585ede88
4178 F20101109_AABGJI hamey_r_Page_037thm.jpg
57130920c71d5e140398723a132bb016
1a77fc4affe11fa8db53b0d54ffd4f8b960461c9
4252 F20101109_AABGIT hamey_r_Page_148.pro
7f211e157ae39d8616b2fc9b47b5f5da
2377ca78ea08f52906255cb4e52753e00516c9df
7376 F20101109_AABGJJ hamey_r_Page_049thm.jpg
b7b61ced1e730e0933c68dd0f2592af2
4280430722a9b7471d026fa7fa61b40a4d992ddc
22394 F20101109_AABGIU hamey_r_Page_101.QC.jpg
9a9dcf10c05828e4cb37a809234cdebe
b4f3bb84d7a386b40e3244f17fa76fd95929736d
3779 F20101109_AABGJK hamey_r_Page_154thm.jpg
aebf7c4a1dd2fdc04350d77ceeb51d79
a8e4b544c78a98dd76adb12c6bdb6e494ab4aa1b
46051 F20101109_AABGIV hamey_r_Page_033.pro
b0f2dc3757ec00a493ee4dcc6560fb51
75c80128f5b608113066f3ad0a9086db6c8f582e
1979 F20101109_AABGJL hamey_r_Page_046.txt
c1540ac906fcbcd032d8c3859632a7c2
af8acd7682e1dde965bf70fb3c043a649c278818
26622 F20101109_AABGIW hamey_r_Page_043.QC.jpg
0535e35d911e0aa5c9b6630beb0d35c5
01bb4d2a3b3d3d052ef41144c72b05547b4a25d3
84949 F20101109_AABGKA hamey_r_Page_049.jpg
e66e22dc4ef2df3817be597dca496be8
a949c41b3c3335e8ca55da1d8c566fa00401f8f5
599662 F20101109_AABGJM hamey_r_Page_037.jp2
11e1e3435fd861cf3d654455a946974e
d6a6d592d0ad6c869cd967a1ab8316b6effb1ba8
2183 F20101109_AABGIX hamey_r_Page_151.txt
ecf9697eed64e295d5ba612f6f189aff
96ac56f31917c324481b69c24b14ce4e556ea396
2249 F20101109_AABGKB hamey_r_Page_074.txt
08780e53358a25de001cc34c15074266
c1e38dc61d44c172a1609afb976c1eb8302b8766
16979 F20101109_AABGJN hamey_r_Page_022.QC.jpg
703906502328149fcb1f08aab999f1b7
bc674db03f3b887f2d5ecf29647058e9228bd38d
6486 F20101109_AABGIY hamey_r_Page_007.QC.jpg
7d903c4f7fa8a4c89e5e67ebf5c65a08
4582ef256885f146a776523273bc1ab74f6e36b1
52709 F20101109_AABGKC hamey_r_Page_147.pro
e08e75a10bad160455cdb25231f3d941
a7470034fba3a6cda7a34c664c61fbcc57d51c0a
2128 F20101109_AABGJO hamey_r_Page_029.txt
072d7258c53e5a093859048a113c8726
52350562791d9778da8f6c5906a2b27e9b351564
10155 F20101109_AABGIZ hamey_r_Page_062.QC.jpg
79bf59de2dab139a83ca6dcc683f69e1
8a94af7da13bb6297636b33753ff5fbdc7b38cf4
15625 F20101109_AABGKD hamey_r_Page_007.pro
de403b7d9666798fc55e4bac51522b0a
5deb6eadf75a0e896599281f6a7c2e3ab42f2a1e
7655 F20101109_AABGJP hamey_r_Page_054thm.jpg
9a956988df03ba9f52595d4c21aabf33
eb9e72e1ebfe1695a46e216e60f0eb6fa133dbd3
3297 F20101109_AABGKE hamey_r_Page_121thm.jpg
1aada43f9d144fd7fdebd152bbaa9f4f
d645db7717c087963b87e773920c3d3dafd4c791
3133 F20101109_AABGJQ hamey_r_Page_039thm.jpg
75e60b7ec87bdbe99aa17d17f1e2b974
40a27f336eea19a167cde7a726b6510353e56717
7566 F20101109_AABGKF hamey_r_Page_146thm.jpg
877012149b67b6f0bef24547c57adc4e
ec6c19f216fc510d7c32c49d92da8a3c0e05ac1f
13544 F20101109_AABGKG hamey_r_Page_068.QC.jpg
1bef7581c6cdfbe41478f2eccbb2623d
5563ef6546d52107a03b6ad8e6a30f788b292403
224 F20101109_AABGJR hamey_r_Page_038.txt
a8896136af0763c75b7f7fa01375e16d
120dd7e694e58a5556e867743c45ff07b46648f3
41550 F20101109_AABGKH hamey_r_Page_154.jpg
e37b635885e4c048c42f6fd85449a29b
d22215df976b46d50366ea60b92aff9f7231d79f
F20101109_AABGJS hamey_r_Page_090.tif
abd9f08b2f4d926167fb3ffdb6dc027f
da6c97d46efd2495d0b53cae1ac2aa33af05a9c3
34522 F20101109_AABGKI hamey_r_Page_099.jp2
190cefe769a77914f01bf5419c9723d9
5ab271c6624ba90164175da3f5c1bfe24184394a
24957 F20101109_AABGJT hamey_r_Page_143.QC.jpg
e255f775b56fb05ab8c06e764e3f71c3
ca9d55e0fc004daef2c557d82b9cd91e47c6f8b7
427183 F20101109_AABGKJ hamey_r_Page_096.jp2
a099255b819fba37cd6600e7be277de1
3e53c3a6867799acb7aa447fa845f334d5e169c1
57117 F20101109_AABGJU hamey_r_Page_145.pro
14d2e6b23ebc3e81c6dec63ed69928d4
1da7eeaa36fdca81e0ad844d69e64a2deb9acc2d
338757 F20101109_AABGKK hamey_r_Page_153.jp2
152b9292d373476fe7f39abdafb10af3
c4c3de9c94ca3b9af792eb3ecfce98a042883713
1051963 F20101109_AABGJV hamey_r_Page_052.jp2
6659f04b37a1b0c90d580d1a285b9183
7a93f73abff8c081f46c85eac46d8f746235f10b
7068 F20101109_AABGKL hamey_r_Page_075thm.jpg
4a59b5783361f18c6ac8fc26e2790910
0914c0c9784079533816263dc8edea85e427dfcf
3134 F20101109_AABGJW hamey_r_Page_066thm.jpg
200c88e28224f62cc2fbb48e81f10b5a
5f0fc3563bfee594951ed4d8791cf2839a296bd2
2968 F20101109_AABGKM hamey_r_Page_108thm.jpg
e1f7d4b5c9c5308262e5b90284efa5cc
7ac2d57d62acfdd2f423df3628581016d91adb57
31064 F20101109_AABGJX hamey_r_Page_038.jpg
39bd4ae6098958fb343e0780d7aa5b54
59b25a7b128c47bd572e7fdf57583c08d936f915
5859451 F20101109_AABGLA hamey_r.pdf
63d4d7095411181b8299f8df70eb1260
88876bd9259cdd6a66d4ab03efb10eae6c820d13
30682 F20101109_AABGKN hamey_r_Page_107.pro
d61b5378304d9f001fe4a1e8e52f887b
84dee8ddae9d77036884d318f6a6e07bd4f8f4a2
44220 F20101109_AABGJY hamey_r_Page_107.jpg
f818a7b347d2a72b1a6d7fac12b7daff
a9a69b009d27f53a1ae015c4710063b84269ab97
1051952 F20101109_AABGLB hamey_r_Page_074.jp2
0d09a1ead9054fa106b4f22c7942bc14
182cfefc8b52849253482873b3fe5d97ccfc6be0
1934 F20101109_AABGKO hamey_r_Page_028.txt
11b71e4849a9830a92aea30672ed76e4
b778f33dd35df4be89409a6cd04cfe5e626fe9f3
7372 F20101109_AABGJZ hamey_r_Page_080thm.jpg
51d003164a2c54bcd07b73798bf50429
e4c81081f6af4f78e13f0cd2ff0a520fa9992cf0
1053954 F20101109_AABGLC hamey_r_Page_099.tif
3d0c2af0f1bc94f4b0f02fa96d0a8d78
a0e4c47073086d51100850ea6cd35e5789a772a9
F20101109_AABGKP hamey_r_Page_019.tif
b1529dfe38e44a339be1b79d69829fd7
903ee11a65b16bac03df4610b9b6d742825c06cd
13039 F20101109_AABGLD hamey_r_Page_109.pro
0a1ddc17059d482ed2367c1f31505c41
83077ad6c9f11ef368314260aa1b538fdfd51ba2
52859 F20101109_AABGKQ hamey_r_Page_018.pro
1313ef60510fe3eddc17fedb41be670a
ff40f506dad50018fa27676726df2f28e4dcb386
F20101109_AABGLE hamey_r_Page_035.tif
707fb4a84b17599019117d5858ae1bd4
c56e214efda70638aaa88108c8515dc14c8cabfe
5824 F20101109_AABGKR hamey_r_Page_135thm.jpg
0769834d127c1330a34c9ce811a67eec
24654b4f50ed50d3ea6a5eda8c0f5ecfe89f46ac
7297 F20101109_AABGLF hamey_r_Page_072thm.jpg
14a0f96980f49110c76ef2483df9fc9b
acaec2eda653a95d94d8f50bccf5b1557af98adf
1051960 F20101109_AABGLG hamey_r_Page_012.jp2
4286fa1a318051b214de4843376f0b3c
57ab0582f7b795f138edb9e2ab5d05dd58575caf
36681 F20101109_AABGKS hamey_r_Page_108.jp2
16e22c4eb136ccec754dc54bc3e481c0
890c11b083e65b213077155799463ab2be4d3735
1932 F20101109_AABGLH hamey_r_Page_075.txt
dd287a6ce7cce2c6e52573f6571e8a1c
cecd4379e834694e046bcc6e4fd7f68b8049cb3c
3316 F20101109_AABGKT hamey_r_Page_060.pro
b87225a1ca62024fa19f566182834d71
26367062ad297285e684423d7156b494369747b6
27815 F20101109_AABGLI hamey_r_Page_080.QC.jpg
9adf4827b157365bed338b20063ec934
acff2f5dea5212f015079112f144e131b433d8cb
48021 F20101109_AABGKU hamey_r_Page_023.pro
ed8b5cdabea301612c77a2794eef397f
48426ac58ce3ef39eab8336346056c91029f332d
7075 F20101109_AABGLJ hamey_r_Page_052thm.jpg
3da9abd56e9217eae3257cdcf0f7a579
df16ce50965190afc062bbaf32e89576adb53216
89887 F20101109_AABGKV hamey_r_Page_044.jpg
3c15a5703f39e26f7265205a89ae573f
78a68fb2b54eb01105778e02f253c23ab7415f94
1051958 F20101109_AABGLK hamey_r_Page_143.jp2
b7b414d8811be7afaf077adf2234559d
08f3d1a6eb15ba5841f85e2d9b1850ae61237f40
7320 F20101109_AABGKW hamey_r_Page_126.QC.jpg
35fe6b6a7ae9cf4e622ed6c95e19f94f
1486d8815dd314c823b5bfc9e8c0ad35d2d20224
12751 F20101109_AABGLL hamey_r_Page_114.QC.jpg
c1b8180e7d49e933392df579f13f8152
e609c85ef6ce70c4ec124f2eede78f10f4e2a452
F20101109_AABGKX hamey_r_Page_121.tif
9115650b65cc51df837b37ba019aaa15
1efadcb14dd9be51651f0d08f015907fb0a113cb
418 F20101109_AABGMA hamey_r_Page_001.txt
d2fa25b7a979e7af0838635c87167bb7
8036b8de1238a3692774d3c3a7b6e648fc1059a4
53151 F20101109_AABGLM hamey_r_Page_151.pro
cc9a36a0bcd288f90cc6036b88c36f90
9322de72aada087182e36c48946ef9614497e2ab
85038 F20101109_AABGKY hamey_r_Page_081.jpg
c5a811d0bbe31205e1c7ce28ef5f18fc
2a219e983b67fb3c422d7b8759b11c06d6477d1e
F20101109_AABGMB hamey_r_Page_130.tif
93c3a97549125a73627667b8de007848
8e9b94e47a0657abefa56a6726bed0a4bb96e40a
F20101109_AABGLN hamey_r_Page_015.tif
fd958a8f6d5b89f5d4eeb7cae79b3057
077f752af14be9af54f2aa56169887a56ab6c297
26062 F20101109_AABGKZ hamey_r_Page_067.pro
2761e5da74d21ed5d3601948247853b2
468a2e6f5abbeb8b73aa364734319e90a4213091
2811 F20101109_AABGMC hamey_r_Page_110thm.jpg
885c1cb59b116cf0668fa550f52008a5
f33875d076d9b183a2e0c4a4d0ba422829ae1c6b
F20101109_AABGLO hamey_r_Page_120.tif
3ae89cf683f416d59791688640d3438f
71b568355e2c97edbdfd93e8a959e5c4fe6d19ac
6944 F20101109_AABGMD hamey_r_Page_082thm.jpg
b828627d537e4acd988a4f00af48bb70
7efab9b228972b10d5311dab10452ba532e9d7ed
2046 F20101109_AABGLP hamey_r_Page_078.txt
c527899b4dac8b9d35874a118998651c
ee52eb4da1d250767f75ff04f7100b6d364f3343
1051986 F20101109_AABGME hamey_r_Page_024.jp2
58b80ff8cecf6611e7243bcf9b94c70f
41743b3f91d18f80b9996dfd29c5465f072f0718
F20101109_AABGLQ hamey_r_Page_022.tif
2a5d7ad721725f761c8d3bd8b27a05dd
70adf6a91675ae9a44e6af850b2e2057d06cb9be
F20101109_AABGMF hamey_r_Page_140.tif
d1c9ebeeb50e21889d24fc8fbefc893a
82a68ba7cd7fb1148223411ee112e46b331b5dd8
41457 F20101109_AABGLR hamey_r_Page_140.jpg
e751c13c7e7cd8bfee8ef72eba129b4e
71b66b911ec7e51a766d2bb5dead78da20e82457
F20101109_AABGMG hamey_r_Page_109.tif
3d04b304852b802b30745af024a2d125
36a509a340977ce54185741c1e85c19f9df86abd
33652 F20101109_AABGLS hamey_r_Page_100.jp2
b15f91f2629af389b539fc99e2cff225
42db77930866ed1746b0c24fa91c6f0aae9c1a0f
7661 F20101109_AABGMH hamey_r_Page_044thm.jpg
6ee8ebd973bcf1f6dfc7190bc2ccc8e9
33b0a3ef11b31f819c81beb16f1560d8d293d6b9
23619 F20101109_AABGMI hamey_r_Page_033.QC.jpg
72649adadb91c7eb77c9270087390959
6cb7c1ee2edfad8366032a28662b58b31c0ebc7b
7152 F20101109_AABGLT hamey_r_Page_070.pro
7ee005b2124d1f778b4d9340feb9f0eb
292fb6223611de3f7f258bb8b01f6bffdea26461
1051984 F20101109_AABGMJ hamey_r_Page_086.jp2
0137871d1efd78b4d11072b9cdebcbbf
e393c4bb23a94ee9e85e270afa9fb23fc9e7f615
25289 F20101109_AABGLU hamey_r_Page_078.QC.jpg
d9ec1ab756fabe94e5429d0983a00f12
3973c6beb2ead1335f91af2a20f8b5dd10068108
9622 F20101109_AABGMK hamey_r_Page_110.pro
e67ce4b4f79a3a2fe2b5bc829b51e867
d1692bfced09427e74750dd6387c569fc7025fcf
28076 F20101109_AABGLV hamey_r_Page_054.QC.jpg
236c20da2d7c386f03d98ef975e084eb
a7601d57101e1a4723fc088167f565a189ea7111
5848 F20101109_AABGML hamey_r_Page_120.pro
00ed80e5d2549674c37eec51f0208519
bc8386ccba62851ae5c4427ac0e3df075902331d
81718 F20101109_AABGLW hamey_r_Page_021.jpg
48bffd0e8cc6f45b0af244cd5f6670e0
fdc54394cd0c21dcb6f89817b2051714723a6f35
462464 F20101109_AABGNA hamey_r_Page_095.jp2
fa663668a8c5381ec19584f0628a870f
5be701ea8243dec223d52c925a42854940251aed
6643 F20101109_AABGMM hamey_r_Page_010thm.jpg
1ef46bb90d694db6a0a90b37315a7508
c1f248b60cfc13e9777c964f544c6656fc805420
F20101109_AABGLX hamey_r_Page_143.tif
8a0342c6e8f00cc95696c0051f277a25
305135b2eb20c6bc94e922764dda29a9b9d4fe27
58083 F20101109_AABGNB hamey_r_Page_004.jpg
e1339077cd46b5eca5179bb641e78d1a
0a500b2573ecd82e228d41f1061d2936457fc329
77503 F20101109_AABGMN hamey_r_Page_023.jpg
e271d1fabc9bcd25734722f99c9bdb2d
e7bcf2fe13e4dbd22e551ef10982a72edc2b8db5
3561 F20101109_AABGLY hamey_r_Page_139thm.jpg
6588b6abe876dc0b6068f615fe7cfecb
814ebf12242609200bc186f4e7f4436e1b871a10
1051961 F20101109_AABGNC hamey_r_Page_008.jp2
171a4398c3bcb8495eedee5c840c517e
d91c3a1115b8d3897fc927990e94afe550d71752
1051947 F20101109_AABGMO hamey_r_Page_025.jp2
1a4b0a93397dbbdb03d2624659157e2d
5f03a97a4af9ad7dc4922da91bed2e460170fb44
F20101109_AABGLZ hamey_r_Page_084.tif
c09a118f07e9950066c97dcf506298ed
94c9750ec044a70d4e1c5723a116827a8052a621
80472 F20101109_AABGND hamey_r_Page_015.jp2
852d1b7dbee28b0b36a491782cd73150
3f638e237ecd87c033aa146204cbdbb2fbb02a8d
30574 F20101109_AABGMP hamey_r_Page_059.jpg
545368772f3f504dcbf467a1d1e8c5bd
de264cb86bfe9a50b7cba4951fdb121cd79abf69
12695 F20101109_AABGNE hamey_r_Page_094.QC.jpg
1309a38cd5ec2c7ded015f226560bbd4
b7bac29b152fddcad2ee68f7c1f13c65769725eb
78432 F20101109_AABGMQ hamey_r_Page_025.jpg
428a0e86c1c6dc43beb70bfa03f5c144
96c63f4fbb4b93b9f0bfaddcf4081b4092afc1ba
F20101109_AABGNF hamey_r_Page_103.tif
ffe1bc6f97e581c55231bf5d63ec1528
028ee9f03551def2d32c52131dfeacfd672bb9d6
2186 F20101109_AABGMR hamey_r_Page_072.txt
ea0d80c38cca3e3d57b826fa8501563f
5d59dc5971d605889580b5bfaf6c38ca39258d96
F20101109_AABGNG hamey_r_Page_009.tif
28747ef4c5cb6eb76cd999b994636582
c5c388d3995208821c0cd33a0de126b3994e05ed
6966 F20101109_AABGMS hamey_r_Page_143thm.jpg
5ca3f559e73e0c9027de27a04b1937a8
13b12c06537122c8efe81b4f9dbe4a7c8062a6d2
296110 F20101109_AABGNH hamey_r_Page_062.jp2
0ca4e74f52f7b60549a07ecb8ad329d5
28c7e410b628569148a346005fd5a62daf8f911a
7368 F20101109_AABGMT hamey_r_Page_124.QC.jpg
8a38dd5ba6001da75d074d704b8e2181
19698b1267e5f73d919ad417f03aca4e76900b91
420720 F20101109_AABGNI hamey_r_Page_114.jp2
6f3ba8e88cd3e16b83036784d2971f8d
cd4c8d34c5bd3aabac1715c428a635141f7e21a2
235242 F20101109_AABGNJ hamey_r_Page_059.jp2
0f77f167fd8a586a848ac24212b07144
0b5e26f496a8176c5d020c0d8036bb85cdf9da6a
6237 F20101109_AABGMU hamey_r_Page_033thm.jpg
679ac36ef7632487ce8533da85be7814
4fad1482542605a592da0af34d374240bdcdedd6
1051974 F20101109_AABGNK hamey_r_Page_005.jp2
20f243296ee40ad1c47c10d703364a3f
d7df303bb7dd842ac2b7ab58b766df4aeddb16c8
8732 F20101109_AABGMV hamey_r_Page_100.pro
334f4fb2230546bbd50034bd5be753fe
8c0eea325b7a27ae037beae8975072a404b24fa3
10128 F20101109_AABGNL hamey_r_Page_115.QC.jpg
787a777e6ae6f5e61d5d7f9798b19e56
9de42f2ba2745a52babf864f7fa8604be7cdbfe4
55572 F20101109_AABGMW hamey_r_Page_087.pro
4c2ccb8b244abda9bc7e5b4bb3228b4d
50e9d98fd9eb96b9c46fd2a2a4b502162feb631b
F20101109_AABGNM hamey_r_Page_102.tif
91b2c7c963e639c0d319b9d555e5da3e
e4e361b527a111454d772cd73ae7316c9e1f0e17
7233 F20101109_AABGMX hamey_r_Page_140.pro
81edeb6f6be6956719d851c1b9b43514
8f8dba2f5a620b1b120d13f5585878fe492def29
913766 F20101109_AABGOA hamey_r_Page_103.jp2
92cfc19b72167a61d8c6e5164185be54
95bb2b176d30daee7ecfd7ba85d07045f518738f
F20101109_AABGMY hamey_r_Page_056.jp2
78d5cef7351717fb57faa5894e04ecee
52c350cd3c789e2c10451baefe05b5058db663dd
52734 F20101109_AABGOB hamey_r_Page_132.pro
89c4e01448366a68bcac8fa0472eb925
44432eab051d7a4a65d4625770d00e428a2eb1b1
44878 F20101109_AABGNN hamey_r_Page_122.jpg
7c5f6de1c37661ae9d7da3033a519a8f
b11a1c62e73b1ede0f8b2f7d2a1fb2aba0bd90bf
1477 F20101109_AABGMZ hamey_r_Page_022.txt
962cbe03781ca20ea94e2c951169b7c3
1bb561a9645c737d05107aa4afd3bf832cb6b479
7283 F20101109_AABGOC hamey_r_Page_084thm.jpg
f6927eab1a46fc8501d8369b9a3c29bd
905394f20d3f1551410036a777b50b0f4f3cddfd
86688 F20101109_AABGNO hamey_r_Page_074.jpg
1a0e292d24d8a7eec7c836a738eaa69a
27040815c734b1dfda39697f783127cd5a5116ea
100438 F20101109_AABGOD hamey_r_Page_008.jpg
a66307743ca63d4b4394dfd5c9ab6a5c
f9e962f6c5d21cbdfea906194f46c7fa2afcdcf2
3424 F20101109_AABGNP hamey_r_Page_095thm.jpg
044358eacd0224a1700b473d8c3000ee
f8b987dc0206c405c056ee78d3f786b8a24214d1
50201 F20101109_AABGOE hamey_r_Page_052.pro
8f8c7a52a9925ca877d73136adaf29f3
baa93a2d9bb4e62967b98844624332b37be5a6a7
7261 F20101109_AABGNQ hamey_r_Page_030thm.jpg
62515cdcf1eb2354d6f8471056dbb2b9
3de6d2b48e5fb6fda6334fb3f57bdcda83047804
667688 F20101109_AABGOF hamey_r_Page_022.jp2
202510c44245fd0664bd37a7778e9776
a55e98443fcc9548022d6f977899e26c69a3e867
5204 F20101109_AABGNR hamey_r_Page_085thm.jpg
d75102bc3a4b7cc0d62e7a882fd86070
24320be541ab8973f189e3164b1ed2511332c716
35145 F20101109_AABGOG hamey_r_Page_060.jpg
5dfdd4ac03234aae289f04959084ef91
244e918f4d86ee18cf8a9c9f1e6b74be5feed4cc
555812 F20101109_AABGNS hamey_r_Page_104.jp2
978023dc7593963a61473d780b9f4b57
89c8c64e42f4510e98895630620429b3d8867d26
826 F20101109_AABGOH hamey_r_Page_137.txt
63dce3fe7596bfcf7de1bcd054ff7a89
4738fada85d2056ce7b1ef3419a61485f7dcd2da
2733 F20101109_AABGNT hamey_r_Page_109thm.jpg
641e9ce10ef5b7501b71a0c4fb565d44
4d4b9d28ac8faf4d6882eb5332ea790f09ae75aa
F20101109_AABGOI hamey_r_Page_084.jp2
8cb957f0b7869656127c63f332f72cf3
b289f621498c025bd5a342dab257e1ddc4d221b7
3434 F20101109_AABGNU hamey_r_Page_140thm.jpg
82f7afb06b96300ae2715c3ca509207b
a19288438d7207a9c46027a15dd9e28fb3c9b0a9
F20101109_AABGOJ hamey_r_Page_059.tif
973e7d3fac5ee499704cd5af2c48f937
71b4d3b85ff5b2c0fbf93c05497e6a3c19206766
F20101109_AABGOK hamey_r_Page_058.tif
c34e006f88a4ed4e6270ea10e84b03b1
7d659fc7a5f2e6f56ba78c12cf1f681762d19f0e
F20101109_AABGNV hamey_r_Page_101.tif
67e5a69fe0039e64bce7f8d32a2fcb9f
525c8e26b17e2aa6c4b82573e409860a063a9fc1
98275 F20101109_AABGOL hamey_r_Page_005.pro
9aa9da6b0a01737db07121b3e5a9d2a9
b1adbfa7e36d9c644400da59b3d173cac99df249
233487 F20101109_AABGNW hamey_r_Page_102.jp2
3f9d5a0b6d54c0b6b2539208bff02ccd
559320895d0ab733d863236eb6f3b98f0ab49e28
4358 F20101109_AABGPA hamey_r_Page_015.QC.jpg
491c17efe9244836f7aee96cf084caa8
d1f0950d8a5936c8b9fa9ceb44885f3a5e4edc87
13051 F20101109_AABGOM hamey_r_Page_154.QC.jpg
c9ae50b829fed2b7a8ef8761753aff80
d375600cf89796451b7b8fdd8242179d07fde168
56631 F20101109_AABGNX hamey_r_Page_027.pro
1c6de73264f446dfd991baf23b02409b
3b478a4e5a99d8cd8ceebac063c6dd05e409602d
13896 F20101109_AABGPB hamey_r_Page_071.QC.jpg
7bcfe75850c77e4d30519c4e52801d1e
4f55d2f40b5589c440f4dd75d9f622db3662131d
16763 F20101109_AABGON hamey_r_Page_066.pro
15958483163cd92598cfe22f42c22ed9
9bfdd15f24a420b19964b69fbcae3a7ffa7101ac
28121 F20101109_AABGNY hamey_r_Page_109.jpg
43a70b3c460f3af7bbf8de88cd5ba1e8
c470e99779e71fefddd645b1ddf69800a760a6c8
46559 F20101109_AABGPC hamey_r_Page_104.jpg
5c909316917bd1d7774a8ed197a4936a
1ad066db20300711a0002d42fa277b6247ebc56f
76648 F20101109_AABGOO hamey_r_Page_128.jpg
f195e13e4e0172e74c508627b8d1166f
65bcc8b5bf123ff32ce7016b9d32b9b5005d5340
82752 F20101109_AABGNZ hamey_r_Page_130.jpg
4c6ef001e97db7e5fac11eab3fcb4c3f
38a62a5ac9c55bbb872eb62f06e62cec869b9918
51829 F20101109_AABGPD hamey_r_Page_078.pro
c0f30dc275c58413e5aa8c386ac288f0
5bb50a1becea8fb63ef506624f7b93c06b3682eb
330574 F20101109_AABGOP hamey_r_Page_142.jp2
a79355a71da5e21359eb289d42f6950d
1aac6896db8469c796d28d908ea7d397d23eeb65
2179 F20101109_AABGPE hamey_r_Page_054.txt
0443fc7f344ec8910adf79cf79721e87
a1e49faa7bfc0f5b4e5ebaece8be6ad38b24b85b
6731 F20101109_AABGOQ hamey_r_Page_042thm.jpg
8fbf2e85c150bee439ac2f0a97212448
811d1cdc8b9db3b7d2b0b3e8c53c185f8071e413
3912 F20101109_AABGPF hamey_r_Page_113thm.jpg
f980147c7c29ae102da94b80ba038606
8fe34ae08a90e0757bf61eb19f759a2307eeaecb
F20101109_AABGOR hamey_r_Page_033.tif
496ee696cbb3df58f658bbecd8cc450c
8e5540581440703fa253c6379ba1e4dc14145d13
F20101109_AABGPG hamey_r_Page_139.tif
9fc277de858776136a1d879054060cff
bcacb4468986557c9e101cc98551692dc381064c
178 F20101109_AABGOS hamey_r_Page_062.txt
c69b6f94cf34eb77b42c7414249160f9
d3a66a0cb19b4e5b93f6c76396ff9293225c4387
42817 F20101109_AABGPH hamey_r_Page_071.jpg
5fcf4a23925bdd2bb3ece6b750989203
d97161c15d69bc9f6002b8954e43dc61390d5d0a
F20101109_AABGOT hamey_r_Page_094.tif
fc03ab5390e881d58911198e1fdd665d
1c30505c8d79a03e092baa50ecc82733c1b10bf5
35522 F20101109_AABGPI hamey_r_Page_036.jpg
66498540238fb54340236f49273165a5
f21a41ca7ebe1f8794d36997ea403becb71c6b14
1051982 F20101109_AABGOU hamey_r_Page_014.jp2
4f70017be6a6e06f68a7105d4b014239
9751ccb6a2a15feb2a58a0da71d2d451881839f3
609 F20101109_AABGPJ hamey_r_Page_096.txt
9a61306ebdabc2e458dbeb95d2dc04f8
9270a16128dc06f686428f1bf5affc0ef7770d45
33264 F20101109_AABGOV hamey_r_Page_106.jpg
6473555f9f90038cdd9f31d333c6ba98
942368988030af81f6702ef73cdee5b83c9b72ea
7905 F20101109_AABGPK hamey_r_Page_142.pro
dfa4b61b94ef8f351fa93066fbb58a61
65d477e1c584b559f9d38273233f988aaa44a314
26020 F20101109_AABGPL hamey_r_Page_151.QC.jpg
95c584b20b8680f88a0ebcc5a1d39b0c
20b55cc97a50b2f01bbd91b89e8520ec31890f3c
1947 F20101109_AABGOW hamey_r_Page_107.txt
9ffe2a04502acd7bee966c1620156f0a
1db8dab0187a4a8e9e61e2e33df7db2102be5c1d
84412 F20101109_AABGPM hamey_r_Page_072.jpg
fe1561fae25272a39314e9a6b1fa63d8
30a29c1aebb4fee59c4579a31dfb87e241b7c5fe
11337 F20101109_AABGOX hamey_r_Page_105.QC.jpg
7856c50466f199560f04a53ad4d3fd4b
fede208675c421789a579f4de60567a0c0cc14a6
12045 F20101109_AABGQA hamey_r_Page_118.QC.jpg
8647d80602148ebc36d47e33d2c9d564
549a6a18f0972df31b7eb56bcedca1a21a2c7917
7168 F20101109_AABGPN hamey_r_Page_079thm.jpg
5518f20e1293f0046b018c216f9e546e
34723af587933195c84fe3dbbcd71510e6fc2638
1051981 F20101109_AABGOY hamey_r_Page_048.jp2
8821b61840739a08ff68cf08dd3eccd4
5426e66a5651b87e052ec06830d116b1616a6c1a
3376 F20101109_AABGQB hamey_r_Page_012.txt
6c73c041275209978fca371cb42eb723
3c35392e300df541e6efef3b5a32f1822fe28fb5
7334 F20101109_AABGPO hamey_r_Page_047thm.jpg
c33d2f9085e1df1abbd54fe10025c4a7
d0f4603cc55d9338cd335127b030ba040e6ea19e
776546 F20101109_AABGOZ hamey_r_Page_060.jp2
1f6e38f56214628ec751f03b851e7e4f
9fa06425b07bd4dd09b76a5d340c7e82bf944ee7
6761 F20101109_AABGQC hamey_r_Page_034thm.jpg
e0623cf9bec5197890db4fb423dffd5b
c68912c9c36eb5fa2c1173627a4cf08bad69ac9b
1051967 F20101109_AABGPP hamey_r_Page_075.jp2
b115669a278027c71a4f36ed696d4cd5
dc9588a1bb75951236d605cb5d69281de516e758
70265 F20101109_AABGQD hamey_r_Page_033.jpg
d7a5f4fd53f6acccff29a1382c974851
2e728b4de934425fb0f17d95818ce2630ae5d9b4
1868 F20101109_AABGPQ hamey_r_Page_085.txt
57c4d8e8a12bcb189f7322da1f37bfb0
941642f63ad349aa4b9842790b6ff04a5d8be72f
18535 F20101109_AABGQE hamey_r_Page_083.QC.jpg
5077dc520e3483208c6fa74af02651d4
8633b0fa5cddcac1520cbee330eb59c2b4ace495
27485 F20101109_AABGPR hamey_r_Page_030.QC.jpg
c0bd3b80e932b00a18b4b7ee1b2fa8d5
29b168200aae82ed852143c72be209d9e21d65b4
431 F20101109_AABGQF hamey_r_Page_100.txt
0df655ebf1a26f504b3984a0a7927b6e
36c99c54e742dd565685b99b65b6e07ca4a01b28
12481 F20101109_AABGPS hamey_r_Page_065.pro
c27a1114afb9666176d39c8d1260d883
d738657c2405b970aeefc8b3931d32781909f558
288513 F20101109_AABGQG hamey_r_Page_038.jp2
e4dce82d33b08d2e1ab3a54e6093fb1c
493623e7b19f4219dafabbf9713f8b0e2f4abfe6
2111 F20101109_AABGPT hamey_r_Page_034.txt
d87616da46b5225c00773f1859bb7b8f
68f5ee2963d6679528217d0f069e5450d4e194ed
7274 F20101109_AABGQH hamey_r_Page_091thm.jpg
f21ac82ac85c7b6bdcff8330f0da0028
2013e78f90a8045fa968554a1c871e7d77ba8593
41309 F20101109_AABGPU hamey_r_Page_068.jpg
8d18e59e7f032c3b0ad9e7f76c3c497a
1983aad655e18a86f00ca0e41f4902ac0c84fd70
2054 F20101109_AABGQI hamey_r_Page_055.txt
6e5f51c9429c5db3d125252688ec43da
2d91d82434576f8d877af126760f5d7685941c00
53972 F20101109_AABGPV hamey_r_Page_079.pro
538366168b0398907ab06a65140e473a
233e715395d0170ec1a7caf739bc72448cdd71aa
7517 F20101109_AABGQJ hamey_r_Page_145thm.jpg
7c7c805c4810c03f8d43f7dc91407e8c
86e36e7863b0b10579ecd472bc3fe3528045e879
27783 F20101109_AABGPW hamey_r_Page_145.QC.jpg
0df0a2cb47b3fd79cdedfdbaa04682fe
4d4b3398c831eb70d42280c1610c966c72f7a4ce
F20101109_AABGQK hamey_r_Page_093.tif
8aebc0662434723f7c32e2b37fe965c9
f8c73911fcc5147113b49fa6376ad11eb9b50980
2164 F20101109_AABGQL hamey_r_Page_090.txt
99f15accfc7da55ee8832f3ec0f5fec7
0bb8389fba1eecc6874b95f4d3c87cf246753932
50108 F20101109_AABGPX hamey_r_Page_046.pro
0efc8737f0fbbeda9168c62aa4e70089
012c3eabb24d16da0f8c3f89cbc17ca9082251f0
55441 F20101109_AABGRA hamey_r_Page_050.pro
35704e41d097ff07f404afce690d710b
c6e47682e5a570550d53195605c0481aa4acccc4
33676 F20101109_AABGQM hamey_r_Page_004.pro
c92ddb02a526a0b95c5babb5673ef370
9801615b70cd2a28f0a8479cd90b14cf054cc3b8
552344 F20101109_AABGPY hamey_r_Page_067.jp2
71083209cb243ddd062f349c604f60c0
6c110122eb0b45634f113543bd1a8bf9ad673451
53358 F20101109_AABGRB hamey_r_Page_072.pro
97dca0a5ec1f241c48ba3fc52603d5a0
1b0d0a2bffa298763e6d19f69c72acce2d8ab907
1051985 F20101109_AABGQN hamey_r_Page_088.jp2
8af55f7ecea954ba00c647a036d01750
505d6295ff4b4146b1e39e31cc7603e9076c4b69
16809 F20101109_AABGPZ hamey_r_Page_117.QC.jpg
d4684efe9ec78c8d89508d3b6e161f27
d9d898a5752466a3b4c79ecb8aff4ae3339a2bcc
85941 F20101109_AABGRC hamey_r_Page_087.jpg
9616d30cc38241427e97c109f74f990f
e98a746974ac2c6bfe868cfa980d48b24899961a
57873 F20101109_AABGQO hamey_r_Page_129.pro
25baf28f58b19f16596bc403bdc3046e
bbbaedf1a717ae348601aed88315506e91c88504
20008 F20101109_AABGRD hamey_r_Page_063.QC.jpg
ad1ec4a483d862f4d2db3c3e76f7bd81
8096423a2a55b570aa1807b263336c6a3f543f5d
22302 F20101109_AABGQP hamey_r_Page_005.QC.jpg
3d542fde8fb83e5a7c68a7d0e2db0061
58475893f4a7cddb87a347ca872570b90b7c8836
26539 F20101109_AABGRE hamey_r_Page_132.QC.jpg
fb92ffda365fafd09e6b74d9a8d58680
f4bfed965f1cc4f2e4da66663bb0b881b507309a
24844 F20101109_AABGQQ hamey_r_Page_014.QC.jpg
1616ab9b30ea590ad4217f2266ab5d87
f247f07e542b4678ecd9b7f211c4786c8dff80a8
25856 F20101109_AABGRF hamey_r_Page_111.jpg
19f96b27bf4caa601b9f50d68e0cc41c
a70ba040770b5603a91e783ea3a2b4df5c5a0d64
F20101109_AABGQR hamey_r_Page_124.tif
dd7986e07cdd3eb37830b3513efd100f
5661ae788d0e357ba6125d30f8cfb7ac7570b4f3
14206 F20101109_AABGRG hamey_r_Page_137.pro
821a9c4950e19fff4427498c8aaa001c
05314aaf0c0c9d871571490a47ba20a73f10e2a6
6808 F20101109_AABGQS hamey_r_Page_014thm.jpg
9bd56ed039ff36b9d87326433358cc55
09bf3e21ca5b0b110aa4e9caaaf639fc89ab4a83
5655 F20101109_AABGRH hamey_r_Page_005thm.jpg
dd3cea9a60905038077bbb457eb1eaae
1c879bab4a043566df8adf21c3fbb8609ce87362
4789 F20101109_AABGQT hamey_r_Page_022thm.jpg
4ff1521c171b2c031a231afe195e78de
5eac048af69ca218ad9e8ebfac3aab5f1fe0f734
F20101109_AABGRI hamey_r_Page_017.jp2
044d2f2503757832a5148bbc5ac7c2d8
da38b5dfd11dc5fda1aa9ab2162642881c34bdb4
766541 F20101109_AABGQU hamey_r_Page_004.jp2
9c0b8e4441b9951aa0d5bd6c2a5f0665
0fd66040dd62dc3b73905b1240308505d0c9746a
F20101109_AABGRJ hamey_r_Page_108.tif
ac9023eeb075cf8b52e3ca4155242333
c6e3564b4162a08d1564fcc57a2169fd0e13f531
F20101109_AABGQV hamey_r_Page_037.tif
252225bfe6a969d7ca14443a696e4eb0
91edba0aada50c8fa3ffd8f457a8bfafe247bec4
56207 F20101109_AABGRK hamey_r_Page_057.pro
2f1760f511410f968b67c83db5829bfc
5f30b437e3792b5aef8b4c491b57b673dabb21a0
52704 F20101109_AABGQW hamey_r_Page_017.pro
7f21ad0247d53cd35d304d818aca3dc4
1cb18fadd1b0993382048c36d9fd72106e7f9cd4
29878 F20101109_AABGRL hamey_r_Page_110.jpg
dbae476730d6a707398b43f65262b5d8
11b314fa3a3b7043119e888afcd23d9e618893e8
2105 F20101109_AABGQX hamey_r_Page_132.txt
19f5da932d4ef5908dc34dc36f549dec
96fe117a5cd878bade02c33f6ccf329cc700c08b
591809 F20101109_AABGSA hamey_r_Page_120.jp2
c2cd8748266636c99b9dbcc7f7c5e6cb
5aff5ab53a45d537890d7b9d129542c436373e09
2197 F20101109_AABGRM hamey_r_Page_026.txt
8b2c0f4db9807a070751dddf704f9c96
c3e6379bd1e48437c5bd30b5a8c7f19b769285ba
F20101109_AABGSB hamey_r_Page_029.tif
2c3a2fe0809a1ac3aed8d54059293293
57fa3789dc9dff09720ce1ee2a75485b4d6dd70f
87822 F20101109_AABGRN hamey_r_Page_054.jpg
a5c5b6eca6a23be9f8e2c77ffbd5edf4
1d9296a04d9b65cdae1924812b82ccc685ae1576
F20101109_AABGQY hamey_r_Page_147.tif
fc262bd3a137fb326297c8bf7f9df923
bfcba43184ed98403fc6cc8a4c247007c4bf2b72
1963 F20101109_AABGSC hamey_r_Page_035.txt
f02792a7cfec18084fbce8c3f60ae495
40a968cca5f170b117faf32a32075e4f7c520de7
461050 F20101109_AABGRO hamey_r_Page_036.jp2
cccd0f73acb7a668b7af8b6deedb2a6c
e05b30f21a9d762893566e8a3a1bd8f824ba88ce
41959 F20101109_AABGQZ hamey_r_Page_135.pro
b9f8ce9711c5e918c090f8935c4dcbe3
5e60e2981e09b23697de4f6f71140f936e494c82
24391 F20101109_AABGSD hamey_r_Page_075.QC.jpg
61d8e46ffe0231d04c78fe6aa09259c8
845d27a630c26afb785325bf6e21e99aec68062e
F20101109_AABGRP hamey_r_Page_149.tif
5e0089e42d2cb32cd2106b09894c3ad6
eb6fb0c7121930d22420b7d43483cff02415b29f
28917 F20101109_AABGSE hamey_r_Page_114.pro
9f1dc74211b8e0b4e3981f23e0407c11
0cff1d691e4c6fc25100814eb32fe604ff744114
54600 F20101109_AABGRQ hamey_r_Page_058.pro
df274721840309ab4c31b4f750d518e6
19d81d4fa4a228cf4151e0779c030a58f18db8d0
25641 F20101109_AABGSF hamey_r_Page_053.QC.jpg
311f315bcde83a6eaa92d5a6b5c13233
a8415cd6132af738ce92b6402eed1ac6e23da4db
6056 F20101109_AABGRR hamey_r_Page_103thm.jpg
935ca0b5a9b78fcd73fae3f069eb2c21
e5221ea1f696741fa7fc68048eee6dc1c4c10465
36389 F20101109_AABGSG hamey_r_Page_020.jpg
b68b8ef54cfc64e1de2576026a5def7a
41fe44f8be88df4cb5508e80da57baafdc1d0dd7
9806 F20101109_AABGRS hamey_r_Page_070.QC.jpg
85d008669fdd9635d557b00f75d0de30
122e1df4fa6ad398cf8d94d98fb3dc03aef03c68
28413 F20101109_AABGSH hamey_r_Page_088.QC.jpg
3be17f8ee505c29d095fee3cf5488919
5fde1de63f2d0e31427d07b1ce12a74a78b14aef
2202 F20101109_AABGRT hamey_r_Page_045.txt
b06b0c3a9534870bae5f8da4a670018f
1539ab2a140504890120ade633dc3920313f9a5a
F20101109_AABGSI hamey_r_Page_036.tif
b9326f66dcb344605676b83d1b410384
acef6ecffebc976d089b5a7561c454f3840e115a
30718 F20101109_AABGRU hamey_r_Page_153.jpg
2b61400e9eae1ad44b52bce1a2c483a2
723ccec061ee08cb4b39183b5938410177dea627
2205 F20101109_AABGSJ hamey_r_Page_057.txt
6164f0e99e941911a3bfe9bb53452be9
83af642222fd71e0b563ef1534a05e5fe9d69081
203 F20101109_AABGRV hamey_r_Page_060.txt
2eb31194dc05d59eef27b4231e2da2da
40545a08637101ad3e25f4c9d51d8c57a8725011
F20101109_AABGRW hamey_r_Page_073.tif
f90179aac6538c500653c9e21b858e41
44e97d732b7d312f01d1bc6dec2f60a3c8edc29c
2282 F20101109_AABGSK hamey_r_Page_001thm.jpg
1f98cffc06d50aea6a24366eee83722f
16ac0d2f075f80a78d9b708a3f059cda1ee802ff
602 F20101109_AABGRX hamey_r_Page_111.txt
4d426bafcf3733de938a3c29371afaed
2b62d43b52e3640511a5af1fa9a5612dab15604a
517 F20101109_AABGSL hamey_r_Page_121.txt
afb98ef057384e434607c86aa50f3302
79832e365cd60923f991d2ed4bfc9a87ee286a5e
7555 F20101109_AABGRY hamey_r_Page_129thm.jpg
a4fa9b6919ab8c6e23c2f02a14ff1367
00a44f7088278a81d6322dfc3e4d0ffde0af9862
481467 F20101109_AABGTA hamey_r_Page_065.jp2
304f137235f5ff4be5dd32ecad721cd6
8df0b1ce1f37ede84c835e6864cfd94de78d73d5
29767 F20101109_AABGSM hamey_r_Page_011.QC.jpg
589237f6207aaebeb5e8fb549870fa00
0aca1036ecdbc00b0506cf8f854833d5a1e2d534
80982 F20101109_AABGTB hamey_r_Page_152.jpg
2201516c076e8b6c8ef1b4830761d3c7
c115587803b629b15db782e79d0acc3c1885df63
2823 F20101109_AABGSN hamey_r_Page_112thm.jpg
740519b603d2ed170c2791d3dfecb1d0
2d681a2f2976522c9a192f5da528780c17018a77
80381 F20101109_AABGRZ hamey_r_Page_008.pro
0d0985adb9b1676205b6a087af7f5d42
3ed45644417c168e46d7437d56b5414bc05a51fa
10013 F20101109_AABGTC hamey_r_Page_097.pro
e0dd807cb2a2ba9e2034e3a1ba6655ac
03a021ab104d5efc6cb587d87e370879f1af6fa1
F20101109_AABGSO hamey_r_Page_067.tif
e7c311173ec3bcd98c9cbf31208ef79f
4703d37ab4d12ce49ccbfd0ff22eaf4d515dd5ed
F20101109_AABGTD hamey_r_Page_004.tif
e6bf8e76d3f69b41c5fcd19296b41ccd
d366e67bcc3bbef8b0efea5ab1abe2a03efb216d
715231 F20101109_AABGSP hamey_r_Page_116.jp2
ac1605e44dc3d95881cf0ec1907c483e
e78da6fdc02967021d73dd282d2953c322306327
2271 F20101109_AABGTE hamey_r_Page_146.txt
facd802680bca71a2c89862a862b6871
1f82583a25cedf238a18075bed28f2170f41b604
498503 F20101109_AABGSQ hamey_r_Page_107.jp2
aa7f5af32c3238c72c77550476a7f157
8b0f6cb2589de43c4a781c089271f6b0495f6bbb
44115 F20101109_AABGTF hamey_r_Page_113.jpg
1f180f1ff47d24e4612505e318452cf5
478cee34931d0f905c470c57ccb206e8b508ae3e
88586 F20101109_AABGSR hamey_r_Page_129.jpg
c5d3dbc2f8a0ad50b1170e329193c593
d8b7a4f811644eb9dda0d992373ba11857dcb200
54750 F20101109_AABGTG hamey_r_Page_086.pro
af16a956c7277ca8b830035dcecb45a0
577881ee72b63a8e7efa2d67b14c0d611ff900f3
49679 F20101109_AABGSS hamey_r_Page_134.pro
d67e07a0b6a159e1e12e21ff87deac00
2be445b4191217108becec6eb8163a2e0e686099
F20101109_AABGTH hamey_r_Page_149.jp2
28e3f6b377ea269c067bd45f83128de6
b6d67518f1d55228dededd36aae56c4eaa05529c
514 F20101109_AABGST hamey_r_Page_119.txt
432615ddf882a635b31a653fba6d6e36
319ee78457a83afa19e53747432ebce02b370c0a
792232 F20101109_AABGTI hamey_r_Page_009.jp2
e4834924e4e8eb8cc4517117cabbc22c
771187e254778fe6e263259f017b387bf3e44811
1660 F20101109_AABGSU hamey_r_Page_015thm.jpg
16d88682b686edf850b476a8f89e6a72
ed2da74c612a775d5a623c94c7df168b47c52762
12316 F20101109_AABGTJ hamey_r_Page_138.QC.jpg
e3d1e6b98f6202f128f2c64319ab3f39
113a12846e43096e2441ffe843ea2167c78911b5
2949 F20101109_AABGSV hamey_r_Page_126thm.jpg
958669e53bdc335ec3f6d2265d0e7d3e
dfbc3866ec9af85992d945ab5695fdcc137359aa
54072 F20101109_AABGTK hamey_r_Page_029.pro
218cd4b4c856d3b06c75e82e9c72f5ee
b713caeb732baf1dc2fb81a44e2e69a1027e770c
516875 F20101109_AABGSW hamey_r_Page_154.jp2
51adc714d6ef71552ee3cdd67441d886
3dcf44f873a018e0cc7501525e87eca8ef18e402
F20101109_AABGTL hamey_r_Page_066.tif
790da5adbce41543da62e91fa56c25ab
1b72fe9f6838c5a93dc9293ab16f09d046f7f55b
F20101109_AABGSX hamey_r_Page_081.tif
c586e506084a8d733b5099f64eb69471
c1848bdc25767d0dbd4e639c95677c0ba65bf217
85401 F20101109_AABGUA hamey_r_Page_058.jpg
5a0130a2e9ac063e41f9587c20453392
81030448825a0028696bdc3fa240b45257009d79
381102 F20101109_AABGTM hamey_r_Page_105.jp2
f38e7fa8d7f4634e5f0cbcf54129d02f
e8c099883c4633d830ae35955362d835001cd804
F20101109_AABGSY hamey_r_Page_151.tif
542a14f3dbf68369d3e924ebdb7529e4
c4ffe484c78dd0475f926e1defb872608d250dfd
15268 F20101109_AABGUB hamey_r_Page_095.pro
7825a5977d83877f59d82bbe25e5b38b
b8a92083328002f8415159edea818570ca48b8f5
3077 F20101109_AABGTN hamey_r_Page_010.txt
33cc173fe83e344085b080106f6b8e6c
acfd16032f528e17899e87f2be14cbf8536aedfa
F20101109_AABGSZ hamey_r_Page_001.tif
e1977577b3ba9f4a1e4aaa6f40de17b8
7a45a307c42b478a96db378229905c21b24aeef0
1794 F20101109_AABGUC hamey_r_Page_148thm.jpg
eb3158cb475977feb51410716a4f8e27
863d0acb55085c941870583be7806bff55224fa8
580713 F20101109_AABGTO hamey_r_Page_019.jp2
4aaeacb14daa3398e05ded3fe48b7c12
44e24b8bb0d5cfa9808ca8b88c537959ab4f9c42
372 F20101109_AABGUD hamey_r_Page_039.txt
151f1aa2a3e327e69d17f2aa1b520b07
bbabc52d1a7334df48dd8f86efa779928eb364d8
49285 F20101109_AABGTP hamey_r_Page_067.jpg
c5d854100bd65fa2028851eb6f61cf49
6feac4841969035399c52092d49941ff956c5648
136 F20101109_AABGUE hamey_r_Page_125.txt
0b42ed79d53d090ee841f3464719b33c
3820c3a55cbe804dfe403e8229ba9230b1eec764
3789 F20101109_AABGTQ hamey_r_Page_122thm.jpg
0329f2d7369cac9ebf3165852dd52e67
414b4042fd1058105c337bedb7a211acfd896e5e
79284 F20101109_AABGUF hamey_r_Page_082.jpg
93ffb7fdd5eea3e3a364fe35d60ce610
82cdf0ff0e2bf5486fea3537da16ec1360ae05a6
46862 F20101109_AABGTR hamey_r_Page_127.pro
0b89497d08b361d974c70699b12d7981
7b9926f17322dce1b3346b37440fd55227fe4ad0
1051920 F20101109_AABHAA hamey_r_Page_053.jp2
8a1bdf9be5a7ef3102b92d7e86494fed
9bd354354d0b1a9a5381e58af6abf4aa4846e49f
5371 F20101109_AABGUG hamey_r_Page_060thm.jpg
57748d74cdc9bcae4b7e3cb9f62d1af6
39b7ae027230a8da50b917408ba20427de87cdb7
86415 F20101109_AABGTS hamey_r_Page_027.jpg
107ecbf3ee8f1cbff64a27ae62b024ec
3ef87264e5140eb690d679729569b82f886d8027
F20101109_AABHAB hamey_r_Page_072.jp2
f90e791b1b41d210d7388ffb7d88e1af
bbd0b42b6ae9d9e5b6e150472b617c29574afc4a
1051924 F20101109_AABGUH hamey_r_Page_031.jp2
084d3e74e8ea27d7ebf010b117bd8e66
e83f86c388b57c086d6b18d698bff7da39a723db
6904 F20101109_AABGTT hamey_r_Page_128thm.jpg
30f23abdabf52813f985580a59fbdb0b
301af3a530fff7fc60abbe80aeaf276434771f8e
1051936 F20101109_AABHAC hamey_r_Page_073.jp2
ad9ee2fc670763dca14b09dc5b3623f7
d8fc1a5b1e495f7a87cb1b37d69b18715a6f3785
27482 F20101109_AABGUI hamey_r_Page_008.QC.jpg
a254ce3e1936e3352fd8a8da72ad2c73
a89c216f448c014fd07d54f858edf4474d0e6059
1626 F20101109_AABGTU hamey_r_Page_113.txt
00adcd1695373128470d7a9796f73f24
d0c7d971dcacc7d5213d97fc0521b619b8d51f3f
980999 F20101109_AABHAD hamey_r_Page_076.jp2
fd015d8c5d9c85c1930f72bd18dca212
acaf0b48c6d53dbbb0cfc5dcb8a188f674309ab6
F20101109_AABGUJ hamey_r_Page_095.tif
a1778192efd7e890bc7663442a174ab3
7394742e1e15f67e780af7f139d353eeb8f74465
26263 F20101109_AABGTV hamey_r_Page_130.QC.jpg
c0ddf1d5d3245a2900df96d85eee2db8
33e6005180a9bc4f8842f329a5f269c64545c60f
F20101109_AABHAE hamey_r_Page_077.jp2
0756182970a67e1c0d984d984bd63ec1
9c9b406bc71d3ea3acd2e6fad33818bc59e97449
15132 F20101109_AABGUK hamey_r_Page_061.QC.jpg
5948345500eb920bd1dcd3a2aa942301
ef272929326e35ad206bb3355339c99d7581cf69
F20101109_AABGTW hamey_r_Page_018.tif
cfbe836b24ed18c527b318b3c88b39c0
4810375e7d74f92ab6ab041e50a99a3d3f50a0b7
1051968 F20101109_AABHAF hamey_r_Page_078.jp2
7ac81f62963e65e68b844b7231f3ce49
8123a6ab1e9cbda3bf924510ad1ec1ba69e02f13
F20101109_AABHAG hamey_r_Page_081.jp2
b53f43e51daa49473b5791554e044030
6bc984d77967f48e9d8b0a82d0941b5fd0c811a5
2079 F20101109_AABGUL hamey_r_Page_017.txt
da36ebc9a6b0b8c73c20e8dac9e8beb2
e5476d54f944f683042241ea328bdc15c8bd8489
11108 F20101109_AABGTX hamey_r_Page_066.QC.jpg
ae90f2882146dc95cf5a1a19268c740c
42e8c8e6bdfc14b221518d504b1ea368c27b5a7c
30147 F20101109_AABGUM hamey_r_Page_022.pro
ea562fa7997976f8490c3fe92d6b06cb
b92868df56929a91efa747b57da8d914f1f51726
3659 F20101109_AABGTY hamey_r_Page_116thm.jpg
b872070a24eaf6f9c5251ffd2cc8844d
40694045976b0ed895d1e4caa5dd0d091e709456
2221 F20101109_AABGVA hamey_r_Page_080.txt
cb71f3531de3b94403c7d9f6c04003be
7c077d69d49ab6308f2dd8f51678c7f41af1c35e
1051979 F20101109_AABHAH hamey_r_Page_082.jp2
02f5cf5ad4f0d57b5d942dea52c87133
a2bc1055ec335ebdd3cb253d77d3e0954dc4cf49
54537 F20101109_AABGUN hamey_r_Page_090.pro
a7564bf54682ccbcbee40078b08780b0
bcdd46ae401aae9852775640913d1fca86759522
783 F20101109_AABGTZ hamey_r_Page_065.txt
cc8bd4f46e79dffd65667be78777abe9
2c6e1965e0da817ccb48bc2731ec15bbb0850f8b
466034 F20101109_AABGVB hamey_r_Page_040.jp2
b520f252b0132c84edee184c1dbd0d4d
87b2979d8acbfcb38447f87b68ca75f99d5642c9
F20101109_AABHAI hamey_r_Page_091.jp2
428752b6262f0e6c8f612cd0a3f637a6
d2dfd8ab2057c06a4d6064275b52db275d490f6a
1051965 F20101109_AABGUO hamey_r_Page_132.jp2
9dc745867ffe0d0ef1867eb28a4a6ce4
039178bd8bf670d346afc3c6e7d7cfdaf092207a
274959 F20101109_AABGVC hamey_r_Page_097.jp2
7d107c7b1fecf7f2fdb9850b1fcec080
4cd9b794e09554bc98c80e1cdf73d66128379e11
F20101109_AABHAJ hamey_r_Page_092.jp2
ddd76bd585aece0e219a2a34ace859c1
214cf013f881f7a9e231c40e298f83fd9acba984
82663 F20101109_AABGUP hamey_r_Page_132.jpg
930b2a1b22444c040088c7f9bde9f84c
cc04c51e6abdd9b18cac3b67e785853a361fd997
F20101109_AABGVD hamey_r_Page_152.jp2
87a5845ffeba2c7bb8bfae47a73e1f1f
8766f57c9fe490e5d0c4dc4363d109d6803fb88d
34334 F20101109_AABHAK hamey_r_Page_098.jp2
af19439fdacf5e3fc6a07046e5cdb130
c2ae15da983e45c5203f6e3e6d75a1c950768268
2106 F20101109_AABGUQ hamey_r_Page_043.txt
616679af1967d5659b9307d90a3fc6f6
68f1f82488272fa8b1fae2f46adf16b880ffa722
14401 F20101109_AABGVE hamey_r_Page_148.jpg
d873c0b1d8a3e0f819cddc94c95ad64d
e198450a88d9c0798c996564f8b4f415823f2aec
1051939 F20101109_AABHBA hamey_r_Page_146.jp2
39a1e20392a9e1c7c1e747d037d4369b
bd6003ec23d3711f6e72144bd6bbff2580aad695
91991 F20101109_AABHAL hamey_r_Page_101.jp2
38c6f821cbcbbbf9379bb177726fb175
f8f2daa58e5031c50f677f642c626fd77cfc82bc
7432 F20101109_AABGUR hamey_r_Page_043thm.jpg
4ee0f99679a7cb0caef7aad6edea027b
0aee875b69ec6b1f23b6373f3ced709d3b0f0f41
F20101109_AABGVF hamey_r_Page_039.tif
3131194bf87ebd541c4034d8c00ae65c
1f24b9ffee5eee73eb64ea3e8a4d26db7c882308
1051894 F20101109_AABHBB hamey_r_Page_147.jp2
1e30e6683878983cde9dc67fe0d0501b
026e5a2a951fc1086e3d97f393463fbf13809640
274240 F20101109_AABHAM hamey_r_Page_109.jp2
3b1fa4bd8158d02138c8d3ce8caf2dd0
a9adb6fa127241a0fa125be7f1e4bff1e273cd8b
7391 F20101109_AABGUS hamey_r_Page_086thm.jpg
eaac683b856a89c810ce7289f1433e0d
483777f6311487cce372ae8759e59f7350dc11a5
1051976 F20101109_AABGVG hamey_r_Page_064.jp2
d8a1cbce0745c0e838f030512034ee54
7c72e35a714d99b03802fa738c5f56d69f39cd4b
F20101109_AABHBC hamey_r_Page_151.jp2
44921c00105cf44d9a40bda0be583cf9
f67273d12bae78d643208b3eb4c5b50a44026ba3
299474 F20101109_AABHAN hamey_r_Page_110.jp2
364fcbb8f7fe654d0f282a97b9d6a0ec
18d5063207655298e5fb7b960c559979dec797f8
6659 F20101109_AABGUT hamey_r_Page_127thm.jpg
a7b9cf8f2e46d277be2ece4c5e4507ed
d3092d98561f6c5262d1f6a7a6a2ea97ea97f2cb
1012 F20101109_AABGVH hamey_r_Page_019.txt
a7566000d4e1ed81961cdfe9a5a3870c
d79ae5c5c09f716bcf32ab75d1bf77b6677c7edf
F20101109_AABHBD hamey_r_Page_002.tif
a9a3f4f14a86e6440aeb6d6b0f21323b
fbe311dffb2e1ae32433f95b1d5a0ec7eb374f46
255776 F20101109_AABHAO hamey_r_Page_111.jp2
a1da41460b7aca037a3e24ebd3f6f291
ec116884ba0e8bd1c8dfbacd07f587fbca20de0c
1051921 F20101109_AABGUU hamey_r_Page_128.jp2
f7959daf72335f95a6cb8ae5a6f637b8
3fdb8885bb9a681094c9d874994f3c3656f94601
27721 F20101109_AABGVI hamey_r_Page_031.QC.jpg
7ddfe3718405435552b2577bfc90c786
b613c0cb1bd05a8a977298fe34bfda12c2f35266
F20101109_AABHBE hamey_r_Page_007.tif
27650e37abec26ef8835ff470656675f
34c309985d53a13ef95b1418c13671a08b3471af
1051966 F20101109_AABHAP hamey_r_Page_117.jp2
71f4a3cc5106611c7a976ce307227772
db1c0f207fce83e1b424081a7deb73de560ddda1
444 F20101109_AABGUV hamey_r_Page_040.txt
d176c06123ce14d3a4cf1f03e18b1cac
7254c4cdbfd6e31a972b627fd4d606a6dd7cd086
27701 F20101109_AABGVJ hamey_r_Page_129.QC.jpg
1a0ae8606c942dd6d5aa43ca18b5a925
27918e6758767396b3d0d69ea95d17993db7e1a0
F20101109_AABHBF hamey_r_Page_008.tif
3504ad6ff349e574d8a713a462eeb177
1b175861dc1bf024d7dd3cab645490cf8f86a218
30762 F20101109_AABHAQ hamey_r_Page_119.jp2
8a228d1504cc77430bf7230b8e5d5b42
0b8bdc12f3f5b735f0d7064a82bfe484d74c6a90
F20101109_AABGUW hamey_r_Page_038.tif
65e9928b880559e94606dd784e076020
c9e8b6124d853daf8e0e4d7eb6ac2511b1a24023
18336 F20101109_AABGVK hamey_r_Page_141.QC.jpg
1f127375deab7e2945ee01bf58a7ca1a
b73fd0f87102713ef369e5cc0d708ae4fd2f9fd1
F20101109_AABHBG hamey_r_Page_010.tif
3fb263115145eeb42f4c54bfb3b5b1b8
e87d313afb72d6498aface0cde0923dca4fdbca8
296580 F20101109_AABHAR hamey_r_Page_121.jp2
9b7c64c832bc99eaaf3358f958a8c6c6
e5b08696315c08080d12d2a13119394ecb25fd1d
51214 F20101109_AABGUX hamey_r_Page_144.pro
982262fb85b1a64be24ab08a9b250494
2c8532846dfb04c40509ab15f0d3291a581dd78e
2150 F20101109_AABGVL hamey_r_Page_073.txt
4886c66fb93d9ac5222273cc9b146b7f
94571cbfe9985a18412184efad5ebd5631af1faf
F20101109_AABHBH hamey_r_Page_011.tif
f83ccba6441616ed96b36b1e9d3763dd
840b24669add3519bbd528dc2c16b6526f2bf672
1051898 F20101109_AABHAS hamey_r_Page_123.jp2
aec4d45c4e80e79d6a2035dcebd5636a
0f3616c5618778e4c36cc5cf33626ed1490bdcc1
28542 F20101109_AABGUY hamey_r_Page_012.QC.jpg
87bc37226c999ef9ec603bf2f18deb0b
20963271ede879fe2003fc809fd56658052fd274
3201 F20101109_AABGWA hamey_r_Page_119thm.jpg
43d5cf739943a5b5453efbd4c8a277b4
341ba0a5186dc9b42977e8d2861559f1a28b2433
84905 F20101109_AABGVM hamey_r_Page_026.jpg
399cbb846b5e7abee75e1a4163007535
ae3a1fb8ac81028c4030335317aca07381f5a0a7
142275 F20101109_AABHAT hamey_r_Page_124.jp2
c91ae4b6699f4ca6c9860ab4c92882f6
7742081a90bbcfd6197660bf30c3b2c69b04af08
F20101109_AABGUZ hamey_r_Page_043.jp2
9d520d8b31379f08a0d38af3a742d0ca
b50819cd2fd40d5cb2fcd6998e5d1a05ecd61323
56580 F20101109_AABGWB hamey_r_Page_080.pro
4988b33bb3373f310e27be0f80fba32b
28592ea760d474ec445ff26964524a7cf9269063
2861 F20101109_AABGVN hamey_r_Page_124thm.jpg
b7a9d52811ee5dade331e9beb8169504
393d9fdde9cb307d103fa181c3f55779c63289c9
F20101109_AABHBI hamey_r_Page_012.tif
f48c66bc8497f7c2b3a17040483bd63a
c2c0a912a361178ddbc48c6e7d5d038fb80bfd11
F20101109_AABHAU hamey_r_Page_129.jp2
01244fb9ba45bd7617aeaf71720e4bc4
fb9a4a13f3d23a3153e5f655d3a748bdc7a470ba
11599 F20101109_AABGWC hamey_r_Page_037.QC.jpg
771750f750672b7653ff25abccca93bb
878869f4689a54777f067b714e9f10a834e3bd00
415680 F20101109_AABGVO hamey_r_Page_118.jp2
78661cfae9c0b1f26823071919e7f2c0
29435aed87b006f9a7b6e267327e40f5860544ff
F20101109_AABHBJ hamey_r_Page_014.tif
8e8053d68154e6d60b22ffdb3602003f
f34117fe26d4f4c039e5871822d1e3fb1e3957a8
1051935 F20101109_AABHAV hamey_r_Page_130.jp2
80076181695a62aaf53d9e3b26496189
fe768eccfee0a0902534438dabac366efd83790d
10609 F20101109_AABGWD hamey_r_Page_106.QC.jpg
80c27bce118cd5c799244f9d480dd5cc
4155dd02264b6dca6780860adbf72c66265547e5
84569 F20101109_AABGVP hamey_r_Page_012.pro
8c877bc37377d08cb3091c04283eff7b
1b7e2dd0b809cc6d6709edba60c6fcd9f402e5ff
F20101109_AABHBK hamey_r_Page_016.tif
1685ea24cf31ddc085f9b333daf746c0
924d1c87d626bea900d2d6abe81a6e40a1cd9edd
F20101109_AABHAW hamey_r_Page_133.jp2
170d851cfd5167f62e9d0f2f59465cdf
a9fa81dbe21a0c6a61ebdec0b95fbd1e4b764102
1051957 F20101109_AABGWE hamey_r_Page_035.jp2
2d76d03f75c72857a61323d0c40df123
99438e0a9bb5e44684907816fcc678ebda5bda51
F20101109_AABGVQ hamey_r_Page_023.jp2
9286e7ec4d3b9776ced46416fdbb7af0
573b66968d9e141929b64a6167ac30d898a47d84
F20101109_AABHCA hamey_r_Page_051.tif
9792377aa0db2df3ad2e3cbbf2acd991
91f0e8c294b98d2528cb0fc5b3f1df7588e32397
F20101109_AABHBL hamey_r_Page_020.tif
07318df80bdbc8a6b0fe14c3a06549c2
f8d80f5f0c3aecb33932e6cdfd7309d7be75875b
656136 F20101109_AABHAX hamey_r_Page_137.jp2
14907683bc19c7087bb5e5c984e77021
0a073ab7c1921892164f17666005c824d02003e1
882817 F20101109_AABGWF hamey_r_Page_061.jp2
d1e21a2e3e774f3edd4e8b1cc1bab6d3
758c0231323b706162d6ab160291f33d8880bf78
3881 F20101109_AABGVR hamey_r_Page_098thm.jpg
843195a07632fd516d498924eb1ac6c9
c7cc7cec39465920a0d796106d5187a18fcca142
F20101109_AABHCB hamey_r_Page_052.tif
04f7acdd168eaffc8c5dc37df3a2fba3
a20c62aeae060933a4608113cf5c20be5b9bb550
F20101109_AABHBM hamey_r_Page_021.tif
ff3799dec25e2e192ead3cfeff1b9d01
3eb612ce284fefc366ea79b7aca04933b79425c2
590642 F20101109_AABHAY hamey_r_Page_140.jp2
ec4577990fc548ac3920b40c717a2d16
ce73683b46473df35f73b3d42132fd780dd8e294
27592 F20101109_AABGWG hamey_r_Page_074.QC.jpg
9a367e4c547763c001b25e32d2c9f353
8d3b7392b6c42e9e00fd880326a0821d20224d72
1955 F20101109_AABGVS hamey_r_Page_051.txt
39b9dad556ac11a431b4aa48874713bf
0dbf3331a5168b7871acfef5ecbfee6ddcd7ff41
F20101109_AABHCC hamey_r_Page_056.tif
b6d074388c439574248d08af8529d05a
0d66ca6b1dd44f27f9a9e7ab39f2ee3b69cd6082
F20101109_AABHBN hamey_r_Page_023.tif
839383128db06e9b6c449c5b42283524
06b0040ccd7c60b95360c7f48db61444b9005686
707518 F20101109_AABHAZ hamey_r_Page_141.jp2
4ea0c9052a1b98146cc9277e054fbf8f
de947a2b003983805e4062e00aaf174bd5b95f92
739430 F20101109_AABGWH hamey_r_Page_068.jp2
ff3c9493825c7c3f38ae84e9285ae084
4ab3dc9450b9144dc81e9f1feeb8a855a3108bf4
5209 F20101109_AABGVT hamey_r_Page_137thm.jpg
6c51d928044f988b54cba95b71e8f385
728d07f0bdd6ccca3ca99f3aa0f2716e95769021
F20101109_AABHCD hamey_r_Page_057.tif
70bc9143a7bf1909cff70c32490ea992
8ff019b8fdad454dac48b4826a2c567820e85fa5
F20101109_AABHBO hamey_r_Page_024.tif
ec6f2ca299bad50007b57a71128d419c
05e33aca09e808db7c8438589a41987283c44f5f
2113 F20101109_AABGWI hamey_r_Page_021.txt
bc85a5532294f85b51320d93c1b7c4e5
4aaf51baa2fe7fbc71286a9782f511a910d6b954
323 F20101109_AABGVU hamey_r_Page_115.txt
27251705a85f62cf25effe3dd5900598
0a942bfca95b409eae3dd2bf140ae704941be918
F20101109_AABHCE hamey_r_Page_063.tif
04973a21b55ff37e1f397c39d2d43a97
717fcdd8980278db5f62621362f39f5bb12ac17d
F20101109_AABHBP hamey_r_Page_025.tif
3ada6a2ef2cff73b416e164146b014ab
b82adce050368e6b462adcf1ae55b315cb5e798d
72019 F20101109_AABGWJ hamey_r_Page_103.jpg
e4b55ce02ef15901d6cc320e9f866dc5
7badcaf46cbeb278987dea097d19d579e1bc1602
26608 F20101109_AABGVV hamey_r_Page_047.QC.jpg
9818a92fc18a0271d17ba9eadb3f0fbf
614137a51dd8f35479a830b45cc5abc58b5ef9f1
F20101109_AABHCF hamey_r_Page_068.tif
ab03c8a3db2d9c627a0de336734100d1
abcae15b62c1b2f6a59270fc5a98cc9012cc2b9a
F20101109_AABHBQ hamey_r_Page_026.tif
5e4eb995663f6b4ac3c09ceb898a2e0a
e4d9c933ed0db0970f91742602e59851fc3ac7c7
55605 F20101109_AABGWK hamey_r_Page_054.pro
2a5af90b98435fe19efe434d2bef8266
88d2a6e8eef545ec13a2485163c040782b7113f7
7180 F20101109_AABGVW hamey_r_Page_001.QC.jpg
ff107aa6c08ed436c1b4e86a25982634
2f62be55431eb952ddfb13e973496dd2645d4262
F20101109_AABHCG hamey_r_Page_070.tif
97b583b8208bc366edd96a8d01856266
5cf38a3f1f2967230f5498f970499e76dc156b75
F20101109_AABHBR hamey_r_Page_027.tif
8c49b1dae7d742523d3376f0ac63f28d
6dd6ff405fb5b95eeeb2b79736e5303a7a706744
F20101109_AABGWL hamey_r_Page_071.tif
1fbcf53080f7e2b5e392551cb56a5fe2
c03f4edd78ca07a84741b7ac0b0e3a358a1418fc
57256 F20101109_AABGVX hamey_r_Page_141.jpg
d18d718e34d287d19cf94357e9d0e718
a73b45113859e9b7a70f4b2c0d4f6e3d1de66b2a
F20101109_AABHCH hamey_r_Page_072.tif
c240fab2016c16839b193a09bf4e2bda
08b005ed9fb56b475d162230a1ee0261072d1982
F20101109_AABHBS hamey_r_Page_031.tif
8fe9fa06edfd2777d0d05ce4c214573e
65ccc4c693fa328ed89de06ca64a67089b0dfcfc
259 F20101109_AABGXA hamey_r_Page_102.txt
342831aa0e7be994e41d418aa3c7d442
c83966dfee15fa77ea43912301e9c4f4d61f5b22
2243 F20101109_AABGWM hamey_r_Page_088.txt
26ae15ee899a603f29a3456498184679
b83be1eb37724dfb1387231a6b9196883c526f31
21315 F20101109_AABGVY hamey_r_Page_007.jpg
9e47b3dbe6c8c3198c8087fa50d19390
81b875af424981091a0412b90ef7b3c56baacd0f
F20101109_AABHCI hamey_r_Page_074.tif
c2475171aab1a4e5e9a75832c5383235
4e699b74dcb8ca5c2b7eaa82357e9c665e42ea79
F20101109_AABHBT hamey_r_Page_032.tif
973ea4581b09f144ba13700481eb08d2
a8cd316d9edf5aea95dad5bf1b6ab2e6f243e8cb
755 F20101109_AABGXB hamey_r_Page_007.txt
9c006fdcc116eabe1df69d7d026dbafe
ed3daa613b923cd9b6a9ab9e86081dfbac829ad6
17691 F20101109_AABGWN hamey_r_Page_085.QC.jpg
212038cab4483da2d5ad77611add2b98
78143c5e13a51881aeeff83b1c0ea3f371ceff53
1051954 F20101109_AABGVZ hamey_r_Page_055.jp2
e7dcb5a62565aec7f906841c98f2f892
0e1072df23d0c9eaf630e8fdc04721461e5aeaca
F20101109_AABHBU hamey_r_Page_040.tif
05e69a044602b6121c5af13778612226
498772c3d561f53299ad8357836c53ae9e5f68cb
176901 F20101109_AABGXC UFE0022125_00001.mets FULL
37ac67e7d56be18527c085e3739e6785
aa9eddeb312fb9ce02788a94c1980446c35f6f58
934630 F20101109_AABGWO hamey_r_Page_135.jp2
e2180e0ca0074bdbbdf9fb59a22cb7c8
c79cba2ad3b9bdd891e60251e378fd7daec886f5
F20101109_AABHCJ hamey_r_Page_075.tif
936de6ea9e5c8367b8015d7af45010c3
4750f8f72c0f06b391a6f03acb871d3054aeeab0
F20101109_AABHBV hamey_r_Page_041.tif
31f79c23df3e25874216fd0499b9451f
1601ee8b337f790019ddf80b49e3a205a4627123
F20101109_AABGWP hamey_r_Page_047.tif
e99e20049b4b068b29f91266cebb9ebe
edb8877eaf6b0ad30d4e1258cc5bfb72af12522b
F20101109_AABHCK hamey_r_Page_082.tif
499ca277c757053885fa169f1442d4db
c93a780c61d5cd4d008f577ee43299bc4aef2bd9
F20101109_AABHBW hamey_r_Page_042.tif
662e00d35ba7282b12f59ca0c2237dc2
9fb55f7e8902331fba6429076faae72a1001722e
28417 F20101109_AABGWQ hamey_r_Page_002.jp2
6ec5057f4b7f402600897cf2925e2b3b
c7787cab72ca26b0b6ec8c2f4093f72186393ee7
F20101109_AABHDA hamey_r_Page_113.tif
b9c1adda97db19b98a8a9f851c8ab99e
e058fe2c8b7be314a1699b033a3c3fe6df20b6f1
F20101109_AABHCL hamey_r_Page_086.tif
3a6a7252ccf24acff56b6434a35757f2
9b2528e99d6fb1d8b485ac740a04befef903ff08
F20101109_AABHBX hamey_r_Page_044.tif
1705552d9dab7e727f353155c7545be9
d24a56271024ccad9a8ac4403602df539153bc79
10271 F20101109_AABGXF hamey_r_Page_002.jpg
aadd8afddd4d7c5a9ee85ad86a5ead89
c3a64f2f3b3af45178870188389dee75142fa72e
7011 F20101109_AABGWR hamey_r_Page_046thm.jpg
b26dba2cc58a8525606685d5901c75f4
2ae2f938b851b20bc1731081c6f9c1c42866466e
F20101109_AABHDB hamey_r_Page_114.tif
3dbcd48e93e8462ff95383edc9ebd1da
daea6c1c86aa524d915e0f497db576bc6f6a76be
F20101109_AABHCM hamey_r_Page_087.tif
999142677ce5fab57997a5ba76df55f8
deeea3804d54ec8ec7228ada5f30bde48a7b606f
F20101109_AABHBY hamey_r_Page_046.tif
7d6e13b5afbab7782b60898c0de6acd0
1b823ace9b656b3cb7c1f10029fe0902c24d4205
11421 F20101109_AABGXG hamey_r_Page_003.jpg
1901b0ef13654c618b814ffefbae073c
06119658b486900543584146e6147829ae286841
1008682 F20101109_AABGWS hamey_r_Page_071.jp2
e6c088aa8c199efc7ce32d8084026a3c
92afcd036d75b853db7249c8e00e0415327b1c85
F20101109_AABHDC hamey_r_Page_119.tif
e6ab2fdd118b7495e767fcc0a6569ff8
93c5f422c4da59c0f8967b924027192b983e1fd9
F20101109_AABHCN hamey_r_Page_088.tif
8094043c835156b6de2692e4726d304e
81dc84ab29132b7d4601e99bf0a3ad1833c1b209
F20101109_AABHBZ hamey_r_Page_050.tif
283365411fa60779e126ae71d856ad96
7053bff81d227cffc1a998e16711f26835dccca3
25063 F20101109_AABGWT hamey_r_Page_041.QC.jpg
f0653f5be8a87498b09bd401e092bd7c
641b251fdf9ccb520becbb0e9d7409be7a61eaaf
27669 F20101109_AABGAA hamey_r_Page_056.QC.jpg
51a38e50414290589852c47aa81552a4
f958f6c7996af9bf273a0779ef48fcb3400ed789
32121 F20101109_AABGXH hamey_r_Page_009.jpg
c28d6899a692bd3f652141d09519bc86
4d44cd477792214cb271dffa22503c5878a1f8fb
F20101109_AABHDD hamey_r_Page_123.tif
a7a6b87462b9f7dbd3caacb6d8b1b760
8ec50676f8b3cd363a7a1fd0f7e9946cdd68a965
F20101109_AABHCO hamey_r_Page_089.tif
ceebaf6a5a4b0e7b53772fd171a5b390
20b30c1f73b612ce2e1ce4ca867f3419b7320a31
7312 F20101109_AABGWU hamey_r_Page_151thm.jpg
9c5ffb46720c87654b7da8132ab956ac
25f0858091d5607a21676fa477fe459078200083
19334 F20101109_AABGAB hamey_r_Page_126.jpg
8b638ff3067f2ea91b026b0b9265baf3
264da8f64e06d6eca78264c6cbfb2fd00649a6e6
84464 F20101109_AABGXI hamey_r_Page_010.jpg
be390c23958b1a6f0bf3451039aaf2ca
689bc7074626d5e2a634981bb6b4e21c890c86af
F20101109_AABHDE hamey_r_Page_125.tif
9d6a2f38f7da14c79d9504ce94b7d5fd
fc4e93c18d8cb145f0ba941101f0bd5bff46a116
F20101109_AABHCP hamey_r_Page_091.tif
d7e391dd2d86b64478ccd4eaba65d7b3
0c38cc91940545179598d52569e9ef3de135e126
29111 F20101109_AABGWV hamey_r_Page_121.jpg
1d09c8f3eba7aa309c637788d6c7b384
5fac57400501210ee0aa8583f730c290ae22cd63
3436 F20101109_AABGAC hamey_r_Page_059thm.jpg
bae629e3275a830c8bec01ff7126f936
05a9b479e0eaab006750682fbf7b25bde3468083
108263 F20101109_AABGXJ hamey_r_Page_011.jpg
0991e3391c1a81f0478668b91c73fddc
c2ff9ccddf6612551cea77f273f7dd0e64dd9400
F20101109_AABHDF hamey_r_Page_126.tif
347c0939a78c030b2ff5f2b75b5f5047
767f0cbd7200b51fd84afac933b42e63363664f1
F20101109_AABHCQ hamey_r_Page_092.tif
786edb54c0685dbf71835523268319b4
301867dc1f30104a6b447383b622a77b1b75bd6e
1051980 F20101109_AABGWW hamey_r_Page_058.jp2
495dab58b4cf4f5866c716455de48735
0a13c02de0441c7a67b67d591628d8443dda9332
12966 F20101109_AABGAD hamey_r_Page_122.QC.jpg
68fff4f7242a8d2bf738bdff190f83b0
db08bb16a8bb28bca84b398fd23277123ef1e062
101033 F20101109_AABGXK hamey_r_Page_012.jpg
1d8202e2464f9157cae67ac6caccc8e7
90531291a674feb4b0a5cf4bf20c7e2cad4462b7
F20101109_AABHDG hamey_r_Page_128.tif
708353c9f5fb95f694e16fc53aa34eb6
cc616d139dbedeb7a6aa8c46d84725d9da72a026
F20101109_AABHCR hamey_r_Page_096.tif
2ea9a9bdcb8720fab7d23c6f54ae5902
41f313c1ab1f9200c893cf5695eeb514e6b8b145
16492 F20101109_AABGWX hamey_r_Page_137.QC.jpg
def38c490f79cbc48b85938164d079e9
598063f146334c335fb2a57c481f8bc1f0086105
F20101109_AABGAE hamey_r_Page_117.tif
1d293ed4125cbb3fff6090e550d4cf66
293915b3df442fb3b72bc213d9d06872193e9685
80521 F20101109_AABGXL hamey_r_Page_014.jpg
ef2c8d5a4835e33e18413441172ec1b2
847459b76d6de123f25c94e6636550d6486180a5
F20101109_AABHDH hamey_r_Page_129.tif
ec69b17756cd347251c2be7c0d80c664
f774044b5bcd1a8b1a8fb5041ae6b1771a4129de
F20101109_AABHCS hamey_r_Page_097.tif
2cbe59bbc3c7a2e86d915f24f8085e92
985e1cae21ce21e93cabeaf83502a5700d1c385f
117691 F20101109_AABGWY hamey_r_Page_006.pro
c044c8a4d9a5e873aacacc9c8262a46c
fd1d6495258d25f13615516c592c8684069ebe09
F20101109_AABGAF hamey_r_Page_150.jp2
284b9f1f5462d8b148585c07e549bbc8
f7c59073d66fee4fde6ea4d49bfabcc26d931f93
86722 F20101109_AABGYA hamey_r_Page_056.jpg
6845dd36fcb456f32c9ccdf45375167f
590a4f7e4d34f2285e4fa207b9d40770af843b59
12974 F20101109_AABGXM hamey_r_Page_015.jpg
27c472d9f0875835c31947c1bd6aba3f
2f0b67b3752ea88b58edef34904f45f21e5120ae
F20101109_AABHDI hamey_r_Page_132.tif
c27a12a8692a3f93492d2dc95fc43f54
5048d8fd602bdf3ff3df6cb4c6e87d34da487c97
F20101109_AABHCT hamey_r_Page_098.tif
59493ac282d36f8855f75c02ec089ca8
cd284d43ba43ebdd8e28240141b44588c492e9dd
F20101109_AABGWZ hamey_r_Page_026.jp2
d27bcd30a341a4d5876b5428be640275
2c05ca8596a1756b0fa2c3226872091ca0074e25
F20101109_AABGAG hamey_r_Page_017thm.jpg
c412e0ddf1129cc458bcc85621feacb1
f18cd4306854cf1590e212b8b29c26bd2b821f3b
28430 F20101109_AABGYB hamey_r_Page_062.jpg
36ad19f0ce5cc936051f0cb1df758c6f
2efebe99c71d1c3990e894b5072fd61dfe39f178
62178 F20101109_AABGXN hamey_r_Page_016.jpg
dc8fd71d2e4041df20ec810d6f08983c
40a464a88214786736c884503f464675eb93f117
F20101109_AABHDJ hamey_r_Page_134.tif
c2a2953f516365fd6128309ccfd10cbb
f87c1a11a77222660b5bd27b2e247d8b9e048349
F20101109_AABHCU hamey_r_Page_100.tif
3e7ea4806510075a07d54028804e5170
5921b694904c336643870fe4fe2cd8980f4ceb90
5001 F20101109_AABGAH hamey_r_Page_117thm.jpg
85030195b6b9bcec7a05780b24d5be7d
2b1fe72ebddeb0dd713df24563fc3862ca770f92
45023 F20101109_AABGYC hamey_r_Page_065.jpg
a9743f1e1d7d5eada77d3ac0d42605bb
96ca12fed2ac70d1c9f7ee1ff6ab4459ecf96368
51102 F20101109_AABGXO hamey_r_Page_022.jpg
46a7b5e4a851f89a7b5b2bdc49d583b7
0a9ef0c947711a8f085c7bb5e62c6af49fef577a
F20101109_AABHCV hamey_r_Page_104.tif
5b56da60b871f79527c4f0ad1163214d
3a6e650b565aaf474a26cb5703df998442bc8a31
29586 F20101109_AABGYD hamey_r_Page_070.jpg
3ff47d006b75121916eb72b69dd480d0
9c27597eab327cf46f270591692a5ff7440bb214
84913 F20101109_AABGXP hamey_r_Page_029.jpg
4bc985cbf392ac72ffc9f765b6cb021d
a2d95ffdb55cf9d525be31189c9586ed1e8d502a
F20101109_AABHDK hamey_r_Page_135.tif
978442655489b46a39eedab1c38afa34
e8c97695336e50ea3be8aff0316ee460525d6cbd
F20101109_AABHCW hamey_r_Page_106.tif
899e9c87cf1aa5ee446499a2e625f95c
cf53b7b5c3eea352e498a4ae15c0f45b0b2bb79f
78751 F20101109_AABGAI hamey_r_Page_089.jpg
b6b1c2ef71cbf66376c6aa0924626739
afa1be9b8b5a18a2e33c3b6146c510c92d70e067
85681 F20101109_AABGYE hamey_r_Page_073.jpg
f9de7510a903b347d943470948efa173
bd44c85c081f14ab9334ebe2a45ce2b05c9d8f8d
87629 F20101109_AABGXQ hamey_r_Page_030.jpg
1046978d0eff4610f7967599b424aefe
934825161132f3fb069ea8174c2e002fef415c58
48595 F20101109_AABHEA hamey_r_Page_028.pro
d672530e9b13bd281b70e6ec930076dc
69e2bcbbf2a5206e2d622ca740d596449e493f91
F20101109_AABHDL hamey_r_Page_142.tif
4d723d4aa85a4e497d22c39e5ce777ef
6bb244a018f61dfc9ae8d25740118a854c2d11db
F20101109_AABHCX hamey_r_Page_110.tif
b169e0ad619dd9fa6879d70eadd37c2b
aad14ed24545d515680783cb670853cf6a7aa33f
2410 F20101109_AABGAJ hamey_r_Page_111thm.jpg
7e72be068bd029c5c934ec167db29e9d
14d0c062a27b7a2cc0e56d5bc388db044b7eba51
77415 F20101109_AABGYF hamey_r_Page_075.jpg
96c10b66893faf1ecc05abdcd64bc2b1
38e24746f166c472d57cebb89884531fd358c422
86286 F20101109_AABGXR hamey_r_Page_031.jpg
34162c08b2267cd821e9a91078ef7e80
8466803321b11904cfdb6de07efc5b0b3ad64469
48236 F20101109_AABHEB hamey_r_Page_032.pro
9e4d3f07c1a7a82cdb7265572bc88b98
1298b266a7ce637429f84cce64538687fa0f1f2e
F20101109_AABHDM hamey_r_Page_148.tif
384d088d0e2677ce2d9479e2799cdd66
f80879953a5d8e2aad2b87392511f5e2ce9a281d
F20101109_AABHCY hamey_r_Page_111.tif
07607e6e1d21a5d879d2c75c7abe8120
76d7ddff1ed657a66020f0ccded666f07304202a
F20101109_AABGAK hamey_r_Page_062.tif
515e408f4b84aa70c3de996fedf3a81e
7cc44441ab1383de6072b116465ddf30f2b6d6e7
77144 F20101109_AABGYG hamey_r_Page_077.jpg
3859c3ab34e52b1717b3f95ef8d3bf15
60b612fb2f40b3f53a3fd31bc4ac82cfae145fd2
77599 F20101109_AABGXS hamey_r_Page_032.jpg
2dab2dace301004946ee5b0d54fa887c
4031e2f1e81b9af38ce24ce2bc9190f985c3165c
11730 F20101109_AABHEC hamey_r_Page_036.pro
587d5226ee1a7aa43cfb3d1847121659
e296d1796f6e8f4dbe9bca839219e0550712bd06
F20101109_AABHDN hamey_r_Page_152.tif
c363d9bae1a57e7b20c7a717efd2d7a6
f42d5275f4b1ea146116d776515ab2105ca97d30
F20101109_AABHCZ hamey_r_Page_112.tif
7091fdbea48319b4a61dd704cff9a72c
7cd979b26e2b1b267ed4bb3a339ad678d27bc3cd
12901 F20101109_AABGAL hamey_r_Page_040.QC.jpg
76fdb3bf3e916a833ad23e9388a1ad3f
981c6afb9e60c8c1d04ae312d479d5be9c4dcc07
83806 F20101109_AABGYH hamey_r_Page_079.jpg
a2760932fe8675e753849b55682274e0
7159d6a9c4784915f4b68cb91c5a0ca59add786d
34894 F20101109_AABGXT hamey_r_Page_040.jpg
4359e32826ce4889f654eed71be22e51
95cfa692c4201be694d1b2b4934636c977ff062e
5365 F20101109_AABGBA hamey_r_Page_039.pro
ca17542b53b0300ccce216ae7d96d8fb
b7c7cb052f885db856689839d0e96ec21dabf721
6975 F20101109_AABHED hamey_r_Page_037.pro
087a72d41735fc04431c6b3bfcb43162
b5b122a1eddf27ba7fcfc18300ae335d4692c2b1
F20101109_AABHDO hamey_r_Page_153.tif
ddf4a1cf77a7e70fde940772de800f89
5519a53d9d9493506f34eb3a98d7925ba07d7da7
25591 F20101109_AABGAM hamey_r_Page_025.QC.jpg
4ff176735f7d7c36c39658e6d938850c
4e2e585df6a033ac87c6464cf218450701312353
86794 F20101109_AABGYI hamey_r_Page_080.jpg
8a7d0479637a1fe3261ac839ff612b50
322b5ac79ee7154d6600be564c6fdb6014c37bcd
73016 F20101109_AABGXU hamey_r_Page_042.jpg
53894c59786777f05a9629bbc2cdf512
c81db51dea6e783bce81089501b0efb6694bb4ea
81566 F20101109_AABGBB hamey_r_Page_144.jpg
952120a6967583ab9ea40abf71471d83
ea15d4c978bf99019e71e34f0e70e2f883031836
46428 F20101109_AABHEE hamey_r_Page_042.pro
a6edcedee010dee532ed124a596ab06a
f7baee2dc5663e632d9aca7db0e5f865e47450c6
F20101109_AABHDP hamey_r_Page_154.tif
38b6c68d0693805b4897ba71e1d4c4d6
5ea3208221c5e13633037a4d0227fc4ff96abe1f
F20101109_AABGAN hamey_r_Page_044.jp2
9c7c977b16b31e18ec43e11e2b232c72
6a410af543210a115d7cc5a75423ea40db6d97da
58949 F20101109_AABGYJ hamey_r_Page_083.jpg
47833bce25d8ccb52607b0d7d0d323e9
ac42de945d576a7f6513bea37d9752570bbf1764
84561 F20101109_AABGXV hamey_r_Page_043.jpg
be2179b4430e5c0407bcea73862ec563
42ece92c8be58fdaa81ce387c75f50e276ee18ce
612 F20101109_AABGBC hamey_r_Page_153.txt
4fc5202efd54fb88125960d533f7909b
1ba5852b397362020befb0b5a2e3230100ce6895
53513 F20101109_AABHEF hamey_r_Page_043.pro
bd2e00bcfa6044eac65d21c9e0f8b79f
066dbf0130a3ff8147ac1db67bfe8726c5f5aa5a
7476 F20101109_AABHDQ hamey_r_Page_001.pro
da9924aa2bb9f0b339413e497f4e147f
fb916d90cf4fd0ac87c4f437417ca5d4b8f276b4
3076 F20101109_AABGAO hamey_r_Page_102thm.jpg
22bd22607c5df83e7daa322b209a962f
ce092648595fa46c5bb8dece652e6037c9736066
81946 F20101109_AABGYK hamey_r_Page_084.jpg
644cffc40f05ed3d8f2e79a5e013bb1a
1a6e9688b3eceaec8cfa6f6cb0b37861fcf3930e
85127 F20101109_AABGXW hamey_r_Page_045.jpg
176dcaa8f97140fe0f9cb1adaa2840e8
1aae706d4983d91b7b685304da9028eb15edda75
24830 F20101109_AABGBD hamey_r_Page_046.QC.jpg
1b99a392e29d5d08ee0de7cfeeb65905
ddc2013b714b66854cebeb54d15915ecbe164ab6
56947 F20101109_AABHEG hamey_r_Page_044.pro
5340ceb7cd50736b12dec682b35509cd
67a4e07d03559244ee3b32c89f93be6de6e5148a
937 F20101109_AABHDR hamey_r_Page_002.pro
d631c66009411178794ac65935561842
4ec93f3b948fbc3962a03f1feb7bfffc6d6dec62
F20101109_AABGAP hamey_r_Page_055.tif
2388b9a9361249755526c67d54820187
eac4b4f43e9e309a90a1aa027a4af796a5b24370
38706 F20101109_AABGZA hamey_r_Page_123.jpg
56329fc19c916d556d73e21254317f5e
2eeb60cd5970de6d8f005e81f333b5be48570a74
53484 F20101109_AABGYL hamey_r_Page_085.jpg
c5b528e4d9b6bcc2e01b5bdb5f5862a9
412de6cff08dc9ad2e96cb633cc22d90e1dcc43e
80468 F20101109_AABGXX hamey_r_Page_048.jpg
40bf347e87bf1b0f6e24d7f561ba8e48
4d37eb3047de23f35d5b2e64eb6233b0b10229d8
25548 F20101109_AABGBE hamey_r_Page_147.QC.jpg
8417919e66fd61647702c50d14faf014
beae072d54c85c33195b6263e587801b33fbdf85
54333 F20101109_AABHEH hamey_r_Page_045.pro
e4ddf590594be95a8f5bfe62e15883fb
14dad971469858d04f495463bac469c1cac51ed7
1741 F20101109_AABHDS hamey_r_Page_003.pro
5bd301a3aff74136fe164e5513937592
cc39c856313604c96cba157e33b5483d1c0a84bf
6787 F20101109_AABGAQ hamey_r_Page_077thm.jpg
6fd59f4402526b648c78ceda1b8ad50c
0c0cc77d66dac06b0b772be92f77fa19b4b6377a
86515 F20101109_AABGYM hamey_r_Page_090.jpg
8f4c8a8b591fab387da92ebc1ec6feb4
0f00950705bc290c8a8585a6bbe248ab6c8b3e0d
86347 F20101109_AABGXY hamey_r_Page_050.jpg
8bfeb5212cb29a8858de3c4864047620
0175346019325b456f7a366a0f2ab56bff131b6c
27868 F20101109_AABGBF hamey_r_Page_092.QC.jpg
391c36810992ccfcd6c5f144abde8a66
b9885290703f79f6f0bd618991e91921c0e84285
53440 F20101109_AABHEI hamey_r_Page_047.pro
0d6b75346b2a3b750eac7f84778fe7e9
c13d7223aec3a8ba91ae852ea832553973f0bfe8
21818 F20101109_AABHDT hamey_r_Page_009.pro
7c8e64a0e19b5f83c5c9290ce19d6ebe
74feccf5f689b451a748fa89cee52d5baa47d3dd
26914 F20101109_AABGAR hamey_r_Page_090.QC.jpg
bef019151c8c90f26b1bc058140e8252
b9f05d83d39e102d798028a668e4c42b3b820a42
18585 F20101109_AABGZB hamey_r_Page_124.jpg
ba7fd66f678641f92b0f900ecb68b3b1
d06533845a072abd45793e2291dbf5caac568d05
88118 F20101109_AABGYN hamey_r_Page_093.jpg
78a75aebd73c0942fc331a6f8c5f6a34
a81acf52e0b7cd7b9f7cf587241f1b1202a08d0f
79045 F20101109_AABGXZ hamey_r_Page_052.jpg
5ea6497bff5499d56c91aeebe1d4c11f
239825d0c8b05f7ea09916cd6cbbebb9d5939bfa
6939 F20101109_AABGBG hamey_r_Page_032thm.jpg
cd27c86e2b3fbc289d6e4f752fa57251
790fade2e4703e02545dc6961f1ac922e00e4d9c
49085 F20101109_AABHEJ hamey_r_Page_048.pro
57fccd09961dba7201465752dfd39574
756bd7e98bba71aeb2f096366357b30349bb2ec0
81141 F20101109_AABHDU hamey_r_Page_011.pro
171117c05dfbd8a401ec97846c4ba98f
57beec9914df4a0c4b1d4b51262ca49a9435cb6b
7564 F20101109_AABFWA hamey_r_Page_093thm.jpg
9ca2ef4680ffcba6cb2bd43f4e805ab7
ee9c3a305897b39401ec4e22791e10a00eb17e38
27894 F20101109_AABGAS hamey_r_Page_104.pro
7eda5a8eda7251298423a162ba9ad8e1
60ffc90f06557bab884452ce91f920a26b2abcb1
88137 F20101109_AABGZC hamey_r_Page_131.jpg
1f099b52331a5918d7640d043f614de9
302ec7909042806088dc72c2268a7625d6d66cd9
36425 F20101109_AABGYO hamey_r_Page_095.jpg
ebe551121897bebc89091dcca2df5224
7f496b396d6b4a1d4d5ae3733c4e91604b3720c1
393579 F20101109_AABGBH hamey_r_Page_139.jp2
78784764ebea4353bc9da96510a577d8
6b8b151edddedd2a5f1c5eb1c487d0f217e8de47
49456 F20101109_AABHEK hamey_r_Page_051.pro
2b5b465b54b5da341e4375f92a9fc4c6
c73f05bdd6f11b885be9811238c87009a4e55cb0
25989 F20101109_AABHDV hamey_r_Page_013.pro
1217c37b2654857ba0e5f206aeb0c4df
39e5001ec82b76b5e0b64c837b3e212b54316833
97046 F20101109_AABFWB hamey_r_Page_006.jpg
cf44ae8169ffba6a20967d54750d687d
54db0c1e1c165be873b650c1e112aea48e842ca8
2237 F20101109_AABGAT hamey_r_Page_044.txt
ea6ec3c67f0d4f7b806badea1f55ea96
8bbe26522416f452e1692e33f5f19029453a6943
85245 F20101109_AABGZD hamey_r_Page_133.jpg
98855bfe5563ff1e390b1a97cef793c6
6f9815319d35ae4a1efa5b58f5afc295807d09d8
28744 F20101109_AABGYP hamey_r_Page_097.jpg
ac48ea21cd10f4de68a4ef3c0ce9e210
df015dee56587fa53197c6f8f70b593f932fe3b0
2009 F20101109_AABGBI hamey_r_Page_089.txt
b1dc7b5857378b9a6ca26f78e5241c61
8c5bb4439401e556bc8ea682ac7d51b4e86aee0f
3446 F20101109_AABHDW hamey_r_Page_015.pro
62d09911c17325ce777e9ba9f1d15049
c247ec06240c95eaa0b73bd5ed0230f2f85850c5
45317 F20101109_AABFWC hamey_r_Page_019.jpg
80802a1a9308928de85a2338c51a9e23
ab74d6f6496f85fd24afd80343f225d63f46337b
F20101109_AABGAU hamey_r_Page_125.jp2
f3508717e460866247f1f0abee9cff98
fc40da9a9126a0620d586ccf03a008cd84c6d24a
79636 F20101109_AABGZE hamey_r_Page_134.jpg
f1f2ee2ca5327fb9b6d5fbf9a006a962
8fab7ae12bf80139c1d41f9a0b371f83aa4cad6b
32849 F20101109_AABGYQ hamey_r_Page_098.jpg
fde89dbf2d02835d9437bc55377233e3
75872053a5b7e6f9cdc66f92bf842ec7789b6a97
57194 F20101109_AABHFA hamey_r_Page_092.pro
70f43ce3338ed8dfbe78c8c76c50c45e
1f45061867573bffcdbc0b079d99e787345f67d1
52097 F20101109_AABHEL hamey_r_Page_053.pro
2c3373c357f55b8e9a766d9a834d7bee
d54e75b18408f5c96e29350838c2dbd7360a9f7a
25526 F20101109_AABHDX hamey_r_Page_019.pro
4d60637564eaacb0af83f3ad43c5cd76
f86f9af3a1a718bba8e2dd678a1408498034e38e
1958 F20101109_AABFWD hamey_r_Page_134.txt
85e3ab74b19df8f8d8cb63b308996609
c7e9beef80fb28290881cc6cdb8283ce9df5cc85
F20101109_AABGAV hamey_r_Page_138.tif
b30be1786c3ca02c3c85061676addaa9
7c125f8a6ff1a0aba0b3c3a0f8f918cd639c7383
68158 F20101109_AABGZF hamey_r_Page_135.jpg
fde35033d113c23cd2ac6c4ef18f013e
75b4de15383f0849c8d2bf652dea2399ec16bfa3
34129 F20101109_AABGYR hamey_r_Page_099.jpg
9c6edac89666018a5a3cb77d40568162
132feed06dee1f862fa63dd960e049a1fec8f07b
24431 F20101109_AABGBJ hamey_r_Page_128.QC.jpg
a58075ef605156c67d9e65b9b0c8dd9a
0ac9a2d742b9463a332875e5c7924451ef723708
56211 F20101109_AABHFB hamey_r_Page_093.pro
de49baaf33113832cb688ce7b5ec6f50
94674f194c4449a96622c42ae47a2b7cec8020a4
52109 F20101109_AABHEM hamey_r_Page_055.pro
fb7c8ddc881d1a216d4d7d36855b7b00
5356318b13ec75bc6f2164efa11183dd442a9d50
2677 F20101109_AABHDY hamey_r_Page_020.pro
c5a51c2b8a6c26e961ce507339eadc42
49e571df4f7a174d02e33eebd81e1ab7448b8b18
F20101109_AABFWE hamey_r_Page_069.tif
fe7f331df3d6194d6dba7c653adece05
487a16bfbd0f5e0f74f34213412b455caa3f0fbb
4843 F20101109_AABGAW hamey_r_Page_136thm.jpg
329f6603ce604329ce913036c9a15de1
83c3dcf4761f3654465202a264c91213b3594476
49096 F20101109_AABGZG hamey_r_Page_137.jpg
cb4d8438416881136dee5d463b0cd9b7
cd5c0ff78f32410c519f1c402fa4a056f68b1e2e
34023 F20101109_AABGYS hamey_r_Page_100.jpg
63fe6b7c09797d7124b45ef1cd2aeebb
b39b2503f432cc89d60f032c090280ccfbe6d5c9
4150 F20101109_AABGBK hamey_r_Page_019thm.jpg
8fbf304e0cab56ba17268b6faf3b97e6
427209e65393ab969af26863f24bbd4e6ef03a9f
22895 F20101109_AABHFC hamey_r_Page_094.pro
7d76df31cdd597906e77b2c9bf8b1143
24232a82267700aeda93cb565ca37066199899bb
56420 F20101109_AABHEN hamey_r_Page_056.pro
bd2e08199e93e5c223f233af5411906e
83297b07b22e7fa1dca8ac5c05e8cfcdeb5d71dc
51833 F20101109_AABHDZ hamey_r_Page_021.pro
9d898682897ded6c9036ead78434aeb4
a8d64d92d9136ce418be46321bed4e237efc128d
2067 F20101109_AABFWF hamey_r_Page_143.txt
4e096f7e579d40f36e41f5695207523a
a63b4d71649bfcc38d94736f4f8670d635065d3f
316 F20101109_AABGAX hamey_r_Page_069.txt
ae190b967a35a949a1099ac9fe85f214
2f8606c74b82e6c9c32356cbfb0e38d029703748
29980 F20101109_AABGZH hamey_r_Page_142.jpg
07afecaf69683154eba664b4642e6585
d343ad9c9b768b2c3d1d47d219e56605a293ea1a
74784 F20101109_AABGYT hamey_r_Page_101.jpg
cdad322f3e08c020aaac08d52c2b88b9
c1b4b037475e94619298715aba514acdb0b468db
24265 F20101109_AABGCA hamey_r_Page_028.QC.jpg
6a7d09f8ed575476331bc28e4acaa269
30310e1d771649463599f98e2cf4db85a315c209
1051915 F20101109_AABGBL hamey_r_Page_027.jp2
f41eeb2191d934d4d5179ea9106257d5
65d81dc913cea3e2756f389c5783447c285a7d6b
11908 F20101109_AABHFD hamey_r_Page_096.pro
bcc7b607720f3a6e12f9972218415e26
ea11a178d62042afd6a5172587b66f930c1a2557
6436 F20101109_AABHEO hamey_r_Page_059.pro
b8c92b913656620d64ad634383951b92
6711e1f9c1ea710f9a2673702b05f517c10c886d
1796 F20101109_AABFWG hamey_r_Page_083.txt
90aa819f6ba35be490bc40c7e3e46fca
ed134f4b2b7d50a7babd094a541340e4e010ba34
51522 F20101109_AABGAY hamey_r_Page_024.pro
3b6219aa501bac8a3a0c75aecf2313ba
ce57d2fcf82115ea91e38d67b1d226c83162e9f0
80819 F20101109_AABGZI hamey_r_Page_143.jpg
6efb2ebeabf69a600bfb9c5e74f5a4e3
f52dcfb3a4c985dc2984076bf1058d7db82fa4b8
34177 F20101109_AABGYU hamey_r_Page_105.jpg
aed8b641deb73ba4a63ffe1ab3c2be4e
f6d87e880312bdc5f9981e0ab53dddbdfffd2c41
F20101109_AABGCB hamey_r_Page_144.tif
5378593d5a27a2e19ef142dc3fd03476
c906341b87e7999b3887d9086bfdcd42b90d33bd
1860 F20101109_AABGBM hamey_r_Page_033.txt
ade8c3e74e626c9ae6af7916ccd339d3
83e2bb94d949b4fad182583c7740d8406e128fae
44854 F20101109_AABHFE hamey_r_Page_101.pro
d477416ebe48fe8cd39970618bb5082b
31a3274e516a25ba58e58b1ac883692f2296c065
2926 F20101109_AABHEP hamey_r_Page_062.pro
a31030e685d60da3e39af7206c83217a
6c7a42dc0e67f5ebbc3a870fe0aa0524e86bdff4
804 F20101109_AABFWH hamey_r_Page_099.txt
14df8d4884c0c74f435574de7cb38ce3
144dd8b6b87d6329aa0cdd2bf0455bc81d9cd8ff
4168 F20101109_AABGAZ hamey_r_Page_036thm.jpg
b569dfa1c3385691aedd9f1ea2906ab5
0b2851d0753cf4881a1c433bf8552f56d434b1a6
88349 F20101109_AABGZJ hamey_r_Page_146.jpg
3d44d108ec5dffc3f90561012b934c4b
3e8502b25f6f57caccad75c2e07e4d62abd68fd9
F20101109_AABGYV hamey_r_Page_108.jpg
a8ebd74b73c95415eeabd2f5ee186dab
c82df8221c96283edfcef5fc8c29181fe4d576f3
28520 F20101109_AABGCC hamey_r_Page_146.QC.jpg
8cd75712cac8c79e9960e349f9b398c8
4f39d94f9077e2d3d2fcc6ddaf0bc981b51272e9
7571 F20101109_AABGBN hamey_r_Page_090thm.jpg
071f2bc3b2731dfa5d364d152605f1cb
87a824cf7c56720b0841c91f62789cf58c72cc52
4557 F20101109_AABHFF hamey_r_Page_102.pro
4af2b484f33a87aa7c68e3f7018efbe3
f892e48bd9fb911aa2ed98370fe4fce9f53b55c4
6807 F20101109_AABHEQ hamey_r_Page_063.pro
bff92de1dd26cb128b3cde76f2f434d7
dc7857fa8d5424a8b042410d174005d9e89e8a34
91344 F20101109_AABFWI hamey_r_Page_088.jpg
86476b53e5aae61f69a113ed77618d5c
2c9326574b6e5bfa200f95a868f84077caa1ea8c
80883 F20101109_AABGZK hamey_r_Page_150.jpg
d14878395430cfefb6e84474c78941d2
065706f020da3d415d07ef6d0ccf360b2865460e
38078 F20101109_AABGYW hamey_r_Page_114.jpg
c8a58df2fa5fe299a6044a34aa64181a
a63ad3926246a98b214c68de69a97846b94ab114
39967 F20101109_AABGCD hamey_r_Page_013.jpg
b3c4cdb478f20c92cde0cdf2d66bc1af
73596bd9f3b39439deb166ac8a3fbfa83649d149
77178 F20101109_AABGBO hamey_r_Page_127.jpg
96c2786889befe594a7b492dfcc088b6
70abcc62565377f489d41035aed2df54e7159844
8711 F20101109_AABHFG hamey_r_Page_105.pro
96b3785df0a438debc2ec8608d3dabab
b12c0cdea19fff9b0718877d745f0e9094ac902d
4226 F20101109_AABHER hamey_r_Page_071.pro
a87697136d0401f57dc7c9239a55d393
32984d8d3f852c309c8d5d99f0629d2f9dbfe8c1
48188 F20101109_AABFWJ hamey_r_Page_041.pro
61b6b79429c3d303452aeb953fa37779
3c5b24ec017bb806e2fa329838b6c361f502dea6
225527 F20101109_AABGZL hamey_r_Page_001.jp2
ea1025958224d7abffbef58823bf04e8
e78ae0c742e08f220d4801cac7f3fab71a39cd16
45625 F20101109_AABGYX hamey_r_Page_116.jpg
8e5655d17a2179121a2e0e2e97f2f385
4e704f7ac58113ce7dfbbd9de8800363e60b8932
8230 F20101109_AABGCE hamey_r_Page_119.pro
58bdb387078c0707ee48d9a5400f39fb
5d08a50af931672eaaa1e3bb21b5d0705f98cb6f
9619 F20101109_AABGBP hamey_r_Page_039.QC.jpg
5f09ffe2dd4b7bf2167bb69b1b80fa9c
47dfc9b9cc41af6d64b21eb528e003ed2bf88fd7
11932 F20101109_AABHFH hamey_r_Page_106.pro
ec4f0680280f9be837af144c51ec4dce
1b8e35c5ccf2a7033c22e4666277a2ed89b12fb5
56495 F20101109_AABHES hamey_r_Page_074.pro
5eb79d32b40a34c50cc06efcf0c277b4
356c794fbfc03a4e065934d00046e060f808abad
27235 F20101109_AABFWK hamey_r_Page_039.jpg
68b70444cdc45c699314a3018341dbd8
ed172d52fe692cba2e33143fda3fc7ec95f70432
F20101109_AABGZM hamey_r_Page_006.jp2
9e65b4a1a56ec3152a7a7f5fc11d954b
03981d20d4d76d9d639f0c8abef629623f426c2f
35622 F20101109_AABGYY hamey_r_Page_118.jpg
bc6694833d138416c31b1ddaf6996f7c
5acd708b9d41a3144a0d54adf01641842f334697
2141 F20101109_AABGCF hamey_r_Page_058.txt
dab97afdcc764b07fc8ad464d6342b9e
58812f18e9da693527c752c77d3da9629bf2a979
29429 F20101109_AABGBQ hamey_r_Page_115.jpg
b13010d1640227a6013922475a35a892
77fc251bdadd5120822fff730c2b4e1754a5ded3
17978 F20101109_AABHFI hamey_r_Page_108.pro
13d53462396b14847271431bfb38ec3b
e0cf835af3afda22e2719cdadcd452ee3e628d80
48953 F20101109_AABHET hamey_r_Page_075.pro
9751d0b770b8bc6d1ddc7a755bf44136
f0d60d5e5820dc3621d1bfb656e3bce722a6c208
27129 F20101109_AABFWL hamey_r_Page_049.QC.jpg
bdbb58e42d0d2f0bbee0145efc3a32f4
8e55de53167f7a2636354ef50e4e84ee55d155d6
390785 F20101109_AABGZN hamey_r_Page_007.jp2
6524a25c4ea6f86d6a6753a9cbdddc77
c51f19a4075e40c85f3dd67d8e9c963ab3248238
26883 F20101109_AABGYZ hamey_r_Page_119.jpg
c9bd6657a7e7c71277af44ec2aa11c3f
b7f758e06aa2b532fb96d24477893f1c10773069
6058 F20101109_AABGCG hamey_r_Page_006thm.jpg
93ad466247ab00bc19a223aaad7e28f8
26e06c3ef8d6045a60a259cebca6196123390859
915 F20101109_AABGBR hamey_r_Page_154.txt
fdd9cb9a4b1727050fa31bf8063052d1
4bd98b2cc9254f1030ea3c7f114fd48dd12de6a5
11441 F20101109_AABHFJ hamey_r_Page_111.pro
0c0405058430ea4d008b7c9a3153162b
e49c929dffa51e2baf0d2e537df612819f4a4703
38853 F20101109_AABHEU hamey_r_Page_083.pro
2332519d439943de3eeff64111ef579e
1a9ce4213a13ccafe2fc4f8ca32a1e7a77807668
10199 F20101109_AABFWM hamey_r_Page_121.pro
a1c168afcd247a2b93bac807bac6baaa
7073bf0d6c2b64efcd063afdf595976efbd3067a
1051973 F20101109_AABGZO hamey_r_Page_011.jp2
d7faef49aa8df51cbbdf176b80208e18
86678ca194a0de9bf8eed60e1a31b78dd4d2ca10
55089 F20101109_AABGCH hamey_r_Page_081.pro
3b55acea40953a7e709ed37b3a9c0a3c
4aec0a7162723e186b6b8a4249738e15870bfd8b
9836 F20101109_AABFXA hamey_r_Page_121.QC.jpg
fc166bc0bb9d08aede23ee8f6b054938
64b677e3b8dc26735951e3ab7e3b91c4969ecb6f
F20101109_AABGBS hamey_r_Page_127.jp2
3b5bb5a74ae6749f33a7491c60f692e4
494f5d51f8ace55c7ac771212f9fe4045b2afef9
9828 F20101109_AABHFK hamey_r_Page_112.pro
cd47b10e3a463995714c052fb244370f
c5bbcfad7fe6d84300be13462888cdf7bac335ae
54008 F20101109_AABHEV hamey_r_Page_084.pro
828aeb4c984aefeecef717a39d5ccd78
69641dd97fa2d3c03102b324ef7a57f74b5b3cd5
3711 F20101109_AABFWN hamey_r_Page_038.pro
d5a8490515f38169cfc723df2efb7cb9
151c8b8ba8801b56e7060bed4f8e0e9e8f3e1254
1044072 F20101109_AABGZP hamey_r_Page_013.jp2
95a69ce939a469532b9c1fa0fb287376
99861b055ec3c3ee8094eb8a8799b5bfddff07e2
761399 F20101109_AABGCI hamey_r_Page_122.jp2
342e1efadb510e6d462bcbb70a9ade0e
3f00d3c723e63efbdb3331368201da3d022f9e94
F20101109_AABFXB hamey_r_Page_013.tif
910395c7a2326c94303c88e34e282bc8
e569942abb2169ea6a1890ee8a60cc8b2b4c76fa
48498 F20101109_AABGBT hamey_r_Page_035.pro
a6485b5b8692b69658b233513295d9ad
17df75a344c336a654044c80273ab513516fc38c
30423 F20101109_AABHFL hamey_r_Page_113.pro
7a2a3b07449d0cd298af923e35673275
95acb770efaefc5e8b44b6d44667b9778ec19374
36619 F20101109_AABHEW hamey_r_Page_085.pro
e12b3cfe519a43dfff16f8ac16a35866
8fdab208d9194a5886cb0921e724e545e74e6a64
F20101109_AABFWO hamey_r_Page_141.pro
87d28a78e40cdccae3ea29c0a00029ec
35bedfd80b7bfe5a575c5a2c8445c6d2a6e6fee7
859932 F20101109_AABGZQ hamey_r_Page_016.jp2
f89d89e2c2c21dc0c51e28a9c4f5cb6f
56ad318f1be06384fa9f7012456f32cd598bddcd
F20101109_AABGCJ hamey_r_Page_045.tif
efd347620b74661232e34c85b59179c6
2749a0f841e531144d5ac21ab6585601091f2f1a
61269 F20101109_AABFXC hamey_r_Page_117.jpg
fdb246491f0a7c0b3bfc63ec09a54d0e
9facdac4811980caa2b0357587ad67cbcb958daf
F20101109_AABGBU hamey_r_Page_045.jp2
cd8722a472572f15290bf608c58607d5
f542d9a6ff58bca6e6afff8293f1a35f80718f83
4056 F20101109_AABHGA hamey_r_Page_005.txt
910ca9f1d8b7a317cb0688aed8a5a86f
22c4c7d5170b9810f32262372e6a11ce2f529d51
57155 F20101109_AABHEX hamey_r_Page_088.pro
61a42cb359f27894574f0e807d0d8132
d0d0683125a102d7275abbe73bc1be93a26382f2
6818 F20101109_AABFWP hamey_r_Page_150thm.jpg
6dee38153a78c34ff10f4be1d10bd63c
be8d6886712e1b445b4871f113d88929d80ac8b4
F20101109_AABGZR hamey_r_Page_018.jp2
9945eccb737e47f89d26ca9056f04ffa
a6c8617bba8eca6415e0c2a68ce46a5d708613b5
17337 F20101109_AABFXD hamey_r_Page_139.pro
64b8de56afbf497c02556a04c45b3bc2
6b3583b3e62de19085fc3587c48cd6bce8022d11
906 F20101109_AABGBV hamey_r_Page_094.txt
390cce18c6d5e11f9c7f9464fb1c8c59
943a2ead36b08d69f7d53bff0485155b7e6b74b2
876 F20101109_AABHGB hamey_r_Page_009.txt
7f183e41d47a8b5662ea5c46fc523765
968071abbf2783446e8a8ad57575303cc9306aae
9196 F20101109_AABHFM hamey_r_Page_117.pro
5f25e7268dbf06198b5d063d8b11082f
afb3174e3709e90fe19bcdd51c67387f8a9a077d
50183 F20101109_AABHEY hamey_r_Page_089.pro
53eaa19654d1c76270f27b2b890653c6
2b5fdfd0cb321538b522ac3bd149db6dd5043eac
14312 F20101109_AABFWQ hamey_r_Page_069.QC.jpg
b8a3f059a04ed95e3abc870764cd26ba
29a37a44af6f05d454f755574b182d2583f0f258
1051964 F20101109_AABGZS hamey_r_Page_029.jp2
9248677c773ed2a055201ffc32bca416
01fdd2be88d0ac7864dda8606ee45ce1bc37e49c
F20101109_AABGCK hamey_r_Page_043.tif
c9bd541e1a7bc088210a7a4552719998
1eeec143cb56dd08d1d5b60e55791f5742c1be15
6952 F20101109_AABFXE hamey_r_Page_021thm.jpg
76b3c4a1c4603a37428ebc3543bf3252
370c1842de36aa6a443bfff1f6f17a3f1e885301
1993 F20101109_AABGBW hamey_r_Page_041.txt
0a5e99e2f3230f0788c75f6beb7cd02f
862a6fb3655f549e07cbebebb73995e2309e2b56
3250 F20101109_AABHGC hamey_r_Page_011.txt
7edf293d2fb2e369bc717522c0b0107e
5201737b9df57a4affe6490c4a17cfc021ea120f
28568 F20101109_AABHFN hamey_r_Page_118.pro
4461df89fb1c87201ac59a9f95a74594
24c0943307c690bfbcacef732041750c906e4de2
53894 F20101109_AABHEZ hamey_r_Page_091.pro
80938ff3ff3dca2104636a5d2f7ecc73
64ce83618583eece4cc0a9a39d7a9895754f4187
F20101109_AABFWR hamey_r_Page_115.tif
6d3f0f30c60025351295a3c41b9eda13
ad5d0f5240edd7b3229c46972be135645bfabed6
1051978 F20101109_AABGZT hamey_r_Page_032.jp2
83881934a3317d1ba504be343b4178f6
713032900a28bdc235e435a3117d37234e84c5b4
796376 F20101109_AABGDA hamey_r_Page_136.jp2
510ff132b9c81606166914d60fd29cf0
abd120803396fe0e3e02570b4c4d51055d982235
334 F20101109_AABGCL hamey_r_Page_063.txt
94f2504d315f3a918d6cab49dee6f686
fb8fb3f32e835517200465d48911b054579a64a2
62300 F20101109_AABFXF hamey_r_Page_063.jpg
fe059a06b9fb4e433dd3437dddf4743a
8808f15a18178cd62b52c9379c53b56d0aba5b5a
11360 F20101109_AABGBX hamey_r_Page_013.QC.jpg
d09d4cda913f6cf9cdbb69bf57d572a7
377bd03ad5d680c2db8990598b8b3ee33f855baa
141 F20101109_AABHGD hamey_r_Page_015.txt
f3f1e70d6f0c8bfc9926cff87875007d
676cf6105145a0937b58c2b88a65e02fd8498437
2259 F20101109_AABHFO hamey_r_Page_124.pro
d7267b58b8cccc237e1093da2f8690c6
1657cb4e3aa08d27692a71e88b40a9bc1f7e0609
14715 F20101109_AABFWS hamey_r_Page_153.pro
6850ca64ec3da2b8a78b41fed108a913
14b329b33df82d88c35a5a7a482930f01a4b2651
F20101109_AABGZU hamey_r_Page_034.jp2
32f49cf1c45c603842d3fac85b3f04d8
061039288c31a297f6707fdb95e71687e7111493
5944 F20101109_AABGDB hamey_r_Page_016thm.jpg
104f94193ecda30e67ed81f3b87a494c
3093b68e86675f38991ca70d356ffdd4e9f52fee
2417 F20101109_AABGCM hamey_r_Page_064.pro
a34ec1de05be5e4c17bb83439d87acc1
fc193c458ff2eb9924fe4932a3043cc37328a883
1051927 F20101109_AABFXG hamey_r_Page_063.jp2
56afdae57ca3ea1e0b451e798b6bcfea
5b2b05fd5df2c1269b301ba148fa87d978975c45
42310 F20101109_AABGBY hamey_r_Page_061.jpg
2d382d1a68a6e210c7478b21abb92ff5
0d167bd495ce8e9fa029f3da26dbbee006aa18f7
1664 F20101109_AABHGE hamey_r_Page_016.txt
9b4616b9772d55425148414480251093
bd1ea7aa8e855796553f8cff9e062e8e95ce423e
3208 F20101109_AABHFP hamey_r_Page_125.pro
972dd4e90c731fa173cc3f522cda04b0
53cfeb8ff15c3a767a821d51ddffc2ecf544ba7d
1936 F20101109_AABFWT hamey_r_Page_048.txt
c08975e974e32597e9ed52ae2b66939c
66651c7e531859b6ff2b542452f88675ca901f3a
386727 F20101109_AABGZV hamey_r_Page_039.jp2
bda9b9336b66358b8a9ce6edc2e8cabc
4caed1089dd20ca4f1e0ed6cc123e6502ba21baa
81029 F20101109_AABGDC hamey_r_Page_053.jpg
731ee28cdcdc546f607fdb3445d4d641
a65c0808da1c2fe1779c45af72c86327715c6431
2122 F20101109_AABGCN hamey_r_Page_018.txt
72d023b8572c5837f39716037ec95571
4bd00a6b05bd45d0f97dfcf0ffc0a26aa2235656
30165 F20101109_AABFXH hamey_r_Page_112.jpg
49e089ae59f3cc8d3f35580b698a86e6
5c8c2ff8d85a3924eed577c62ebdcbabbfff53c4
F20101109_AABGBZ hamey_r_Page_064.tif
2a97d2bcc16cdecfd150bc92b8f5a0f1
7598d434c97f30a711c53fd2c78188952390ab0d
1913 F20101109_AABHGF hamey_r_Page_023.txt
2f6824f8fe3d17a6160b4ffef8d73ee5
8330dcba32b5dac3eed55c10dc8ee85241344334
2909 F20101109_AABHFQ hamey_r_Page_126.pro
61ca7f006b6a5d688c14a2b3ed93803d
8e95b18128f70b1c233fe896a7abe81109580cf6
416 F20101109_AABFWU hamey_r_Page_136.txt
865b25cbf0655b32542236573683995b
02f936d32c39637b6cef408f8f1cae451ec55123
F20101109_AABGZW hamey_r_Page_041.jp2
44ed68bc72f0277ba73d7980d5010d1b
3902636ac0fbdcd52968a3aba2df24774546e524
50652 F20101109_AABGDD hamey_r_Page_025.pro
ae6004f86c156e6267c3d79ce5a8b423
4f8820d110754d68ad68db327f2fb3366aea304d
1054 F20101109_AABGCO hamey_r_Page_013.txt
fee66442dd487a8dea9642ada8f5022b
3cd85aca68f42f72ba19dca8db96f1e94ee952a3
7281 F20101109_AABFXI hamey_r_Page_144thm.jpg
77fe50e3cfb389c2c0cd43ad72073fb5
0409a66afc81f44a6be8f6ea8e7b6548c3e6a8f2
2064 F20101109_AABHGG hamey_r_Page_024.txt
cb01c8aa5f8602a584e8317630251524
7c1ce8b9cd2928b0ff9eed08a0670f648d7f4a11
54619 F20101109_AABHFR hamey_r_Page_131.pro
2ee711bd564bf0a6fc46a0c0af848d96
557cbf76b4ca62d73e5f9369c3e8bab2ddfce97e
1006479 F20101109_AABGZX hamey_r_Page_042.jp2
42d9b2007539144ded296e181cffc411
fed1b14d68a65d002d60684ac1e4d5d0d00049f6
F20101109_AABGDE hamey_r_Page_060.tif
af11a5368723b64fbe1c31e876c5203f
b36736356f22398d65b1b63a533d6dc5cec9c1af
F20101109_AABGCP hamey_r_Page_080.tif
2a4d6dd95c3a225dcf8242f3122b7f3e
cf0270e46c4e4c92db8d4929049d679ebb87d938
F20101109_AABFXJ hamey_r_Page_131.tif
993ed737353e73fa1190717f44fc8f9c
d817d55d662a67f8b178c2351daddd59898dcb21
F20101109_AABFWV hamey_r_Page_061.tif
7141a674c95f17602479e8590542a11d
135e4c86671ef62f75e32e8c5ba647da509a53e5
F20101109_AABHGH hamey_r_Page_025.txt
1320188362ae2ad6d6b211bc2eafcee9
5adc22f3e40dc1eec1232af8ebbecd655bee2b03
54446 F20101109_AABHFS hamey_r_Page_133.pro
2407c340d030e0bd63fdbd3850249160
a5cc593a8311f21c86878aced05cd9946c2dd60d
1051951 F20101109_AABGZY hamey_r_Page_050.jp2
533b46334b37f46d6d0dbb85877aff85
4877d093a21f2d89182f30747206633cac432d06
F20101109_AABGDF hamey_r_Page_049.jp2
e880c24c94730dfdd38c31f2a8b5f60d
b91330df68c4f254f07138964dd641ae9586e715
4524 F20101109_AABGCQ hamey_r_Page_107thm.jpg
00d97b50bc75ad35d4b31a58657265a0
a1a917ba042adc8a70c093b5eb5ee3ff65c2ef6e
27170 F20101109_AABFXK hamey_r_Page_073.QC.jpg
88b691d5e8a997f91faec87487851af5
0ea826a1e55a940a5fc0175e71ad3987409daad0
7353 F20101109_AABFWW hamey_r_Page_087thm.jpg
f99c43d060cfe3a3d1fbf05c0a574d24
e0072da4b71373f5050da59f67e883ceeeffebec
2231 F20101109_AABHGI hamey_r_Page_027.txt
f84ecc6ce8d1097f66501e668dfc2d39
57a7392f1d0f7d92a9685c870c5ea2b27dba6b46
49354 F20101109_AABHFT hamey_r_Page_143.pro
f05b6ec522fec159111776e7c1b347a7
8df90a2adf09faa2e403c2c6109f97145f360574
1051953 F20101109_AABGZZ hamey_r_Page_051.jp2
64a5c28a5762a20ed715eeade43c2bab
5859e169a8feba85346c293f23bc33d88300a57a
23076 F20101109_AABGDG hamey_r_Page_001.jpg
11802534862c8c98271e7f080d05d7bd
765b79eddb8112ba529abaafa7df8d67677bcd5e
F20101109_AABFYA hamey_r_Page_003.tif
11dc5a5cd1e25749f9d3d3999e961447
c9c4fa1a9d05d613f13cf6af53dab91823f5a42c
F20101109_AABGCR hamey_r_Page_054.tif
6336ac430cb9cd99593dfb949978d02e
a4de3990ec1936c08bd9b14cfa6cd9f6de548202
12011 F20101109_AABFXL hamey_r_Page_116.QC.jpg
2c8897c61b9b229a8eb9a5b83f513dd3
28b5c41ba9877286d500f2d9f0fb76ae4a045f90
34741 F20101109_AABFWX hamey_r_Page_096.jpg
b3d27d794420a078925b21698cfb8ac6
037dd00b02daa8d751b1458fcb7bf3451e2d1230
2206 F20101109_AABHGJ hamey_r_Page_030.txt
5a363db19ee3a324611b06834121db76
3ea7bac50412c6a23195af238a814288886081bb
56919 F20101109_AABHFU hamey_r_Page_146.pro
ae2c53e452b2164a5a01aa75b9c47889
946a82dd0477a18c1832b2d28b0a91de57927bca
712 F20101109_AABGDH hamey_r_Page_122.txt
4c8fd487f5ea753f3219569dd9ca6515
a95c4e196f64417efac423754faf9552a67436e0
2151 F20101109_AABGCS hamey_r_Page_053.txt
c72ab9f51482896c04f317bdd976e433
e46c9a787aac36144d5d6a032fe4a7e2b9331666
F20101109_AABFXM hamey_r_Page_008.txt
6efea6e650df228b6059203daa7d61e8
1e5d165a3a315a1d45b5cb0a716032c218a5327a
F20101109_AABFWY hamey_r_Page_048.tif
9547e0d362f9a48c237c4e9c8aa14268
bd85580e6844b63f78ca4ac26e356325e4ac52fd
F20101109_AABHGK hamey_r_Page_032.txt
71d80079aaac8d8ad448f618f8cca78d
a2f63e193b8a569dc1faa6402521034c90f9b1b9
49796 F20101109_AABHFV hamey_r_Page_149.pro
b69c2375a0f506758c802aeeeb073c21
66594ff73c83a5c06bae324f6371c9cc05ec9dea
40885 F20101109_AABGDI hamey_r_Page_094.jpg
4fcd23998d5c8102ec64f5a21da0e565
be3910562f7523584a4a8516d63a3b185662b146
F20101109_AABFYB hamey_r_Page_054.jp2
77f7f8967a890b9fb29df5665af7b571
18c90bdba09b3f25b0ee3012293f3ea3efc8771e
7084 F20101109_AABGCT hamey_r_Page_055thm.jpg
278551d5546a21929f3b8720cceb3462
d444252a26a3f20dafe6ce2b74b311357469f44c
F20101109_AABFXN hamey_r_Page_017.tif
49b7736417b5c28b35d81574c9ec3f94
688c8878fec2dbcddfd91d837b5433df92608e77
F20101109_AABFWZ hamey_r_Page_079.jp2
e2c333694568e8f58eae4397e638b571
d4fcfffed4a663e828689af261c2e1d96596f4fc
F20101109_AABHGL hamey_r_Page_037.txt
737260459682b1c3ab724bd2bfa87b21
26768e962823871c2fa814d0556c9b26ce3ecf25
50394 F20101109_AABHFW hamey_r_Page_152.pro
9bb12e0795fbb5a35fd08cd914cb225b
bb6af1ee8082894becc71f73f7ba57d6b98c8a12
78905 F20101109_AABGDJ hamey_r_Page_041.jpg
7ed6cff296013760db7a662a789c25ba
cf2c25de7c7fcc0bc936a1990733cc411a59f2e9
27327 F20101109_AABFYC hamey_r_Page_029.QC.jpg
45f284793ce715d86decd07279bcc753
1cbee121653fffc38382d002dbc3d60898169aa9
2547 F20101109_AABGCU hamey_r_Page_123.pro
1fe4de40db51810d729182e56096c136
3d01c95ebcd6ebbb57bf6ac0cc646170e54b1f2b
F20101109_AABFXO hamey_r_Page_057.jp2
0eb7a12131029cb783cc604ee86792e9
65fa182447e65833ba719eb669a8fa13533687eb
1018 F20101109_AABHHA hamey_r_Page_095.txt
8af30bfd8af29f930c8bde5b088f798f
474d5a4cf4c45292e44f57fb51493c5be8347e9b
1865 F20101109_AABHGM hamey_r_Page_042.txt
ba9349738f413c6cf9a4391822dc19e2
33da1da51e9e0c3fee0150639db84b3ccf50a2bd
22104 F20101109_AABHFX hamey_r_Page_154.pro
d93088e4935cb298c19f84f05195d6bc
e800146dee924522b1d7e185789da30b60dd5c61
3616 F20101109_AABGDK hamey_r_Page_114thm.jpg
3e228ccae6573b98eaa4d53dcd02d5d8
5f75f1dd8d7230afb6c87ed6a5120a6cfb6f842e
F20101109_AABFYD hamey_r_Page_030.jp2
af4b4499d21117a1b469086d2a277f51
e1592c11bfbc0696aafa5dd04e596f4e60e87ebf
8734 F20101109_AABGCV hamey_r_Page_122.pro
078fea248561e1717a898dca5505bc94
5e50b136d56b3fab3bb5c27bda16cee3a660d3de
83133 F20101109_AABFXP hamey_r_Page_017.jpg
6bb727157fdecc38320125dfbcf2941b
33aadcfe7d64b25f4f52e0d480c9a440120a9e7b
635 F20101109_AABHHB hamey_r_Page_097.txt
4e6c303af0743810abbd16f345ba2fdd
47158b96e6c35ee9d9882edbcf3c428236a043c0
94 F20101109_AABHFY hamey_r_Page_002.txt
ca533c37319aa7864975d5840e4043a4
e6329137fecb5b4df75138d03bf3a5ce9aff7c8a
F20101109_AABFYE hamey_r_Page_021.jp2
09be568514385879437d9a081861b36f
5bb218a1a923db5f467c79551209dd4013bc65a5
87686 F20101109_AABGCW hamey_r_Page_145.jpg
8f7493dc94f3df1ad36cc35de04221bf
66ca40fa1b7cb442702ecdee154b7ea87610b69f
F20101109_AABFXQ hamey_r_Page_078.tif
c710151fa3ef5a41c173123aa3246ddf
baa7cfa1c27b311afcb6d8c75d5349eea4244bc4
2882 F20101109_AABHHC hamey_r_Page_103.txt
e7a63d6aa2563e2b216f1865acc43fe9
43755afd828bdf6f24ad91500d8c3008449b44bb
2103 F20101109_AABHGN hamey_r_Page_047.txt
fef86bf2e583e67bcf2a1616301c765c
b44f40c79d6c769df687f859ded9c508be9d7e12
1392 F20101109_AABHFZ hamey_r_Page_004.txt
801ee4618a4c4c0183b231cfe01fa804
b4a25a3e2af866c4ad6799a261274f492f31ca3f
F20101109_AABGDL hamey_r_Page_105.tif
c65dc7a5fef27c2c63bfade212d30aca
e7717b1a5e9cf8fc52b16f9b9acf28bf3691b21a
48646 F20101109_AABFYF hamey_r_Page_150.pro
47a5058407cd0ffa284f823140791585
2c6c86362cfa4fffc3120dfd1e5fcd20c97f072c
F20101109_AABGCX hamey_r_Page_089.jp2
d8830f176a8e9408db80b7dc9589772b
126eb61e25e91c4c1d0edad7ae88e4ab630feade
F20101109_AABFXR hamey_r_Page_076.tif
ffbc67ed7a78d80217af08c639dc395e
c824156b7160dbf7d6954fb1315cefd7a0dbaed4
14339 F20101109_AABGEA hamey_r_Page_019.QC.jpg
49142eb7611a6d52f1c35cd673bb2409
7dfcb2bb020e77ab96976f8ebe8eadf876f3f3a1
315 F20101109_AABHHD hamey_r_Page_116.txt
637504d2d56c3b81ed5fbf02eb387b20
4e574b932882f7b28b801898d9b0f0101ec7d6f6
F20101109_AABHGO hamey_r_Page_050.txt
2d4cf82bb31158480fa8eca0b925123d
8014b3181ef1e0a4958208ef91f5fe338ff6e489
55045 F20101109_AABGDM hamey_r_Page_031.pro
a687ad0403b1fc920a729120aeca4b3e
808f45f1eae3b2d63888f819c17d076d0067e6f9
8805 F20101109_AABFYG hamey_r_Page_119.QC.jpg
ba810af55fbf313dfcd90b6988327869
b1ae866833b73af0761396a3e694950dd5d7b47a
F20101109_AABGCY hamey_r_Page_080.jp2
24afa9c7208d023d114afaf0a94de30e
7571b429b9eefbbaee9e78e814af146e6de1c9ff
23190 F20101109_AABFXS hamey_r_Page_102.jpg
93b001bea4801d0d268e60e93274425d
cca30eff0eb2109ad3f4fc7922be3e9421422ddf
4091 F20101109_AABGEB hamey_r_Page_040thm.jpg
ee33ceb55d65c16625ad4e4ec873167e
b28fad4c598fcb957ecaf47fce0a15d5b59319dc
182 F20101109_AABHHE hamey_r_Page_123.txt
71362477b52e538a9fe9b04e9dce7704
9c6bc1474c783e094b8a0e76e04a82d0eac24c79
2015 F20101109_AABHGP hamey_r_Page_052.txt
96f78109248d1f7e0add37338089a370
9e079368960b5ea42db15883f3f104438e76614f
33861 F20101109_AABGDN hamey_r_Page_066.jpg
f9d2e3ec4cb2286a154d87e1ea50869e
35e87958b15232e72835659875cff3e6f64be2dd
299 F20101109_AABFYH hamey_r_Page_061.txt
4c7bc9f112cf81b7bcaad566e2e0a201
47e36a1216295eea206f3bd4d84b91d2d64126ef
3125 F20101109_AABGCZ hamey_r_Page_097thm.jpg
751cff7c8d4f4875fad35f0c2ec2bf0e
7995e82142a6c916a8ff2540739c7ef296870b8f
87169 F20101109_AABFXT hamey_r_Page_151.jpg
2ad936cd578e83b0d18a8b7a89429338
4d910140ee11a16c5ebd2c9676d0ecfc64da2a25
7492 F20101109_AABGEC hamey_r_Page_011thm.jpg
6a213014a7d0550de65c4da85cab7c38
5705c4ac2776ac79d0c5eb17f2dfe7a8211124b2
248 F20101109_AABHHF hamey_r_Page_124.txt
baa1b721bf395e754d1cef2ae87323b3
18bc71ca1da938e5a85842edea3f5faf92238d24
2209 F20101109_AABHGQ hamey_r_Page_056.txt
1c0620a0b8f20be556225a8c87742310
3ed5ec079231dadcb633e1462cb7d80276951dc0
4645 F20101109_AABGDO hamey_r_Page_069thm.jpg
a17995cc0f347dc797374b08c886cb87
32d265aeaa85c43898b285c4b27948e0423933a7
25079 F20101109_AABFYI hamey_r_Page_082.QC.jpg
c7ca2a7cfd4e0c4a30317b106eeec2e2
009ff38f9c8645e902d2dbe2ebaaadbe2e2a2795
1789 F20101109_AABFXU hamey_r_Page_067.txt
b4c0ab5e46f99923a49d27fe1646362b
81e0f21da98674e9ac20aec680543eefe0151583
52529 F20101109_AABGED hamey_r_Page_130.pro
ca82dc1915eb84111e0fbc38381f9270
7025eb5567d6120f3f8a852bb865a2664c799b44
271 F20101109_AABHHG hamey_r_Page_126.txt
68966642212c1f78e0916bf4c4520dd6
68e739cf1a5e726a725bb0e271ab426030414b5a
283 F20101109_AABHGR hamey_r_Page_059.txt
c158aff33ae73dc314bd0d3c1aed81f8
ac8bc04eaf99c83ae45f1c38db19c007753cd709
53494 F20101109_AABGDP hamey_r_Page_049.pro
f4d60f948847911abf8b3d1eb81dd5d5
08130e0c38675a2c8dee8150066e64bd6193a888
83982 F20101109_AABFYJ hamey_r_Page_086.jpg
f10603523b5fd05109436bb1c5640e58
4ec96dd288a540e60fa7acaafc2147c1be7bd7e9
1051944 F20101109_AABFXV hamey_r_Page_047.jp2
bf83da9252643bb0406e8f1ddeeb5375
bebc8149c9ee0e48658a67e59ad21a8a40301de8
545 F20101109_AABGEE hamey_r_Page_036.txt
7ac3d0b1dfb4c36bbb89e4484e004071
ccabc28f5895862a2c06adc533bf726cec796863
1973 F20101109_AABHHH hamey_r_Page_127.txt
654ed0bfcab11afc02e91daf86018141
9b6a976795a70239770bb12e486dd4c2d0373731
452 F20101109_AABHGS hamey_r_Page_070.txt
e0635cd49dd7e8fd4635aa8a1aabecb3
b7134bc44501e1288d85d6c2dcf16448c988a22c
11383 F20101109_AABGDQ hamey_r_Page_095.QC.jpg
7f054dda9b7e6423dd67690538f96cda
a751827284d04d29e88bf562e223269a64da6380
654149 F20101109_AABFYK hamey_r_Page_138.jp2
d9688b0823fb51ac1b508ba892e37c3e
e554d21195cbd4624033ddf85018295df1f59c20
151 F20101109_AABFXW hamey_r_Page_020.txt
f4fd54b8aa2bc25179e6477c8f169c7b
5318eb6f3bb69dce4d44d5cb3eeafebda488d803
27176 F20101109_AABGEF hamey_r_Page_086.QC.jpg
e9e88d828b8e8cb923126a675242c5c5
5ecf9a7b2dabb9a4a4a65b84f574bc7ae74cac93
2268 F20101109_AABHHI hamey_r_Page_129.txt
93b5992bc2f4017402d22a6bf5debe8a
59b099d6acc8e235d2a672b4b9ae2c44cd0883e2
221 F20101109_AABHGT hamey_r_Page_071.txt
ac7a501bcb087ef317af8bc3d9779dcf
963dc944df9c238a6cb9e842b30c25558a152801
81143 F20101109_AABGDR hamey_r_Page_078.jpg
e13a73416eca0d358961f82353f0ef0c
d3690fabaf3cba39cb68682de7b059f732757c39
83262 F20101109_AABFYL hamey_r_Page_018.jpg
46b0cce6ccb38611d51106aa6f649980
ad0a51619553936dd8f44746c226f2931796cb5a
F20101109_AABFXX hamey_r_Page_087.jp2
875e51ed4c381aba3edf93e4d141e34c
a2ff59de26a8cc18e3dfb3cfd5331731e44af0ee
71121 F20101109_AABGEG hamey_r_Page_076.jpg
c4fe651e50d018e7b6adda3762b08fad
a7ee77c2e79319fb382764a2949d2017516b3451
F20101109_AABFZA hamey_r_Page_116.tif
2d3a296347aa56692acb313baa4745e7
2421ca0354b54602aa86f8e02496294a8b2dacf9
2072 F20101109_AABHHJ hamey_r_Page_130.txt
b8cefb4c0756a06fccb04c9c26c9f350
bb354ba8a2629edcfa8ea4498e6a26deed1d4e8c
2019 F20101109_AABHGU hamey_r_Page_076.txt
be982af9b90d9271623d2e08fc93dbb3
195e298bd4dec8bcdaf80a797a602bf7df716286
367465 F20101109_AABFYM hamey_r_Page_106.jp2
30084c6390bdc648977616031d439386
45451c2beefcf1a1262b25d64ec9991d5650d47d
2210 F20101109_AABFXY hamey_r_Page_084.txt
3262f33f884d9de7f05339b076b21fb7
7be939fd8f77867e7007a8838529e0b757e28e73
47432 F20101109_AABGEH hamey_r_Page_076.pro
1121ff24f603564b415f1c4aefb64095
cb2c85403c63da898a49c6fdcb2b8fbdb0451eb9
11379 F20101109_AABFZB hamey_r_Page_100.QC.jpg
f4719060df102817763d7e25d42d3bb6
9400193027ba0b78ee735d070d026ac67689c93d
14442 F20101109_AABGDS hamey_r_Page_060.QC.jpg
85c422f888dc77a04de4720b5da1a8b7
0793607253b7755adf80beedfa74125815962157
2136 F20101109_AABHHK hamey_r_Page_133.txt
20d1dc674d8b7e6c78f7d109d620040e
fa12dc8a897d938bd662535b4a2a740d33d6a8c1
2140 F20101109_AABHGV hamey_r_Page_077.txt
db6f17e40eb7513e25149c2ca5b02eb3
3be7d99755b4768b4167d0cb95e11f6ba03225d2
2023 F20101109_AABFYN hamey_r_Page_144.txt
034f98e1cb1821868087fd93c812a932
f9bb97df57d64e177d1d3168aad4671071213849
588 F20101109_AABFXZ hamey_r_Page_068.txt
b2091a96292a54fc8a5d1eb1a3221d7d
08957826f5bc686b0176216f194c7792631ba847
F20101109_AABGEI hamey_r_Page_028.tif
0dfc701c58fd025b3929c840243af124
1c401a63c6d5595cdc76cfc4025b675c200ac9ce
7316 F20101109_AABGDT hamey_r_Page_058thm.jpg
db296699b20dd45b045676ccc85e8ca5
6acf086d4790564547a26cf50868e943e622e897
1708 F20101109_AABHHL hamey_r_Page_135.txt
7d3fe00c2b07bc3ef4878cc58ea351bc
fd22ca3caf3893fe953f914cb484042c761054ec
2190 F20101109_AABHGW hamey_r_Page_081.txt
d03e25e65afc34f424f592b56b018077
a81cea3af8403ce748da7e3af51c08bc3548f713
10101 F20101109_AABFYO hamey_r_Page_112.QC.jpg
10cd72a162957dea216c7a5d891b1414
4ebc2ce16ddb44581b6b19da419f301b031d56c5
2155 F20101109_AABGEJ hamey_r_Page_086.txt
dab135f722d39b28560fe2bb6c3a8a4d
1afa5788ed861d86be037d530b5926d6fbbfc1a8
44883 F20101109_AABFZC hamey_r_Page_125.jpg
df3d3acbbcf2e3d043507cba0699b754
1747b820c84174c99aba8a643a9321f173adc920
38346 F20101109_AABGDU hamey_r_Page_016.pro
1107cd6b8ac02a8779f301ebe57d0cbb
2a8df6c0e0abbb5a7eebc79cac6a070fbbf46008
24449 F20101109_AABHIA hamey_r_Page_127.QC.jpg
a4878858dba1b33a0068d0d5c8fcd53a
ba4a4fa1725ab999890819614f932036b8f15223
443 F20101109_AABHHM hamey_r_Page_138.txt
e0d70f2f6e00ac336f46b7c48adcf74e
1f90a0f31e61b8baf65162c99d46e4fde1d23793
2115 F20101109_AABHGX hamey_r_Page_082.txt
9cb96656171ce057b9abd0bf496a3def
fc9b6f3cfa19e83f3e3fe930647c5b1bae1a4785
5154 F20101109_AABFYP hamey_r_Page_004thm.jpg
310aa4818e45db1fcc6940511dcbb357
216268e6cf88aed192b34b86eff5b06daae8082e
101035 F20101109_AABGEK hamey_r_Page_148.jp2
c7a4d8e8166b6640a16b506039631bd6
de294f55e39e89b2df3b542a9f86503644516195
7487 F20101109_AABFZD hamey_r_Page_029thm.jpg
22e9519a3123479034bce8a728b2d091
541d6b168d28cefe374d1cc4c9203d72e9c5dadb
54706 F20101109_AABGDV hamey_r_Page_073.pro
2bada825ce70b2dc54af7f0c9eeddbbc
bcab683260fa984dc616a6a2ad68f2b51dab8f48
9099 F20101109_AABHIB hamey_r_Page_109.QC.jpg
c698e26b6c579d4cbd8185bddd5054cd
5f2d6152c3694c728a66f37a8dc8542d1e8efadf
484 F20101109_AABHHN hamey_r_Page_140.txt
283f1a9eac51a4c218087dd8948d483e
642788e8c19695d74c2bb1b5fbe15a1baa6c0240
2189 F20101109_AABHGY hamey_r_Page_087.txt
eba73477c0f5edf5ac50b26c5c477895
d0d9ad87b277ffc56a4b5fe3535d8828f7bc091f
7553 F20101109_AABFYQ hamey_r_Page_056thm.jpg
528da331430b4645180191fea854b4ba
b19849bf4038ef1dc142f53513d0834b23369a96
78945 F20101109_AABGEL hamey_r_Page_051.jpg
2aef1eb32c3202c1cd187a899ad75d9b
5abb0a82903f7b66ff5a82571f3382df35ca0a0b
6199 F20101109_AABFZE hamey_r_Page_115.pro
37a7d98dfb52bca62a905896b281f1e2
4e80391df5476abc687d48668180176417caa323
6371 F20101109_AABGDW hamey_r_Page_076thm.jpg
6d0e0e8b844e8b63d1ced83b1b18b5d5
0786d9d1d5e1f84e667ea4558bb2f42f64f46a4f
15569 F20101109_AABHIC hamey_r_Page_067.QC.jpg
8a87ae00973b6e71aa66f62089542bb7
0a87aad384cf3614a587e1af5258802a923d0db7
2248 F20101109_AABHGZ hamey_r_Page_093.txt
5202208129c458239c98cbc68de10261
f9976bd670b6638f43edb7334a040f600677eac1
20653 F20101109_AABFYR hamey_r_Page_064.QC.jpg
be1c698cac805bfe2442a73c1e0a02c2
2e21b8b7bbcc60b6fb352afcdf3d5b62ed1671ca
7332 F20101109_AABGFA hamey_r_Page_069.pro
f0d2c50232082b84ce2a99c29a89de88
62acc05be92b8431049ef66df9f43850ab7d5712
25042 F20101109_AABFZF hamey_r_Page_051.QC.jpg
3ca2b3d699f50a83cc9009c104963a1c
0fe42136e7083e43406746484e0803e136b5cb13
25662 F20101109_AABGDX hamey_r_Page_134.QC.jpg
68f884a9806ec0b9466d30678901a5c8
f2920aab57960db6b6ce9486ce66ade040c59030
1625 F20101109_AABHID hamey_r_Page_003thm.jpg
f271b005111aa7d582b3733238c32699
f93861295e0a83acc9a1b856e1fc549fe46ca959
977 F20101109_AABHHO hamey_r_Page_141.txt
3d1e075ed4a0ca50134a6432baae2780
921b98d81b23ea63e1b9f014342505ac04c7e2c0
123 F20101109_AABFYS hamey_r_Page_003.txt
00f4aaed61b4304ce0f35c049c5c2c11
859a106bc5d6721eaa28d85467a5ebdbd15a42ef
517544 F20101109_AABGFB hamey_r_Page_094.jp2
b5be2f9d23666ffe3145e0643d2c0c99
810de17e5d70e3a62c8433372490169ae6c05c37
14311 F20101109_AABGEM hamey_r_Page_107.QC.jpg
383aca07dd41f24bf1f71506419b7449
3d3d13a8f18eba4ba8f274ac5190202043b099b1
77506 F20101109_AABFZG hamey_r_Page_046.jpg
c7f3e6dad18214408c7b7707554954e0
529e28aff4fb8d11c59fecba2d1ce880667d3160
F20101109_AABGDY hamey_r_Page_014.txt
fc23a17da2319994c6f206ce04f8561a
e0d461dce791fa7d381e482bd970ca579004358f
3327 F20101109_AABHIE hamey_r_Page_096thm.jpg
ce508e725bb49286f8e27b1117b73b32
c415c99f5f7e572850ed240fd755141d79efaef2
2082 F20101109_AABHHP hamey_r_Page_147.txt
c250b6df5dfbb523925c7f9194ef8ae6
a1c4e3842ce1f747cd26ffaee5e2a9903211c356
581 F20101109_AABFYT hamey_r_Page_110.txt
874a4f37d4ebe94a0b6c70c000539e15
54a5b2051ed646b50a779b6a8e9ea51abc252cff
3263 F20101109_AABGFC hamey_r_Page_062thm.jpg
7e706d5d0361df219b63dd844e5103e7
4147b80d1b415157cd24cf2e69696a042d84d264
77849 F20101109_AABGEN hamey_r_Page_034.jpg
4316e7664e2950883809d145b0bde28b
e2fb04b91246b40a52cd8822811bd2950286cd54
F20101109_AABFZH hamey_r_Page_127.tif
e723236e545ed6c13f1fe718547d84af
8205354d795a80d8b9337c291ca83ecfc7303c49
87564 F20101109_AABGDZ hamey_r_Page_057.jpg
082c27f0ae602aff3cfba2065797f550
ad19ca5145d83c026dfe72b2c8951710459b70d2
7448 F20101109_AABHIF hamey_r_Page_073thm.jpg
a98e33e124101fc76550b5baac5f1e0e
6e96f6000c0ea95c324729f0ab04d3a4b7fc3f1e
F20101109_AABHHQ hamey_r_Page_149.txt
1ce1324209fedfcce4ae3f3c020c89ca
8c088e1ff3fe029d3c0e896e845a121ef119467a
24041 F20101109_AABFYU hamey_r_Page_150.QC.jpg
6363c77c2a8e6469637080d42c167f33
a234c8840703223c32287b3822da01d7f72dc464
1076 F20101109_AABGFD hamey_r_Page_066.txt
93a232f58dfc6ac350a9b33ee9edfdce
a87f618b773a26b4827da46a0ef62822d02d42db
992 F20101109_AABGEO hamey_r_Page_098.txt
a880c59c559d25741631bbb11e028fc9
0a02520a1b15ab9d57393faff6757cbeee8e53c7
F20101109_AABFZI hamey_r_Page_136.tif
6d4a6c47ecb3d9478f8a019176333276
83717e0959a6ebeb110273194bf189fc7b2f5117
25496 F20101109_AABHIG hamey_r_Page_024.QC.jpg
b9389d98824bb71085cf10ad69ea8ea1
8c207502043b5236434e60225096c58cd74263aa
2006 F20101109_AABHHR hamey_r_Page_150.txt
cbe6eda44da81d091d533b2e2186689d
868607a285b04f6431c1513dd23d6f33c2f061fc
32965 F20101109_AABFYV hamey_r_Page_115.jp2
f50067e3e5eb6dc72a4193d578334304
4af7ed5126ade35613f4717406773fa6b57f6e7d
3227 F20101109_AABGFE hamey_r_Page_120thm.jpg
dd4e8aa4b4fa05221d5238bceb8e5c53
d80cb04f327c66693cd3610eb9dfc7180a85a679
1838 F20101109_AABGEP hamey_r_Page_118.txt
4ebb5b5b8b53f0df4545ebc9ad9a8510
8979e3c44b6215dee1dfb1269734b6a773bc8d7b
77276 F20101109_AABFZJ hamey_r_Page_028.jpg
8eaede14f406b469553e67f156d070b5
7f2931b1c67324b613ec0a2c3ff8a777f1e51b89
229006 F20101109_AABHIH UFE0022125_00001.xml
6349af23a70d14fa6691feae5352bcf1
3ad1373893f56b719ccb08a042eb4d2fa72112bf
2042 F20101109_AABHHS hamey_r_Page_152.txt
12c74da5089efe90149fd4d153c80a53
c45a386fa16c9e6bf79d4d9d17ceb799f14f5885
730844 F20101109_AABFYW hamey_r_Page_069.jp2
b5187e6528d1a115ca25688ef3e2853d
39e6fe3a57fcc18f79d7fc7f52a5aba3a5d11470
F20101109_AABGFF hamey_r_Page_081thm.jpg
ea6362ca22d6e66e430bc9c653f14421
5b6fe8ae1020b6c7bcfa0bb3a91acb827d23849a
F20101109_AABGEQ hamey_r_Page_107.tif
a0a1a914898d175f07ed5fed5910b6c6
4b3093dc794b54b99507f44638203e3b3aedead4
7715 F20101109_AABFZK hamey_r_Page_088thm.jpg
91657550d2fbc164457ecb2086094ce2
79c601982b815964ba59d53d9fe00ffcb0209e7b
F20101109_AABHII hamey_r_Page_002.QC.jpg
aa88793247885bcb1aace4410dc37048
0d1afe974d95423fca8cadb754294df15a80c509
7069 F20101109_AABHHT hamey_r_Page_130thm.jpg
d521a8db0dd0407d8476cbe76ba83f49
fe3ba8171c6088777860850a06f6433cbf08911a
692 F20101109_AABFYX hamey_r_Page_117.txt
1e11d306805402413d3783d53946b874
398eb3c9c8f070818112ab8a02c00c539e105b28
F20101109_AABGFG hamey_r_Page_141.tif
108801b4f1511d64eb59d2d298896847
2e530141411f6e7a213858d766b4dff26f5d62ff
49762 F20101109_AABGER hamey_r_Page_014.pro
9cfccc18c953079b0c99e6c8466713f8
fdde02372fe6af5113262d84602db224e6919ef3
88038 F20101109_AABFZL hamey_r_Page_092.jpg
587f9bd287557fc1cca68350962a73af
b397931a70861d7ebefc2e4f2a387e15b95847fd
3443 F20101109_AABHIJ hamey_r_Page_003.QC.jpg
b5fd0a8a03aad71beda3865b4a690471
5a273668d1a0a7c0f0c3ac10f9b7b0a73f0e65f1
6872 F20101109_AABHHU hamey_r_Page_134thm.jpg
d1d09db7529ccdff4b10d0427ece9b8e
526698551bdee2391f56b236b3664767ed59248b
9699 F20101109_AABFYY hamey_r_Page_068.pro
78bbac553d7377fd6590a4eb8f33104e
799bb0b6913e90a9e8d911b0228c7986fe0d3b55
26924 F20101109_AABGFH hamey_r_Page_081.QC.jpg
0da47516209c350da81dc6777833cab9
ae7bab814be25fd5410de1dc79e9dc12a3b37275
28035 F20101109_AABGES hamey_r_Page_050.QC.jpg
9bae1a75b7d455d53bcc79c2a9190b8b
9df02e6567ff4a39ef590a82df0cbbb762222846
F20101109_AABFZM hamey_r_Page_085.tif
d1929e4c23b42e3782537af4ae0b8c60
d4ef17ab120f75b1b8596be4f4d8056338e7eff8
2213 F20101109_AABHIK hamey_r_Page_007thm.jpg
69e549755ffedceb95b770f23dbed3fd
8f8580df0679fa4090511e9ad2411455b1418c74
7272 F20101109_AABHHV hamey_r_Page_026thm.jpg
06c34165f116ba8639d49a496212a77d
8bfa651f69d63b97b2b7f382193f1935ab884ff3
867 F20101109_AABFYZ hamey_r_Page_108.txt
1b70329b233749a35a9e7a255ec9c2a7
2d3166d63c0d83da1b28f4f056138eaf0214ed95
10200 F20101109_AABGFI hamey_r_Page_038.QC.jpg
eb9661c4351828437c0d1146dfb097b4
ee989087ccff4d078992b0e9f19e32e9b5540261
49511 F20101109_AABGET hamey_r_Page_128.pro
1bd2a9e6a270db45b7f460c0e7c34284
ba7aa81f5c149e75319100a3c436002ab1f1812a
24382 F20101109_AABFZN hamey_r_Page_034.QC.jpg
aa705d47133fc5c2035cf2ccf09e5855
851377e018d043a7a88d28978df92d51f3fd2e7e
6951 F20101109_AABHIL hamey_r_Page_008thm.jpg
780b2bc56d1a29f5d10bcbb8de76cf75
e91d9167d126a9b02785aa56c811a5eb345f2916
6883 F20101109_AABHHW hamey_r_Page_023thm.jpg
803cf384fb249fb82e7815cdc93e7077
bff33a640a12589bf04098fd2128c01ecfb78719
26928 F20101109_AABGFJ hamey_r_Page_079.QC.jpg
f4f99b699ddce3e43a5aa48114354645
9c0be946f3f6c920bf0e09f2daede76a9b101182
13298 F20101109_AABGEU hamey_r_Page_099.pro
894d117b42d29414b205c56aa353816e
e924cabee91be1718f40c1b073796f61b44c1f80
13721 F20101109_AABFZO hamey_r_Page_113.QC.jpg
1fc300da14de51b2e637bcbdb1a34413
0e5fcfbc859720f4d30438331f7295275f08dad2
27224 F20101109_AABHJA hamey_r_Page_027.QC.jpg
2a2b55ac53f2882779bb80d1efe10141
17fe3fe2893442d596db8ff0d416a774cd4dd95b
9303 F20101109_AABHIM hamey_r_Page_009.QC.jpg
845e31d5cbc96307a21f2578c89ae1e3
64cea78d15f0f4b0177efdce30c22fc7c4547140
3902 F20101109_AABHHX hamey_r_Page_099thm.jpg
9a69c2252731b2046091964409a60019
5295590b03c41a24cc4ad911ec196824e4451135
81771 F20101109_AABGFK hamey_r_Page_055.jpg
0a6bd6df434833fb1e34e5929dbe65f1
84a3fb1a1535b6d55c2c8e9a4267428b32084acb
76506 F20101109_AABGEV hamey_r_Page_010.pro
f0a9b7eb5be164f2cebaf38dc5dc803c
7faf48da9226b92acdf037341af78a465f11105e
2152 F20101109_AABFZP hamey_r_Page_091.txt
2da4650105cb7c267ddb220308c37459
5bfa894be741be2e2ac104a5b1b83828b5a93688
7440 F20101109_AABHJB hamey_r_Page_027thm.jpg
97fad18ffc91d2595abdd14c25c262c7
2ce3acfa976ef45b02d846fd5a18c5e85733d420
2930 F20101109_AABHIN hamey_r_Page_009thm.jpg
22c17bf11f52e6e89705ad6d16b9acb8
49448be2f12c038769f93f265cc26898e67949ed
25138 F20101109_AABHHY hamey_r_Page_035.QC.jpg
9bbbfd20565a06b880223a9ddf9ac391
271bc6dcf36cb81568e2e30f08fede328baf2529
F20101109_AABGFL hamey_r_Page_118.tif
695b655c3a1ad4b3ebe7d07273954b4a
1b24cb9325b9cb2a6dbe8a6b907b3f7f6b543394
56236 F20101109_AABGEW hamey_r_Page_030.pro
b342bc68a9256edb07119b4b94ce0197
58e4d7506904f4952207d5b1fe2256533b7e624a
20947 F20101109_AABFZQ hamey_r_Page_135.QC.jpg
becd842f6666583e7100bd52a466075a
eaf9ac3581e2689b3c3d45fa7bc8fac1cca9801b
6600 F20101109_AABHJC hamey_r_Page_028thm.jpg
3ef0fbac5812069bfe97ef13cd8ebb9b
5fb50c317977e8c6b339da4f95e0685ed10efd2b
24676 F20101109_AABHIO hamey_r_Page_010.QC.jpg
49843c0129637b27286d0787157252db
66bf2bfd65a02ad616c655b5b841e3097ae19580
6411 F20101109_AABHHZ hamey_r_Page_101thm.jpg
59eb98aef533d0128948205ca03a85ad
12b873a26cda47cf74ef9f3851adad1e433f87ea
3442 F20101109_AABGFM hamey_r_Page_061.pro
d8879a82c2d6449c0d0f08c125f3911a
6a3cbe12a3e7c7e77ef78a103d60e757d58e5d5d
359 F20101109_AABGEX hamey_r_Page_120.txt
34fc570299e1ce33a8ec0064f686349b
5e175ec70c283edb41464902a84a266223f87a03
2056 F20101109_AABFZR hamey_r_Page_101.txt
20efa2c49a0f08a78cd70f105b6103c9
26741f7f0ab5d13ad33d65b9e6f5d0fb1594087e
2240 F20101109_AABGGA hamey_r_Page_145.txt
87bceef81cf38fc75dc4df39b6bda806
a203c17f9f466152d7f12c1c1ab2d8d297267003
7463 F20101109_AABHJD hamey_r_Page_031thm.jpg
5c06ad70a9c6e6716f95a04d5d656c46
a9dd918c6bd9e114c354fe3a0ca957b39f858e6c
24771 F20101109_AABGEY hamey_r_Page_089.QC.jpg
9109d8a2e044c3dacd09a8bd00990b75
3255f01a3e263746c1cf1f2e384ad4322a353a7e
F20101109_AABFZS hamey_r_Page_077.tif
1716781e09e04922d0236f76286263d3
e9a17f46a4c905baf0c3f4f55e74efa7af0cc5b6
25400 F20101109_AABGGB hamey_r_Page_052.QC.jpg
2d85ea698ead3085e1c1c3ddba713fb4
72775dd1dbad85945fd18e2856a9a0e1c7cb9009
25274 F20101109_AABHJE hamey_r_Page_032.QC.jpg
29821e12227d9e0c3f95a1df9d78fa94
1caca01d9f21874e9fe5a4902fd06a0c978af40d
7489 F20101109_AABHIP hamey_r_Page_012thm.jpg
680e7e183f748464be3a12fd75091610
f01f953112546bcebc7101c503bd13d844fb0026
554153 F20101109_AABGFN hamey_r_Page_113.jp2
7dc410d68ca3ae8211c736649c283225
464e8d3818fdae3ad0ab9b5e8130262fa8aa1619
851 F20101109_AABGEZ hamey_r_Page_106.txt
e8e822b33b49da37bd08a556c08424db
e55e4a25271ebca93398f9d3169e9c720dd51470
561 F20101109_AABFZT hamey_r_Page_105.txt
45fd17b2155c73a42f0eb26957b23750
25208f0bc2235a02972502a041588a91bdb8f416
50858 F20101109_AABGGC hamey_r_Page_034.pro
8d6b11b68501b4199a6520436178519c
83b9d8e7966cce7ada50395509d4698a9c86ab96
12796 F20101109_AABHJF hamey_r_Page_036.QC.jpg
6085dea0df366b47a0c326d6ba5fc52a
6021d071c164b7ff7fd4ac987409eb3de11ef6f0
3381 F20101109_AABHIQ hamey_r_Page_013thm.jpg
9527f6168c3a661220eb81c97561fba1
4b4e3f6261930b74582d65dd242df68a6328791d
46449 F20101109_AABGFO hamey_r_Page_003.jp2
20b69b918bfb65383c16de47b32e96e1
687c08f4c35ecf1c4b5c7391d1a46b9f0d98c54f
39979 F20101109_AABFZU hamey_r_Page_138.jpg
c686737b3b4eceedf58d6e777317b0b8
45079b84a5161b0c93bbd50b0df2b6cd263f0731
8990 F20101109_AABGGD hamey_r_Page_097.QC.jpg
d8eef4a8be29cc2bb38dd93b22e16666
434ba72e766e5d15b4f784ee695ccdb1aa67c238
3544 F20101109_AABHJG hamey_r_Page_038thm.jpg
4ed6adfcbeabb3e491084c505c1f6e9d
0727c8b950d78611713e9ef26c0b5828ae504adf
25838 F20101109_AABHIR hamey_r_Page_017.QC.jpg
a334e2530bacf812e2afbbe344e3ead7
b339a08bd0fa3882ef903c3d36e1299239f8f49a
7120 F20101109_AABGFP hamey_r_Page_152thm.jpg
04f4d9dc64de2f5ebb8357fc43f5f5e8
6ee4c2bb00bf30877f9850c194fbc7f846555d68
499 F20101109_AABFZV hamey_r_Page_112.txt
c8add6cfb7b51db0d776351368ff718b
79f9adf8f01d7bc73f65a2e97eb12b51d8c850f9
11266 F20101109_AABGGE hamey_r_Page_096.QC.jpg
d56a56bdd358e82f65227eb56aa952fa
3749804bb69d08732d8c6e91103643684dd1c5d8
6752 F20101109_AABHJH hamey_r_Page_041thm.jpg
4d0c4d5eb6ed67ec00ec97d430bbcaa0
aa7b485d8d0a8d178638c1133cfbfb07680dfa82
26634 F20101109_AABHIS hamey_r_Page_018.QC.jpg
1f1b716901f987df80da0b0d2cf34176
17dcf3fb4dcf3bf62631bb7a145255e16671edf8
7007 F20101109_AABGFQ hamey_r_Page_035thm.jpg
96224d10698960101bb932b41bcfbd32
d41446d44d6d7ea77484d6ca3e22c1febc5c83c0
F20101109_AABFZW hamey_r_Page_122.tif
2189407f3222e02c3ecb8f60bc9056fc
e30df11c90b7e6bf19c25d99f5c1e2b601029bfa
10024 F20101109_AABGGF hamey_r_Page_142.QC.jpg
77c1f70406553f6563a3c95b708ab59d
06f92f4b54fcc924e0d9c9f3c85e42882f175133
23546 F20101109_AABHJI hamey_r_Page_042.QC.jpg
2f4c49314173c1c2ff1a3831789b9a41
637be0348d2eab249cb5b865bff3db21300a66fb
7498 F20101109_AABHIT hamey_r_Page_018thm.jpg
94317555f8a68621f1838ba83d61c751
64bfbd8ed44592ceff4fb1dcd2f630028b97eed1
999734 F20101109_AABGFR hamey_r_Page_033.jp2
930abafd619fe49abdada04b18ce952f
d17a4d4a19a210ea8721ad7b35c171a703c5ba29
1051959 F20101109_AABFZX hamey_r_Page_010.jp2
1467abc2fbbe38aacee893152e830529
2d2e5dfb0fc415b2dadc9e22e08bf6dbea56b68f
F20101109_AABGGG hamey_r_Page_131.jp2
df2b7974dafded7a287511c40067d035
697d5db30266a9a8b6bd4186b97c2057d14aa9c1
28481 F20101109_AABHJJ hamey_r_Page_044.QC.jpg
fc2a51bf75d52b3babe56b432e0cdd54
c675ead7985ce9b1c8f94fa74848775313b5f6aa
12274 F20101109_AABHIU hamey_r_Page_020.QC.jpg
cfb324c6fb24bbfbb9260ee10f04f5fc
26f12a53ca9f8475aa09001938e92642f13f34ec
376567 F20101109_AABGFS hamey_r_Page_066.jp2
ca166f249acf4c4b6704b8c2a12b0a01
fcbd06c2d4c16333209199bf031cac613b863fe0
833 F20101109_AABFZY hamey_r_Page_109.txt
ea69e6ad7cc557ff938fd5ae8027035b
f6ab2b765c3178d690d0cf869970e1e7fc57e2d2
4215 F20101109_AABGGH hamey_r_Page_068thm.jpg
fa9ec2a21a7032a676381963cd9cdf9e
866b135c26731367007094c579037576f81e5246
27050 F20101109_AABHJK hamey_r_Page_045.QC.jpg
54799ac793e155eb2ac02309b99f7693
5992b422ab85721f6f2719f6c7fbe9c490e21b6c
F20101109_AABHIV hamey_r_Page_020thm.jpg
9c73889a42e54cac7ff34c803b7262b6
dde53cd135019682045459236f16da32ccfd211a
463022 F20101109_AABGFT hamey_r_Page_020.jp2
458ca1694acc89852d2257f3c5285d4d
c691c680aaf60f5b24611f7e0ac01f29d9bb6d5d
10826 F20101109_AABFZZ hamey_r_Page_059.QC.jpg
2188c348806074bfadcf7b734930ccba
5f03115666406255f09a263a802b49b312a421f3
80512 F20101109_AABGGI hamey_r_Page_024.jpg
058ae5e3b5bbd8b6f1209f51e6280b72
c71858f41032feafc3a36a609cc9e86e3a972e6a
7600 F20101109_AABHJL hamey_r_Page_045thm.jpg
bcf905d600a71dae263287bf00e6354b
aa4ef69a86c55009ca04d6c6892957a9fd5d6a19
25059 F20101109_AABHIW hamey_r_Page_021.QC.jpg
e4602beba4fe996c7f0656f7e7e327e5
0f92d68f57d015961faf969e1cc560aaa924b68b
9599 F20101109_AABGFU hamey_r_Page_040.pro
0232a9cc0e87ee3684a55c778a989804
dc5d4e9f9fa7513765d56b1a8d158a55d846715d
1069 F20101109_AABGGJ hamey_r_Page_139.txt
390bb0ff0d2f4da19d25a3dbcd1055ee
68265926c9ae76b406df0296af80b55014835d19
7369 F20101109_AABHKA hamey_r_Page_074thm.jpg
7574f7f7cb5016624632ff7785342227
04b2f759bfe2b6fc60c5587c8013d6d6addb0f40
25558 F20101109_AABHJM hamey_r_Page_048.QC.jpg
8580a13238a9f9c03d81f927b2a136bc
289b245722fca5f8b6eccef6e5bfa8b716960bd0
24697 F20101109_AABHIX hamey_r_Page_023.QC.jpg
a8b9a388e5dffe90829ec4c238ace3ca
4f628070a4849f0b33bcc0f3a7edeb468520a935
82662 F20101109_AABGFV hamey_r_Page_147.jpg
1bca238bc5bbc32838e67c2eeacaa514
1d86ace2d1999e4fbe41d17a7ea797acd993bdd6
35483 F20101109_AABGGK hamey_r_Page_120.jpg
c570c801610082ee2108bc4167ebb640
6c07d4b966336b9efb030fb2182b9d034c240fca
22288 F20101109_AABHKB hamey_r_Page_076.QC.jpg
11735d020cf007db48c1f66da54c1864
1abcd123eeecb33f1049d7033e3722f27b50df00
7190 F20101109_AABHJN hamey_r_Page_048thm.jpg
dde4c721729af0eb29d84c828db8c7c0
bf9b9c2230693e5229a7e7df19a80f282919a2f1
7438 F20101109_AABHIY hamey_r_Page_024thm.jpg
9d66b19bc26375d9dff635aecad54703
b3f40c33fac0c3f3cdde8b15b7aa6677b13148d7
11647 F20101109_AABGFW hamey_r_Page_123.QC.jpg
78eb13ae1b6640946a32b37738dd080b
b306d2c03115213dea6d790ed2c0c7ecaa3d205a
F20101109_AABGGL hamey_r_Page_049.txt
097f4522e24f557f4ff7873fb299dc22
e10f96e4b81c610b4ce38152056b99f5d1d804d3







PARTICLE DEFORMATION DURING STIRRED MEDIA MILLING


By

RHYE GARRETT HAMEY


















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

2008


































2008 Rhye Garrett Hamey


































My family and fiance Yun Mi Kim, for their encouragement









ACKNOWLEDGMENTS

I would like to acknowledge the friendship and the wisdom of the late Professor Brian

Scarlett, whose instruction and support were the reasons for the pursuit of a Ph. D. Dr. Brian

Scarlett, taught me how to be a better engineer and a better person.

I am grateful for the help of Dr. Hassan El-Shall for his countless hours of guidance

through out this study. Dr. El-Shall was a constant inspiration. I would also like to thank Dr.

Mecholsky, Dr. Fuchs, Dr. Whitney, and Dr Svoronos for serving as committee members and for

their guidance and discussions during this Ph..D. study.

A special thanks goes out to all the friends I have made during this study, Maria

Palazeulos, Scott Brown, Milorad Djomlija, Vijay, Krishna, Stephen Tedeschi, Nate Stevens,

Kerri-Ann Hue, and Dauntel Specht. I would like to acknowledge Marco Verwijs for being a

great friend and roommate for so many years. I would also like to thank Yun Mi Kim for her

love and support.

I would like to acknowledge the financial support of the Particle Engineering Research

Center (PERC) at the University of Florida, The National Science Foundation (NSF) and the

Industrial Partners of the PERC for support of this research. Thanks are extended to the US

Army for their support of the project.









TABLE OF CONTENTS

page

A CK N O W LED G M EN T S ................................................................. ........... ............. .....

L IST O F T A B L E S ...................................................................................................... . 8

LIST OF FIGURES .................................. .. .... ..... ................. 10

A B S T R A C T ................................ ............................................................ 14

CHAPTER

1 INTRODUCTION ............... ................. ........... .............................. 16

Im pact of M killing M etallic Pow ders.......................................................................... ...... 16
Research of Particle D eform ation during M killing .................................................................18
Method for Understanding Particle Deformation during Milling................ ..................18

2 B A C K G R O U N D ...................... ..... ............................ .. .. ......... .... .. ....21

C om m inution E quipm ent............................................................................... ... ............22
C ru sh ers ...................................... ..................................................... 2 3
G rin d ers ...................................... ..................................................... 2 3
U ltrafine G rinders...................... ........ .. .......... .. ............24
Cutting M achines...................... .. .................... 24
Particle Deform ation and Particle Breakage...................................... ......................... 24
Particle D reform action .................................. .. .. .. ...... .. ............24
P article B reakage ... .......... .. .. ..................................................................... 27
Cracks and Defects .......... ......... ...... ...... ..... ................. ......... 28
Particle D eform ation during M killing ............................ .............................. ........... 29
Effect of Milling Parameters on Particle Deformation............................... ...............31
Particle Breakage during M killing ............................................................................... 32
Effect of Milling Parameters on Particle Breakage...................................................33
Su m m ary ...............................................................................................3 5

3 MATERIALS, CHARACTERIZATION AND EXPERIMENTAL PROCEDURE ...........41

M a te ria ls ................................................................................. 4 1
M e ta l P o w d e rs ........................................................................................................... 4 1
M killing M edia.......................................................................................... 4 1
E xperim mental P procedures ........................................................................ ... .......................42
M e d ia M ill ................................................................................................................. 4 2
M killing Procedure .................. ............ ...................... .......... ................ 42
Drying of Aluminum Slurry and Milling Samples.......................................................43
Sample Preparation by Dispersion of Dry Powder.............................................44
Statistical D design of Experim ent ....................................................................... 45









C characterization .......................................... ................ ...................... ........ 45
L ig h t S c atterin g ............................................................................................................... 4 6
M icroscopy and Im age A analysis ............................................. ............................ 47
Electron m icroscopy ......................................................... .... .......... 47
O optical m icroscopy ............................................ ................. ........ 48
O cchio optical particle sizer......................................................... ............... 48
Surface A rea A analysis (B E T) ........................................................................... ...... 48
Inductively Couple Plasma Spectroscopy (ICP) .................................. ...............49
Fourier Transform Infrared Spectroscopy (FTIR)....................................................49
L ight Scattering R esults.......... ................................................................................. ... .... 50
Results of Aluminum Powder ............................... .........................50
R results for A lum inum F lake .................................................................. ...................51
Summary of Characterization and Light Scattering ................................. ...............52
M killing C condition Selection ......................................................................... ....................52
P relim inary M killing Study ................................................................... .... ..................52
Prelim inary M killing R results ................................................. ............................... 53
Establishing M killing Conditions......................................................... .............. 54
M killing m medium .............................................................................................. ........54
M ill loading and grinding aids ........................................ ........................... 55
M killing tim e and tem perature........................................................ ............... 56
M edia properties and rotational rate.................................... ......... ............... 57

4 RESULTS AND D ISCU SSION .................................................. ............................... 72

Experimental Results of Stirred Media Milling................................. ....................... 72
E effect of M ill P aram eters ................................................................... .... ................... 72
K inetic Energy M odel ................................................ .. ............. ......... 73
S tre ss M o d el ................................................................7 5
D reform action R ate ....................................... ........... ...... ........ .. 79
Strain Energy and M killing E efficiency ......................................................................... ...... 81
S tra in E n e rg y ............................................................................. 8 1
M ill E fficien cy .............................. .... ... ............ .. ... .. ........ ................8 3
Empirical Modeling of Particle Deformation through Statistical Design of Experiment ......88
M icrostructure A naly sis........... ....................................................................... ...... ............ 92
S u m m ary ................... ...................9...................3..........

5 CASE STUDY: MATERIAL SELECTION AND DEVELOPMENT FOR INFRARED
OB SCU R AN TS ......................................................... ................ ............ 127

Introdu action ..................................................................................................................... 12 7
Development of New Obscurant Materials ............................................. ...............128
Material Characterization and Obscurant Performance Measurement Method ...................131
Production of Infrared Obscurant ........................................................................ 132
S u m m a ry ..................................................................






6









6 SUMMARY, CONCLUSIONS, AND FUTURE WORK ................................................143

Sum m ary and C onclu sions ..................................................................... ........................143
F utu re W ork .........................................................14 6

L IST O F R E F E R E N C E S ............................................................................. ..........................149

B IO G R A PH IC A L SK E T C H ......................................................................... .. ...................... 154















































7









LIST OF TABLES


Table page

3-1 Statistical data comparison for light scattering and image analysis performed using
the O cchio particle counter. ....................................................................... ...................66

3-2 Stirred media milling equipment variables......................................................................67

3-3 Stirred media milling operating variables................................................... .............. 67

3-4 Experimental parameters used in the hexagonal statistical design of experiment.............67

3-5 M killing conditions for prelim inary study ........................................ ........ ............... 68

4-1 Calculated stress intensity for experiments performed in this study. The table
indicates that it is possible to achieve equivalent energies and different milling
condition s ......... .... .............. ..................................... ........................... 10 1

4-2 Maximum mean particle size of milling experiments. Even though the kinetic energy
of some of the milling experiments are equivalent the maximum particles sizes
d iffer .......... ....... .......... ................ ............ ................................. 10 1

4-3 Force, radius of contact area, and stress acting on the milling media as a function of
media size and rotation rate (calculated using Hertz theory)........................................103

4-4 Number of particles stressed in a single collision between grinding media and the
stress exerted on an individual particles. .............................................. ............... 103

4-5 Stress frequency as calculated from Kwade's model. The frequency is the number of
times milling media collide with each other. ........................................ ............... 103

4-6 The time required to reach the maximum particle diameter. As milling speed
increases the time it takes to obtain the maximum particle diameter decreases............104

4-7 Total number of particle compressions in 60 minutes, as calculated using stress
frequency and number of particles, Equation 4-2 and 4-6. ..............................................104

4-8 Strain energy per particle for deformation study. ................................. .................107

4-9 Total strain energy for deformation study. The 1.5 mm media milling experiments
resulted in the largest amount of stain energy. ..................................... ............... 108

4-10 List of constants for equations used to calculate milling power...............................108

4-11 .Milling efficiency as a function of percent amount of strain the material has
experienced at 1000 rpm and varying media size. The 1.5 mm media reaches over 60
% of the maximum strain while milling at 37% efficiency. The last few percent of
the maximum strain result in the most inefficient milling ................ ....... ...........113









4-12 Experimental design for central composite design 1 .................................................. 114

4-13. Analysis of variance for the central composite design 1 described in table 4-12 ............14

4-14 Experimental design for central composite design 2 ................................ ................118

4-15 Analysis of variance for the central composite design 2 described in table 4-10........... 118

5-1 Experimental parameters used in the statistical design of experiment used to
determ ine m killing param eters. ............................................... .............................. 137

5-2 Experimental parameters used in the full 2 factorial (22) statistical design of
ex p erim en t .............................................................................................13 9









LIST OF FIGURES


Figure page

1-1 Methodology for understanding particle deformation during stirred media milling.........20

2-1 1) Particle stressed by compression or shear, 2) particle stressed by impaction on
rigid surface, and 3) particle stressed in shear flow ......................................................36

2-2 Milling efficiency variations by machine type. ..................................... ............... 36

2-3 T ypes of grinding m ills .................................................................................... ........... ..37

2-4 Diagram of a stirred media mill, energy is supplied to the grinding media by the
rotation of the agitator .................. ................................ ....... .. ........ .... 37

2-5 Face centered unit cell for alum inum .......................................... ........................... 38

2-6 Slip planes and directions for FCC alum inum ........................................ .....................38

2-7 True stress- true strain curve for a material undergoing plastic deformation and strain
hardening ................................. .................. .. ..... ........ ............................39

2-9 Effect of impact ............... ................. .......... ............................. 40

3-1 Particle size distribution of as received H-2 aluminum, along with particle size data,
obtained from C oulter L S 11320. ........................................................... .....................59

3-2 Image of Yttria-stabilized zirconia grinding media obtained form an optical
m microscope, m m in diam eter ............................................................................. .... .... 60

3-3 Union Process stirred media mill use in study ............... ...........................................61

3-4 Phase diagram for carbon dioxide, indicating the region of supercritical fluid
form action ........................................................................ ................... .... 62

3-5 SPI pressure vessel used for supercritical drying of powder samples taken from
m illin g ................... ........................................................... ................ 6 3

3-6 Galia partial vacuum chamber used to disseminate powder............................................63

3-7 Scanning electron microscope image of as received H-2 aluminum powder ..................64

3-8 Differential volume and number distributions for as received H-2 spherical
aluminum powder measured by light scattering. ............. .........................................65

3-9 Volume particle size distribution of milled aluminum flake as measure by light
scatterin g ........ ........ ........................................................................6 5









3-10 Volume particle size distribution of milled aluminum flake as measured by the
Occhio image analysis. ................................... .. .. ... ..... .. ............66

3-11 SEM image of US Bronze "as received" brass flake used in preliminary milling
experim ents. ................................................................................68

3-12 Surface response of figure of merit with respect to media loading and RPM .................69

3-13 Stearic acid a saturated fatty acid found in many animal fats and vegetable oils.............69

3-14 Cumulative undersized versus particle size as at varying milling time for H-2
alum inum .................................................................................70

3-15 Scanning electron microscope images of a) as received aluminum, b) milled for 240
minutes aluminum and c) milled for 600 minutes aluminum.........................................71

4-1 Volume percent particle size distributions for experiments using 1.0 mm grinding
media at 2000 rpm. Bimodality in the particle size distribution is indicative of
particle deformation and particle breakage occurring simultaneously. ..........................95

4-2 Volume percent particle size distributions for the experiments with 1.5 mm grinding
media at 1000 rpm, a visible shift in the particle size distribution to larger sizes is
seen due to particle deform ation. .............................................. .............................. 96

4-3 Maximum mean particle size achieved for each milling experiment versus media
size, the 1.5 mm milling media produced the largest amount of deformation...................97

4-4 Particle deformation versus milling time at a milling speed of 1000 rpm for varying
grinding m edia sizes. ...................................... .. ....... .......... .... 98

4-5 Particle deformation versus milling time at a milling speed of 1500 rpm for varying
grinding m edia sizes.. ............................ .......... .. ....... .......... .. .. 99

4-6 Particle deformation versus milling time at a milling speed of 2000 rpm for varying
grinding m edia sizes.. .................... ...................... .. .. ...................... 100

4-7 Mean particle size versus milling time as a function of rotation rate .............................101

4-8 Contact between colliding elastic spheres. Aluminum particles are caught in the
contact area betw een the m edia. ........................................................... .....................102

4-9 Diagram of the change in dimensions of particle and the measurements used to
calculate stress and strain ............. ........................... ......... ..... .............. 104

4-10 Stress-strain behavior for experiments performed at 1000 rpm............... ...............105

4-11 Stress-strain behavior for experiments performed at 1500 rpm....................................106

4-12 Stress-strain behavior for experiments performed at 2000 rpm....................................107









4-13 A plot of milling efficiency as a function of milling time at 1000 rpm, for all media
siz e s ................... ............................................................ ................ 1 0 9

4-14 A plot of milling efficiency as a function of milling time at 1500 rpm, for all media
siz e s ................... ............................................................ ................ 1 1 0

4-15 A plot of milling efficiency as a function of milling time at 2000 rpm, for all media
siz e s ................... ............................................................. ................. 1 1 1

4-16 A plot of milling efficiency as a function of milling time for the 1.5 mm at varying
rotation n al rates ................................................................................................ .... 1 12

4-17 A plot of milling efficiency at failure, right side of figure 4-13 used for comparison
of efficiencies as a function of rotational rate............................................ ..........113

4-18 Contour plot for central composite design 1, a minimum strain of 2.17 can be seen at
approxim ately 0.9 m m milling m edia ................................................................. ....... 115

4-19 Surface response for central composite design 1, a trough exists at the 1.0 mm media
siz e .................. ............................................................................ 1 1 6

4-20 Standard error associated with central composite design 1, error increases at the
lim its of the study .................................................... ........ ...... 117

4-21 Contour plot for central composite design 2, shows a maximum at approximately the
1.6 m m m edia......................................................................................... .. .. .... 119

4-22 Surface response for central composite design 2, a peak can be seen at intermediate
m e d ia siz e .......................................................................... 12 0

4-23 A plot of the interaction between the media size and rotation rate, it can be seen that
there is only a slight interaction at low ........................................ ........................ 121

4-24 Standard error associated with central composite design 2. .........................................122

4-25 Transmission electron microscope image of "as received" aluminum particle.............123

4-26 Diffraction pattern obtained from particle in Figure 4-20. ............................................124

4-27 Transmission electron microscope image of sectioned flake after 4 hours of milling. ...125

4-28 Single crystalline diffraction pattern of sectioned flake after 4 hour sof milling ..........126

5-1 A chart of relevant material properties of materials for use as an infrared obscurant.
High values in each property are preferred for candidate materials .............................136

5-2 The performance of various milled metals, the copper has the highest performance to
due to the properties being the highest in all the desired area. ......................................137









5-3 The results of the initial design of experiment, indicating that higher media loading
increase the perform ance of the m material ............................................... .................. 138

5-4 Plot of FOM for as received and ground brass at specific milling condition of 480ml
media volume and 700 rpm rotational speed ........................................................139

5-5 The results of the affect of varying media size and rotation rate on the performance
of a milled brass flake........................ .......... ............ 140

5-6 Differential volume particle size distribution at increasing milling times.......................141

5-7 Figure of merit versus IR wavelength for increasing milling times ............................141

5-8 Mean particle size as a function of milling time for different density milling media,
zirconia (6 g/cc), stainless steel (9.8 g/cc), and tungsten carbide (16 g/cc)..................142









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

PARTICLE DEFORMATION DURING STIRRED MEDIA MILLING
By

Rhye Garrett Hamey

August 2008

Chair: Hassan El-Shall
Major: Materials Science and Engineering

Production of high aspect ratio metal flakes is an important part of the paint and coating

industry. The United States Army also uses high aspect ratio metal flakes of a specific dimension

in obscurant clouds to attenuate infrared radiation. The most common method for their

production is by milling a metal powder. Ductile metal particles are initially flattened in the

process increasing the aspect ratio. As the process continues, coldwelding of metal flakes can

take place increasing the particle size and decreasing the aspect ratio. Extended milling times

may also result in fracture leading to a further decrease in the particle size and aspect ratio. Both

the coldwelding of the particles and the breakage of the particles are ultimately detrimental to the

materials performance. This study utilized characterization techniques, such as, light scattering

and image analysis to determine the change in particle size as a function of milling time and

parameters.

This study proved that a fundamental relationship between the milling parameters and

particle deformation could be established by using Hertz's theory to calculate the stress acting on

the aluminum particles. The study also demonstrated a method by which milling efficiency could

be calculated, based on the amount of energy required to cause particle deformation. The study

found that the particle deformation process could be an energy efficient process at short milling

times with milling efficiency as high as 80%. Finally, statistical design of experiment was used









to obtain a model that related particle deformation to milling parameters, such as, rotation rate

and milling media size.









CHAPTER 1
INTRODUCTION

In today's market there is a need to develop new and improved products at an increasing

rate. Particle deformation by milling is important to many areas of product development, such as,

powder metallurgy (P/M), paints and coatings, and military applications. Understanding and

predicting particle deformation during milling, will lead to more rapid development of products

in which these materials are used.

Impact of Milling Metallic Powders

Generally, milling is used to reduce the size of a material or for the processing of brittle

materials. However, large quantities of ductile metals used in the powder metallurgy industry

(P/M) are processed through milling [1]. Powder metallurgy is estimated to be a $1.8 billion

industry annually with an estimated 70 % spent in the automotive industry [2]. Below is a list of

reasons for the milling of materials in the P/M industry [1]:

* Particle size reduction (comminution or grinding) for sintering

* Particle size growth

* Shape change (flaking)

* Agglomeration

* Solid-state alloying (mechanical alloying)

* Solid-state blending (incomplete alloying)

* Modifying, changing, or altering properties of a material (density, flowability, or work
hardening)

* Mixing or blending of two or more materials or mixed phases

* Nonequilibrium processing of metastable phases such as amorphous alloys, extended solid
solutions, and nanocrystalline structures









During the processing and subsequent stressing of metallic powders, several phenomena

may occur; the metal can fracture, deform, coldwork, coldweld, or transform into other

polymorphs [3]. Coldwelding is used to mechanically alloy (MA) and refine the microstructure

of immiscible metals [4]. Coldworking is used to strengthen metallic particles. This study will

focus on deformation of metallic particles during milling. This technique is used to make

metallic flakes for paints, coating, and obscurants [5-9]. The metallic flakes in paints give what is

termed a "metallic look" which means the randomly oriented flakes in the paint will give a

"shiny" appearance. Radiation shielding in devices such as cell phones will use metallic particles

to form a conductive coating. The US Army uses metallic flakes in smoke bombs to obscure

infrared light (IR). The conductive metallic particles are capable of attenuating the IR signal

which provides a shield against missiles and IR imaging devices.

The most common metallic pigment is aluminum, which when milled deforms into flakes

with highly reflective surfaces. In order for the metallic flakes to be effective at scattering and

reflecting light, the particle size, size distribution, particle shape, and particle morphology must

be controlled. The particle size and size distribution affect the optical reflectance and light

scattering properties of the flake. If the particle size distribution of these paints is broad the

reflectiveness will be diminished and the paint will appear dull. The particle morphology and

surface roughness also affect the optical reflectance i.e. the smoother the surface the more

reflective the pigment. An additional difficulty to this processing method is that most of the

aluminum pigments are produced in batch milling processes and there exists large batch to batch

variations due to lack of process knowledge [10]. In the end, these variations lead to additional

processing costs from classification.









Research of Particle Deformation during Milling

Most milling studies focus on particle breakage with limited research on particle

deformation. The particle deformation studies have focused on the mechanical alloying process

and do not investigate optimizing the flake dimensions, which are necessary for paints and

coatings. Zoz showed how the dimensions of flakes are affected by the material hardness[l 1].

Several researchers have investigated the change in the microstructure and material properties of

metals during milling [12-14]. None of these studies has shown the effect of milling parameters

on particle deformation. The research to date does not provide any methodology for predicting

the magnitude or extent of particle deformation during milling. In order to improve the

production and quality of the products it is important to understand the effects of milling

parameters on the materials being processed. Even though particle breakage during milling has

been studied extensively, there is very little knowledge of the particle deformation process

during milling. This study will attempt to apply some of the knowledge gained from particle

breakage during milling to the deformation process and to fill the gaps in the theories between

the two processes.

Breakage during milling has been studied at nearly every length scale [15-19]. Numerous

researchers have modeled milling kinetics and energetic for particle breakage [20-22]. Others

have used population balance modeling to predict particle size during milling [19, 23]. The

response of the product material to nearly every milling parameter has been investigated for

particle breakage. Wear on equipment and process optimization studies have also been

researched [24, 25].

Method for Understanding Particle Deformation during Milling

The goal of this study is to understand and predict particle deformation during stirred

media milling. Figure 1-1 shows the methodology that was used to accomplish this goal. The









study investigated the affects of mill speed and media size on the particle size and morphology of

an aluminum powder. The study shows how milling breakage models fail to describe particle

deformation during milling. New models were derived from the stress-strain behavior of

aluminum during milling to describe particle deformation. Hertz theory was used to calculate the

stress acting on a particle during. The stress acting on a particle and the affect it had on particle

deformation was then determined. Statistical analysis was used to determine the magnitude of the

interaction between milling parameters. Design of experiment was used to construct a model

that relates milling parameters to particle deformation. From these studies a more comprehensive

understanding of particle deformation during milling was achieved. This may ultimately lead to a

method to better predict and optimize milling processes where the deformation of metallic flakes

is desired.






























Figure 1-1. Methodology for understanding particle deformation during stirred media milling.









CHAPTER 2
BACKGROUND

Comminution is arguably the oldest material processing technique. Earlier civilizations

employed simple mortars and pestles to reduce the size of agricultural and medicinal products.

Even though there have been significant changes in technology, these simple devices that were

employed thousands of years ago can still be found in nearly every lab today. A simple definition

of comminution is the reduction in particle size through the application of an applied load. In

order to study and improve a milling process it is important to understand how the load is applied

and to analyze the response of the material to that load. Rumpf suggested four modes by which

stress can cause particle breakage [26]:

1. Compression-shear stressing
2. Impact stressing
3. Stressing in shear flow
4. Stressing by other methods of energy transmission (electrical, chemical, or heat)

Figure 2.1 is a representation of the modes of stressing a material as indicated by Rumpf.

The first two modes are caused by contact forces and are important for milling. Mode 3 is

important for the dispersion of materials under shear, and only exerts enough force to break apart

weak agglomerates [26]. The fourth mode is not applicable to milling where mechanical

stressing is important. There are several types of machines designed to apply stress to a material

by modes 1 and 2.

The beginning of this chapter will describe the types of milling equipment used today. A

review of literature on single and multi particle breakage will be presented. Followed by a

discussion of the response of a material to the load applied during milling. This information will

be used as a basis for this study and it will shed light on the deformation process during milling,

which has not been studied as thoroughly. This chapter will conclude by discussing material

deformation and reviewing deformation during milling processes. Ultimately, this study will









compile what has already been done to study particle deformation during milling and use that

knowledge, combined with particle breakage and further experimentation, to more accurately

develop the understanding of particle deformation during milling.

Comminution Equipment

Comminution is a diverse process, found in nearly every industry. Therefore there are

numerous types of milling equipment. The equipment must be able to operate in rigorous, high

throughput environments of the mineral processing industry and in the ultra pure, contamination

free pharmaceutical industry. McCabe, Smith and Harriott identified four classes of milling

equipment each with subclasses [27]:

A. Crushers (coarse and fine)
a. Jaw crushers
b. Gyratory crushers
c. Crushing roll
B. Grinders (intermediate and fine)
a. Hammer mills, impactors
b. Rolling-compression mills
i. Bowl mills
ii. Roller mills
c. Attrition mills
d. Tumbling mills
i. Rod mills
ii. Ball mills; pebble mills
iii. Tube mills; compartment mills
C. Ultrafine grinders
a. Hammer mills with internal classification
b. Fluid-energy mills
c. Agitated mills
D. Cutting machines









a. Knife cutters; dicers; slitters
The differences in equipment are due to the mode in which stress is applied to the material

and the final achievable particle size. Some of the equipment may work in clean environments

and process small quantities of products while other may process large amounts of material in

wet and dry environments. Schonert has shown that some of these machines can operate at high

energy efficiencies >90% while others operations are a fraction of a percent efficient. Most of

this inefficiency is associated with the increase in energy required to achieve smaller particle

size. Examples of the efficiency of select mills can be seen in Figure 2-2, obtained from

tabulated data given by Prasher for the energy consumption associated with milling to a specific

size [28].

Crushers

Jaw crushers, gyratory crushers and crushing roll mills are used to process large quantities

of materials. They are often used in the mineral processing and cement industries to reduce the

size of rock and ore for further processing.

Grinders

Grinders are one of the most diverse sets of milling equipment. They are most often used

to reduce large aggregates to a powder. Grinding mills can supply a load to a material by

impaction, attrition, or compression. Hammer mills apply a load via impaction by dropping a

weight repeatedly on a bed of material. Attrition mills wear away a material by the breakage of

small fragments from the surface. Roll mills and ball mills apply a load through compression by

trapping material between colliding media or in between two heavy rollers. Figure 2-3 is a

diagram of a ball mill and a roller mill. Ball mills have been the subject of a considerable amount

of research because they can be used in wet and dry environments and their large capacities. Ball









mills operate by rotating a drum containing hard dense balls and product material. The balls fall

from a height equal to the diameter of the drum and compress the material.

Ultrafine Grinders

Ultrafine grinders have recently become very popular due to there ability to produce

nanoparticles. The most popular ultrafine grinders are fluid energy mills such as jet mills and

agitated mills such as stirred media mills. Jet mills stress a material by entraining particles in a

gas stream and impacting the particles on a hard surface or against each other. Jet mills are used

extensively in the pharmaceutical industry due to their ability to produce fine particles without

the wear of mechanical parts. Another class of mills used to produce fine particles, are the stirred

media mills, which are similar to ball mills except they contain an agitator, which supplies the

necessary energy instead of rotation of the vessel. The agitator allows the media to collide with a

much higher force than is possible in the ball mill. Due to their importance in this study, stirred

media mills and particle breakage during stirred media mills will be discussed in more detail

later. Figure 2-4, is a diagram of a stirred media mill. Products in a stirred media mill are often

ground in a wet environment to allow for easier stabilization of the product.

Cutting Machines

Cutting machines are used to reduce the size of resilient materials that are not easily

fractured [27]. Some examples of materials that are processed using cutting machines are tire

rubber for recycling, paper products, and plant materials. Cutting machines apply a shear stress

to materials, which causes them to fail or break.

Particle Deformation and Particle Breakage

Particle Deformation

Plastic deformation in a metal occurs when the material is stressed beyond its yield stress;

the atomic bonds inside a material are broken and then reformed as planes of atoms are moved









past each other. Plastic deformation results in a percent change in the shape of the object being

stressed. Plastic deformation is affected by the stress (c), strain rate (dc/dt), temperature (T), and

the microstructure of the material [29, 30]. This study focuses on the deformation of aluminum

particles when subjected to a compressive stress.

Aluminum is a crystalline material with a face center cubic (FCC) structure with atoms at

each corner and in the center of each face of a cube. Figure 2-5 is a sketch of a FCC unit cell. For

such a structure, deformation often occurs by slip, the process of plastic deformation by

dislocation motion, along the preferred crystal planes. The 111 } plane for aluminum is the most

atomically dense plane and it is in this plane, along the <110> direction, that slip occurs [31].

There are 4 slip planes for aluminum and 3 slip directions this means that there are 12 total slip

systems for dislocation motion to occur in. Figure 2-6 show the highlighted { 111} plane and the

<110> direction is denoted by the arrows beneath the highlighted areas.

When dislocations encounter each other and a strengthening of the material will occur

resulting in a loss of ductility. This strengthening is called work hardening, cold working, or

stain hardening; and is due to the increase in dislocation density and decrease in the allowable

dislocation motion. Equation 2-1 is the work hardening equation, ,y is the true yield stress, E is

the true strain of the material, Kw is the strengthening or work hardening coefficient, and n is the

work hardening exponent. In the equation the stress and strain are true stress and true strain. The

exponent and coefficient for the work hardening equation are generally obtained experimentally.

Many researchers have calculated the exponent for numerous materials under numerous

conditions [32-34].

ya = Kw 2-1









According to this relationship, higher stresses are required to obtain greater amounts of

deformation. The material will eventually work harden to the extent that no additional

deformation may be achieved and the material will ultimately fail.

In a milling process, a material will undergo multiple loading and unloading cycles. This

behavior can be characterized by a stress strain curve. Figure 2-7 is a generic representation of a

stress-strain curve for plastic deformation followed by failure. When the material is unloaded

after experiencing plastic deformation, it will remain in the deformed state. The area under the

elastic region of the curve is called resilience. It is the strain energy due to elastic deformation of

the material. The total area under the curve until fracture is called toughness. This is the same

toughness discussed earlier for particle breakage. The area under the curve represents the amount

of energy the material can absorb before it will fracture, or the strain energy due to plastic

deformation. It should be noted that the area under the linear part of the curve is the energy due

to elastic deformation.

Strain rate, d cdt, is the rate at which a material is deformed. Schonert referred to strain

rate as stress velocity and theorized that it could affect material breakage. The strain rate also

affects deformation. Several researchers have investigated the effects of strain rate on the

deformation of materials [35-37]. Tome and Canova investigated the stress strain response of

aluminum and copper at different strain rates[37]. They showed that over the strain rate range

tested, copper did not appear to have strain rate dependence. However, aluminum did have strain

rate dependence; lower strain rates resulted in less deformation before failure. Nieh observed

superplastic-like behavior, also known as extended ductility, of aluminum composites at very

high strain rates. Meyers has also shown that at very high strain rates it is possible for a material

to even recrystallize and form new grains [38]. Hines also observed this result in copper [39]. In









order to achieve a greater understanding of particle deformation during milling it is important to

review particle breakage and to correlate these results to deformation.

Particle Breakage

Numerous authors have studied single particle and multi particle breakage [40-43]. They

have found that particles can break in different ways depending on the method in which the load

is applied. In a stirred media mill, load is applied to particles trapped between colliding grinding

media, media and the wall of the vessel, and between the stirrer and media. Therefore, particle

breakage is a complex phenomenon with numerous stages to study. Schonert extensively studied

breakage, crack propagation, and crack velocity for many particulate materials and found that the

strain rate or stressing velocity can influence material failure [44]. Polymer materials, for

example, are strongly affected by strain rate. Metals also have been shown to have some strain

rate dependence, however, the effect is much less than that of polymers [44].

There are three modes by which a material can fracture, which are a function of the

material as well as the stress applied to the material.

(1) Opening mode that corresponds to an applied tensile load
(2) Sliding mode where the force is applied in the plane of the defect or crack
(3) Tearing mode where a force is applied in a direction out of the plane of the crack or defect

For most particles, breakage is considered to take place in the first mode of failure. This

may be due to artifacts in the techniques developed to measure particle breakage. For this reason,

most techniques for measuring particle failure involve stressing the particle in tension. If a

normal load is applied to a particle a load equal in magnitude is acting internally in the direction

perpendicular to the load and can then be considered to be a tensile load. Tensile failure is more

likely than compressive failure for many materials.

Particle breakage can take place through many different mechanisms of loading. These

mechanisms of loading as well as the material being stressed can affect how the particle will









break. The mechanisms of breakage were alluded to earlier when discussing the types of milling

machines and now will be discussed in more detail. Rumpf suggested there are three ways by

which a particle can break, which are fracture, cleavage or attrition [26]. During fracture a

mother particle will break into daughter particles or fragments that have a wider size distribution.

During cleavage a mother particle will break into smaller daughter particles of roughly the same

sizes.

Attrition takes place when a mother particle has surface asperities removed from the

particle. This leads to much smaller daughter particles. Figure 2-8 represent the three different

mechanisms of particle breakage. If a brittle material experiences a normal load such as

impaction in a milling process it would undergo a fracture or attrition mechanism of breakage. If

a ductile material experiences a shear force, it is most likely to fail by a cleavage mechanism

[45].

Cracks and Defects

Understanding crack formation and propagation is of fundamental importance to the study

of particle breakage. Particle breakage is preceded by crack growth, which is preceded by crack

formation. Crack formation is enhanced by local defects in a material. Defects are separated into

four classes, which correspond to their dimensions.

(1) Point defects (vacancies, interstitials, or substitution)
(2) Line dislocations (screw or edge)
(3) Interfacial
(4) Volume

The strength of a material is related to the number, size and types of defects present. Most

materials are only a fraction as strong as the theoretical strength would predict due to the

presence of defects. A method by which cracks are initiated is when the applied stress exceeds

the strength of the material. Following crack initiation is crack growth, which leads to particle









breakage. Sch6nert extensively studied crack growth and found that crack velocity can vary

widely depending on the material, environment, and mode of stressing [46]. Sch6nert also

studied particle breakage during milling, however few researchers have studied deformation

during milling [47].

Particle Deformation during Milling

A significant amount of work has gone into investigating particle deformation during

milling, however, that work has focused on the mechanical alloying process[l]. This process is

important because it is used to make alloys that can be made in no other way. Beside the

formation of novel alloys, mechanical alloying can be used to refine the microstructure of the

materials to increase their strength. Mechanical alloying and microstructure refinement involves

adding a metallic powder(s), fatty acid/dispersant, and grinding media to a ball mill. The mixture

is then milled until an equilibrium particle size is reached [4]. The particles generally produced

in this process are semi-spherical agglomerates.

Benjamin is the pioneering researcher of mechanical alloying. The author was the first to

form an indium alloy [48]. Since then numerous authors have studied mechanical alloying (MA)

[49-53]. Most of the studies on MA have focused either on alloying or on refinement of

microstructure. Weeber studied the effect of different ball milling operations on the final product

and found that the milling equipment can affect the final product [52]. The author found

microstructure can vary using different types of ball mills. Gilman investigated the effect of

additives on the alloying process and found that if too little quantity of dispersant was added that

the metal would coldweld together and not effectively alloy[4].

Even though mechanical alloying and the production of metallic flakes involve similar

equipment and similar physics, the final product for flake production is very different. The

desired product of the flake making process is a powder with a large major dimension and a









narrow particle size distribution. It is not possible to run a flaking process to an equilibrium

particle size, as is the case with MA. Further milling will cause the flakes to break, no longer

scatter light effectively, and perform poorly as metallic effect pigments. The major differences

between the MA and flaking process are the use of dispersant and the grinding time. Many

flaking processes mill in a solvent to prevent the metal particle from coldwelding together and

the processes are run for shorter lengths of time to limit breakage.

The production of metallic flakes for use as pigments has been ongoing for over 50 years.

Some of the first available works are patents for devises made to produce metallic flakes [54].

More recently metallic flakes have found increasing use in other applications such as conductive

and electromagnetic resistant coatings and infrared obscurants [7, 8]. It is important for flakes to

have a large major dimension in order to function in the specific part of the electromagnetic

spectrum desired for these applications. Breakage during milling of these metallic particles

ultimately leads to poorly performing obscurants and paints. It is necessary to understand the

relationship between particle deformation and milling.

Moshksar has shown that when milling aluminum powder that the aspect ratio of the

material [major dimension divided by minimum dimension (d/t)] increases to some maximum,

then begins to decrease at longer milling times [55]. The decrease is due to particle breakage and

a subsequent reduction in aspect ratio. The results suggest that an ideal product may be produced

by monitoring the aspect ratio as a function of milling time. The author also found that the cold

welding and agglomeration of the powder particles was hindered by the fracture of the flakes at

longer milling times. Zoz showed that the maximum value of the d/t ratio varies with the

hardness of the metal being milled [56]. The lesser the hardness of the metal and the greater the

d/t ratio, the more likely the material will form flakes. Material selection and strain rate can









significantly affect whether the material breaks or deforms. It may also be possible to relate

material properties such as hardness and strain rate to the aspect ratio. This relationship would

enable the development of a model that relates material properties, and milling parameters to

particle deformation.

Effect of Milling Parameters on Particle Deformation

The deformation process has yet to be studied as extensively as the breakage process; the

studies that have been performed are mainly focused on a specific application. Much of the

research that has been performed is directed at understanding the microstructure development

during mechanical alloying (MA) [4]. For the paint industry, changes in particle morphology are

important and not the change in the microstructure of the material [57]. Few researchers have

investigated the milling of metals for use in the paint industry [5, 56, 58]. Zoz's work was

discussed above [11]. Sung studied the affect of particle morphology on the reflectance and

quality of the paint.

Hashimoto studied the effect of milling media diameter on the total energy consumption as

well as the energy per impact on a metal powder in a simple vibratory mill [14]. From this, the

average energy per collision was calculated. The author found that the total energy consumption

was higher for larger media. The author also found that as media loading increased, the average

energy per collision is reduced due to ball-ball interactions. Kwade's models of stress frequency

and stress intensity neglect ball-ball interactions. Hashimoto measured the change in the

hardness of the metal with milling time, the work hardening of the metal, and found that the

increase in the hardness of the powder was greater with larger media. This increase was

attributed to cold working of the material during the milling process. Riffel determined that an

optimum media size existed when mechanically alloying (MA) Mg2Si compounds. The author

determined that if the media was too small it did not provide enough kinetic energy to alloy or









reduce the size of the material [59]. Riffel's results for MA also agree with the Kwade's

predictions of an optimum stress value (milling conditions) for particle breakage. Kwade and

Riffel's result suggest that it may be possible to use particle breakage equations to describe the

particle deformation process.

Rodriguez milled aluminum in a stirred ball mill, and measured the hardness as a function

of milling time [60]. The researcher found an initial increase in Vickers hardness which reached

a maximum at a value of 130 kg/mm2. The increase in hardness measured by Rodriguez agrees

with the increase also observed by Hashimoto. Huang also measures an increase in hardness with

milling time in copper [12]. Huang used the change in the diffraction pattern measured from the

high-resolution transmission electron microscope (HRTEM) to determine the lattice strain and

measured a 0.2% increase in lattice strain. Hwang also used powder x-ray diffraction data

(XRD) to investigate the effect of milling time on lattice strain and determined that it

increased[ 13]. Huang and Hwang's results agree and provide other methods of characterizing the

work hardening of a material during milling.

Particle Breakage during Milling

Prasher gives a chronology of proposed laws relating energy for breakage to simple

parameters such as changes in surface area or change in particle size [28]. Two of the most

notable relationships are Rittinger's and Kick's Law. Rittinger states that, energy E, to break

particles is equal to the new surface area times a constant as described by Equation 2.1.


E c 1- = K1As 2-1


The terms xl and x2 describe the particle diameter of the feed and product material

respectively, K1 is a constant and As is the change in surface area. Kick's Law, Equation 2.2,









finds the energy to break a particle is equal to a constant, K2, times the log of the starting particle

size divided by the ending particle size.


E= K log( x1 2-2


These relationships apply to a limited number of cases with precise particle sizes and specific

materials and have not been shown to apply to processes where particle deformation is desired

[28].

Figure 2-9 shows a comparison of brittle materials and ductile materials trapped between

impinging grinding media. Brittle particles are compressed and fractured by the media, whereas

ductile particles are plastically deformed and flattened. Both processes represent an increase in

overall surface area, however, during particle deformation the volume of the particle is

conserved.

Effect of Milling Parameters on Particle Breakage

Choosing the proper milling parameters can have a significant effect on the breakage rate,

energy consumption, and efficiency of the milling process. Grinding media density and size, for

example, directly affects the magnitude of the load on the particles, and thus can be expected to

alter milling. Many authors have studied the effects of milling parameters on particle breakage in

a stirred media mill [21, 25, 61, 62]. Kwade formulated a simple equation (eq. 2-3) to

mathematically relate media size, media density, and rotational rate to the stress intensity.

Kwade's stress intensity term is merely a representation of the kinetic energy of the media and is

not a representation of the actual stress acting on the particles. Stress intensity (SI) is directly

proportional to circumferential velocity v, media size Dp, and media density pp less the fluid

density pf through Equation 2-3 [63]:









SI = D p pf)u2 2-3

Kwade then showed that at a specific energy input an optimum in stress intensity existed

that provided minimum particles size for a minimum energy input. The optimum energy input

for producing a maximum fineness exists between a stress intensity of 0.01 Nm and 0.1 Nm.

Kwade also derived an expression that relates the stress frequency (SF) to media size [63]. The

stress frequency is a representation of the number of media-particles interactions, i.e. the number

of times a particle is contacted by the media, and described by Equation 2-4.

(D 2
SF = c), D T 2-4
DP

Here, T is milling time, Dp is the diameter of grinding media, D, is diameter of the stirrer, and os

is angular velocity. Kwade showed that the greatest stress frequency results from the smallest

media size. Large stress frequency also results in the smallest particle size. The author was able

to provide a method for optimizing and even scaling a milling process. However, these results

have only been shown to apply to a system where particle size reduction is desired. The effect of

stressing velocity on the deformation of materials in a milling system has not been studied.

Stressing velocity and strain rate in a mill can be assumed to be equivalent terms as previously

discussed by Schonert [44]. Using Kwade's model of SI it is now possible to relate specific

milling parameters (media size, media density, and rotational rate) to the strain rate. This would

be the first step in obtaining a realistic model to describe the deformation process in a stirred

media mill.

Eskin et al. developed a more sophisticated approach to modeling the energetic of a

milling process. The authors calculated granular temperature, and used it to represent the

velocity of the grinding media in their process [20]. The authors then calculated the force, and









stress of the grinding media based on Hertz's theory and rigid body impact. Equation 2-5

demonstrates how force, F, is related to media size, R, granular temperature(velocity), 0, and

materials properties, such as, Poisson's ratio, r, density, p, and modulus of elasticity, Y [20].

2

F= 1.96 (p)5(R)2 () 2-5


Eskin derived a relationship determining the number of compressions a particle would

experience as a function of media size. This relationship is similar to Kwade's stress frequency

term. Eskin concluded by calculating a milling efficiency term based on the energy input into the

mill. The work by Eskin was found to be important to this study and the equation and derivations

of his work will be revisited in more detail in Chapter 4.

Summary

In summary, Kwade has provided models that describe the stress intensity and stress frequency

in a stirred media milling process for particle breakage. These models may also be valid for

describing the stress intensity in a milling process where deformation is desired. It is known that

some metals have a strain rate dependent behavior and that greater deformation may be achieved

from a lower strain rate. Zoz showed that the amount of deformation a material can undergo in a

milling experiment is also related to the hardness of the material. Researchers have also shown

that the extent of work hardening varies with media size, concentration and milling time. The

researchers have provided three methods by which work hardening can be characterized,

including indentation, XRD, and HRTEM. From these studies, it is now possible to begin to

construct models that can predict material deformation based on the hardness of the material,

deformation, milling time, media size, media density and rotational rate of the stirred media mill.









-k


4-


SU ---
Figure 2-1. 1) Particle stressed by compression or shear, 2) particle stressed by impaction on
rigid surface, and 3) particle stressed in shear flow.


Fluid Energy 2%

Pin Mill 5%


Free Crusher 100%

Roll Crusher 80%


Ball Mill 8%

Ball Roce Mill 13%
Swing Hammer Mill 22%


30 40 50 60 70 80 90 100


Figure 2-2. Milling efficiency variations by machine type.


0 10








B)

0 /PO





Ball Mill Roll mill

Figure 2-3. Types of grinding mills, A) ball mill and B) roll mill.








Stirred media
Figure 2-4.Diagram of a stirred media mill, energy is supplied to the grinding media by the
rotation of the agitator.






















FCC unit cell

Figure 2-5. Face centered unit cell for aluminum.


X

Figure 2-6. Slip planes and directions for FCC aluminum.








Stress


Plastic reion
GUTS



1 .[Unloading

Elastic region









Strain

Figure 2-7. True stress- true strain curve for a material undergoing plastic deformation and strain
hardening.




















Cleavage


Attrition


Fracture


Starting particle size distribution


Attrition


Increasing relative particle size ,

Figure 2-8. Relative particle size distributions for attrition, cleavage, and fracture, respectively.


a)b)
Figure 2-9. Effect of impact (a) brittle single particle, and (b) ductile single spherical particle.









CHAPTER 3
MATERIALS, CHARACTERIZATION AND EXPERIMENTAL PROCEDURE

The goal of the study was to understand and predict particle deformation during milling. In

order to achieve this goal it was necessary to develop experimental procedures and

characterization techniques for performing and assessing particle deformation. This chapter

describes the milling equipment and procedures used in this study. Characterization techniques

and validation of the results are discussed in this chapter. In addition, the chapter describes the

statistical analysis techniques used for predicting particle deformation during stirred media

milling. The chapter concludes with identification of the milling parameters used in this study.

Materials

Metal Powders

The aluminum powder (product name H-2) used in this study was purchased from Valimet

Inc, which is generally used in the production of metal effect paints and aluminum based

explosives. Valimet manufactures the aluminum by atomization. Molten aluminum is sprayed

through a high pressure nozzle and solidifies when it is exposed to argon atmosphere forming

semi-spherical aluminum particles. The aluminum is then classified to obtain the desired size

distribution. The powder has a purity of 99.7 weight percent aluminum and a median particle

size of 3.2 jtm as measured by the manufacture. In house measurements of the particle size using

light scattering determined the median size to be 3.8 jtm, and a mean of 4.2 jtm. Figure 3-1

shows the size distribution of the "as received" H-2 aluminum as measured by the Coulter

LS13320 light scattering particle sizer.

Milling Media

The milling media used in this study was yttria-stabilized zirconia, purchased from Advanced

Materials Inc. in sizes of 0.5, 1.0, 1.5, and 2.0 mm. The density of the media is 6.0 g/cm3. Figure









3-2 is an image of the 1.0 mm grinding media used in this study. In order to maintain a consistent

size distribution the media was sieved after each experiment to remove fines and fragments

caused by fracture and wear [1].

Experimental Procedures

Media Mill

The mill used in this study was a bench top Union Process attrition mill as seen in Figure 3-3.

The mills operational speed is between 500-5500 rpm. Numerous types of milling vessels and

stirrers can be equipped on this model. This study used a 750 ml stainless steel milling vessel and

a 4 bar stainless steel agitator, the bars of the agitator are approximately 60 mm in length. It is

important to note that the length of the agitator arm and the rotational rate was used to calculate

the tip speed of the mill and ultimately the kinetic energy of the grinding media. The vessel is 90

mm in diameter and 90 mm in height. Figure 3-3 is a picture of the milling vessel and the

agitator used in the study. The mill is attached to a recirculation reservoir that maintains the

temperature of the mill contents at 25C. The mill is attached to a digital controller, which

provides speed and power control over the mill.

Milling Procedure

The mill was assembled according the specification provided in the mill manual. To assure

experimental continuity, the volume of material used and the distance between the bottom of the

agitator and the bottom of the milling vessel (1/4 in.) was maintained in all experiments. The

contents and milling sequence were performed in the following order:

1. 25 grams of H-2 aluminum powder
2. 5 grams of stearic acid
3. 300 ml of isopropyl alcohol
4. Initial dispersion
5. 400 ml of grinding media
6. Start milling experiment (sample at appropriate times)









The milling medium used in this study was isopropanol with 5g of stearic acid dispersant. A

more detailed description of these materials will come later in the chapter. Prior to the loading of

milling media, the mill was run for 5 min at 500 rpm to disperse the metal powder in the medium

(liquid suspension). Following the 5 minutes of dispersion the mill was ramped up to the desired

milling rotational rate. It was then shut off. The milling experiment was then restarted when the

milling media was added. Samples were taken from the mill at times of 1, 3, 5, 8, 12, 15, 30 45,

and 60 minutes then at 1.5, 2, 4, 6, 8, and 10 hours. In order to reduce the effect of the change in

volume on milling kinetics the samples removed from the mill were less than 1.0 ml. Samples

were diluted with 20 ml of isopropyl alcohol, sealed and then stored for either drying or

characterization.

At the end of each experiment, the mill was stopped and dissembled. The contents of the

mill were pored through two sieves, one of sufficient size (i.e. 1.0 mm mesh for 1.5 mm media)

to catch grinding media and another of 90 |tm to catch fragments of grinding media. After the

slurry was separated form grinding media it was placed in a Nalgen bottle for storage and later

drying. Every part of the mill that was exposed to the aluminum slurry was rigorously cleaned

with soap and water and allowed to dry in an oven before performing another experiment. The

grinding media caught in the sieve was placed in a 1000 ml Erlenmeyer flask and washed with a

3 M hydrochloric acid (HC1) solution. The contents of the flask was then emptied into a sieve

and placed in a sonic water bath. The sonicator was flushed repeatedly with water to remove

HC1. The media was then placed in an oven and allowed to dry.

Drying of Aluminum Slurry and Milling Samples

There has been a considerable amount of effort put into establishing the best way to dry a

liquid suspension. A review of the work can be found in a dissertation on dry dispersion by









Stephen Tedeschi [64]. This study will use a supercritical drying technique to dry samples

obtained from the milling process. When liquid CO2 is heated to above 31.2 C in a confined

volume, CO2 becomes a meta stable fluid termed a supercritical fluid Figure 3-4. The

supercritical drying method involves solvent exchanging the suspensions fluid with that of liquid

CO2, then heating the pressure vessel above 31.2C and releasing the meta stable fluid. The

importance of drying across each phase boundary was discussed in Tedeschi work; it was found

that supercritical drying reduced capillary forces that are present in conventional drying

techniques [64]. Figure 3-5 is the pressure vessel used in this study, it was purchased from SPI

Inc. The vessel is capable of sustaining a pressure of 2000 psi and a temperature of 50C.

The method we used is as follows. The particle suspension samples removed during

milling was placed in a 25,000 Dalton dialysis bag and diluted with isopropyl alcohol. The

dialysis bag was used to allow diffusion of solvents, while capturing the particles. The bags were

then placed in the pressure vessel, which was then filled with liquid CO2 and allowed to sit for 2

hours. After two hours, approximately 70% of the fluid inside the vessel was released and the

vessel was again filled with CO2. The step of emptying and refilling the vessel was continued

until all the isopropyl alcohol was removed, which took approximately 10 washes. After the final

solvent exchange, the vessel was heated to above 31.2 C and the meta stable fluid was slowly

released as a gas. The samples were then removed from the vessel and placed in Nalgen bottles

for storage and later analysis and testing.

Sample Preparation by Dispersion of Dry Powder

For many of the characterization and analytical techniques used in this study the milled

product must be dispersed in the dry form. A particle vacuum chamber was used to disperse the

powder. Samples prepared in this way were used in image analysis, obscuration tests, and mass

measurements. They were also used for calculating IR obscurant performance. Obscurant results









will be discussed later in the case study. The device used for this study was a Galai partial

vacuum chamber. Approximately 2 mg of supercritically dried sample was dispersed in the

chamber using 5 bars of vacuum. The amount of sample used was found by trail and error, and

was determined to be enough to give a near monolayer of particles. Figure 3-6, shows a) the

Galia chamber, as well as, b) a glass slide used for mass and image analysis measurements, c) a

silicone wafer chip used for scanning electron microscope measurements, and d) a double pass

transmission (DPT) slide used for Fourier transform infrared spectroscopy measurements

(FTIR). The exact use and importance of the samples prepared by this method will be discussed

in the appropriate characterization section.

Statistical Design of Experiment

Statistical design of experiment (DoE) is a powerful tool used to optimize and analyze

many industrial and commercial processes. This study first used statistical design to understand

and choose the parameters that are important to particle deformation during stirred media

milling. The variables studied were rotation rate, media loading, and media size. Additional

designs and analysis were used to determine the interaction between milling parameters and

define their significance. A central composite design was used to analyze the measured material

response to milling (strain) as a function of milling parameters. The DoE software used in this

study was purchased from Stat-Ease, and is called Design-Expert 7. The software is capable of

performing statistical analysis of the data, based on analysis of variance (ANOVA). It is also

capable of developing and experimental design that can then be modeled by the software.

Characterization

Characterization is a vital part of any study, it is important to understand and relate process

variable to product qualities. By characterizing and interpreting the product materials, it is

possible to determine and predict the behavior of a material or process. This study used









characterization techniques to measure the response of aluminum to process parameters before,

during, and after mill. The characterization techniques used in this study and the sample there

importance are discussed in the following sections.

Light Scattering

Light scattering has been used to rapidly and accurately measure particle size and size

distributions. The main principle of light scattering for particle size analysis is that light scatters

at different angles based on the size of the particle. A detector measures the angle and intensity

of the scattered light and is able to reconstruct a particle size distribution from that data. One

limitation of light scattering is that it assumes all particles are spherical for size calculations. This

limitation will be addressed in more detail in chapter 4. This study uses the Coulter LS11320

which is capable of measuring particle sizes from 40 nm to 2,000 im. The Coulter measures the

particles size distribution of the samples taken from the mill. In order to achieve accurate

measurements, the samples from the mill are further diluted with isopropyl alcohol to a

concentration fit for size measurement. To achieve an accurate measurement from the LS11320,

light must be able to pass through the sample, the approximate concentration must be <1 %. The

data from the Coulter is used for calculations of stress, strain energy, strain, milling efficiency

and strain rate, these results will be shown in chapter 4.

The LS1 1320 is capable of providing data output in number, surface area, and volume

distributions. The differences in these measurements are due to how they are calculated. If the

particle size distribution were narrow, these three methods would provide equivalent. However,

if the distribution is bi-model or broad a number would appear to have smaller particles size and

a volume distribution would appear larger.









Microscopy and Image Analysis

This study produced high aspect ratio metallic flakes with major dimensions as large as 10

|tm and minimum dimension sizes as small as 10 nm. This led to challenges in obtaining clear

images from one specific device. Some of these challenges were getting both quantitative and

qualitative flake thickness measurements due to difficulty in getting the flakes to stand

perpendicular to the surface. There were also challenges in getting quantitative particles size

measurements of the flake diameter to verify light scattering data. This challenge will be

addressed later in this chapter.

Electron microscopy

Most of the images used in this study were taken using a scanning electron microscope

(SEM). The SEM functions by focusing a beam of electrons onto a surface. The beam of

electrons then excites and emits electrons from the surface that are then detected and used to

reconstruct the surface topography. The microscope used in this study was a JEOL 6330 cold

field emission microscope. Samples for the SEM image were prepared by disseminating the

powder onto a silicone wafer and attaching the wafer to an aluminum stub with carbon tape. This

procedure allows for a conductive contact between the flakes and the instrument. SEM images

were used to verify light scattering data using image analysis software called ImagePro. The

software was capable of thresholding the images and counting and measuring the particles. SEM

was used to study particle morphology and determine when particle breakage occurred.

Transmission electron microscope (TEM) was used in this study to measure particles size,

and crystallinity. TEM image materials by passing an electron beam through a thin slice of

sample. TEM can also determine crystallinity and microstructure of materials. Two TEMs were

used in this study, the JEOL TEM 200CX was capable of imaging at high magnification, and the

JEOL TEM 2010F was capable of measure crystallinity over a broad size range. Samples were









prepared for the TEM analysis by mixing the dry powder with an epoxy. The epoxy was then

allowed to cure. After curing, the epoxy was sectioned using a microtome to cut slices

approximately 100 nm thick.

Optical microscopy

Imaging was also done on an Olympus BX60 optical microscope equipped with a SPOT

Insight Digital CCD camera. The microscope was capable of magnifying and image between 2 to

100 times. This microscope requires very little sample preparation to obtain images. Optical

microscopy provided a quick way to estimate particle size and morphology. However, it was

challenging to measure particles of less than 5 |tm due to the low magnification and low

resolution at that length scale.

Occhio optical particle sizer

The Occhio optical particle sizer was also used in this study. The instrument passes a

sample below a fixed high magnification CCD camera capable of imaging thousands of particles

in seconds. It then analyzes the images with Calisto software. The Occhio instrument specifies

that it is capable of measuring particles as small as 0.5 microns and as large as 3 mm. A

comparison of data between the Occhio and light scattering will be shown later.

Surface Area Analysis (BET)

Surface area analysis by gas adsorption was first achieved by Stephen Brunauer, Paul

Hugh Emmett and Edward Teller in 1938 and was later termed the BET technique [65]. BET

involves measuring the amount of gas condensed on the surface of a sample at low temperatures

and pressures. This study used the Quantachrome Autosorb 1C-MS to measure the specific

surface area of samples taken before, during, and after milling. Samples for BET analysis were

allowed to outgas under vacuum at 600C for 12 hours before being placed in the BET for

analysis. The particle size of the before milling samples were verified by image analysis and









light scattering techniques. The after and during milling samples were used to study the change

in particle size as a function of milling time and were used to determine IR obscurant

performance.

Inductively Couple Plasma Spectroscopy (ICP)

Inductively coupled plasma spectroscopy is used to measure the elemental composition of

an organic or aqueous liquid. The ICP ionizes atoms is an argon plasma, the ionized atoms emit

light when electrons return to the ground state. The light is detected and correlated to specific

atoms. The instrument used in this study was the Perkin-Elmer 3200 Inductively Coupled Plasma

Spectroscope. It has the capability to measure contents to less than 1 part per million. ICP was

used to verify the purity of the starting material and to determine the mass used in the case study.

As mentioned previously approximately 2 mg of sample was disseminated into the particle

vacuum chamber. However, only a fraction of the would end up on the DPT slide, this lead to

difficulty in accurately determining the mass-transmission relation necessary for measuring

material performance as an IR obscurant. This was resolved by also disseminating on to a glass

slide of known dimensions then measuring the concentration of metal dissolved off the slide.

From the mass an accurate measurement of mass per area was obtained, which was then equated

to the mass per area of the DPT slide. The measurement of mass and obscuration values were

then used to determine the performance of the material.

Fourier Transform Infrared Spectroscopy (FTIR)

Fourier transform infrared spectroscopy (FTIR) is most often associated with

measurements used to identify compounds in liquids and on solid surfaces. The equipment used

was the Thermo Electron Magna 760. The FTIR is capable of measuring the transmission, and

reflectance of light at wavelengths from 400-4000 cm-1. This study used the FTIR to determine

the percent obscuration of light in the infrared (IR) spectrum for the development of IR









obscurants. The development of IR will be discussed in detail in chapters 7 & 8. As discussed

previously, sample was disseminated onto DPT slides, which were then place in the FTIR. Light

was then passed through the sample and a percent transmission was measured. This value was

then used to calculate the performance of the material, the calculation will be shown later.

Light Scattering Results

In this study, it was important to utilize techniques to characterize the response (particle

size) to varying milling conditions. This study used light scattering to measure particle size and

used image analysis to validate the techniques.

Results of Aluminum Powder

Experimental studies were performed on an aluminum powder (H-2) that deforms as a

function of milling time. It was also important to confirm light scattering particle size results for

the powders used in this study. Figure 3-7 shows an SEM image of the as received spherical

aluminum powder. The powder is roughly spherical with few non-spherical agglomerates. The

image also indicates that the particle size distribution for this particular powder may be broad as

can be seen by the large number of particles that are less than a micron and the large number that

are greater than a micron. The particle size of this powder was measured by light scattering.

Figure 3-8 gives the differential volume and number distribution data from this measurement.

Figure 3-8 also shows the error bars that represent the standard deviation obtained from three

separate measurements of the as received aluminum powder. The error bars are barely visible

due to the very low sample-to-sample variations in the measurement technique. A difference in

the volume and number distributions is apparent in Figure 3-8. If the particle size distribution

was narrow and mono dispersed the number and volume distributions would overlay each other.

However, the distribution is broad, as indicated by the SEM image. In addition, when there are

many fine particles present the number distribution tends to shift to the smaller particle size









range. The mean particle size diameter as measured from the number and volume distributions

are 1.0 |tm and 4.2 |tm, respectively.

Surface area measurements were performed on the "as received" H-2 aluminum powder

using a BET. This measurement was compared with the calculations of the specific surface area

based on the mean particle size obtained from the light scattering data. The mass of an individual

aluminum particle was calculated from the density (2700 kg/m3) for particles of mean diameters

of 1.0 |tm and 4.2 itm. The calculated specific surface areas for the mean number and volume

particle sizes of 1.0 |tm and 4.2 |tm are 2.22 m2/g and 0.53 m2/g respectively. The measured

surface area from the BET was determined to be 1.55 m2/g. The measured and calculated values

are close only for the 4.2 micron particle indicating that the BET and light scattering techniques

are in close agreement.

Results for Aluminum Flake

However, this study focused on disk-shaped particles obtained from the deformation of the

spherical particles. It was therefore necessary to verify the light scattering techniques used in this

study with image analysis. To obtain a more accurate measurement of the aluminum flake we

employed the use of the Occhio particle sizing system (specifications and functionality

discussed in chapter 3). Figures 3-9 and 3-10 are the particle size distribution data obtained from

light scattering and image analysis using the Occhio. The data is for aluminum flakes that

were milled for 10 hrs. The figures show an agreement between the distributions obtained from

both measurement techniques. The mean particle size data obtained from the Occhio and light

scattering measurements were 4.6 and 6.9 |tm, respectively. The Occhio counted over 16000

particles; this makes the measure more statistically significant [66]. Table 3-1 gives a









comparison between the light scattering and image analysis data. It can be seen that the dl0, d50

and d90 have a close values.

Summary of Characterization and Light Scattering

In any application where particle size is a critical parameter, it is important to determine

what the desired form of the particle size distribution is for a process. If the number of particles

in the process is a key variable for quality control then a number distribution might be

considered. However, if the mass of the material being processed is an important process

condition the volume distribution may be important. This study will show later how the volume

of the particle is important for modeling the milling process. In this study the volume distribution

obtained from light scattering data was determined to be the most applicable.

Even though there are challenges with using light scattering to measure the particle size

and size distribution of disk shaped particles, the above work validating the technique shows that

can be used in this work. Particle size distribution of aluminum flakes obtained from light

scattering and image analysis techniques were comparable. This indicates that light scattering is

a valid technique for characterizing the major dimension of disc shaped materials.

Milling Condition Selection

Preliminary Milling Study

Stirred media milling is a complex process with numerous operating and equipment

variables affecting the performance of the mill and the final product. Some of the equipment

variables that can be manipulated are listed in Table 3-2. The vessel size, geometry, stirrer size,

and geometry were fixed. Table 3-3 gives a list of operating variables. Those variables were

media size and density, solvent, material, and surfactants. However, many other operating

variables need to be manipulated or determined in order to conduct a milling experiment.









Preliminary Milling Results

The best way to understand the effect of milling parameters is to vary conditions during a

series of experiments. A preliminary study of the effect of media loading on mill performance

was carried out using brass flakes obtained from US Bronze. The flake is the current material

used in infrared obscurant. Scanning electron microscope images of the starting material can be

seen in Figure 3-11. Statistical design of experiments was used to determine the effect of media

loading on deformation. A two factor design was used in this experiment. This design involves

varying the factor A (media loading) between five different concentrations (levels) and a second

factor B (rotational rate) between 3 different speeds (levels), then measuring a response (particle

size). Table 3-4 gives the experimental details used in this study. Experiments 4a-4c were used to

confirm the repeatability of the experiments. The other parameters such as milling medium and

medium volume, milling media size, material concentration and surfactant usage were held

constant; Table 3-5 gives the specific milling conditions used in this study.

The response variable used to measure mill performance was the figure of merit (FOM).

Figure of merit is a measure of how efficiently a material obscures infrared light (IR) and is an

indication of the specific surface area of the material. Equation 3-1 describes FOM where

transmission, T, is obtain from FTIR measurements, mass, m, is measured using the ICP and SA,

is the surface area of a microscope slide. The material constant, Cm, is considered proprietary.


LN T
FOM C 3-1
2*
SAshde

A surface response plot of figure of merit versus the factors listed above can be seen in

Figure 3-12. The figure shows that as media loading is increased the figure of merit is also

increased. Higher media loading results in a larger number of grinding media that can impart









energy on the particles. The highest figure of merit was obtained at 21% media loading. These

results suggest that even higher media loading may further improve mill performance. However,

it was found that if media loading was increased to greater that 70% that the vessel contents

would overflow at velocities higher than 1000 rpm due to the turbulent environment inside the

mill. It was then possible to maximize mill performance by maximizing media loading, this also

allowed for the study of addition milling parameters by fixing media loading at a constant value.

Establishing Milling Conditions

This study also focused on using a specific mill and milling accessories described in Table

3-2. All milling experiments were conducted using the same mill geometry, vessel size and

stirrer geometries. Therefore, all the equipment variables used in this study were held constant.

We have established the importance of some of the operating variables on mill performance from

the preliminary study. However, numerous other operating variables can be manipulated and

may have a significant effect on mill performance.

Milling medium

The milling medium is also an important parameter for a wet stirred media milling process.

The medium affects the viscosity of the system, which in turn affects the energetic of the milling

process. The medium can also affect the ability to disperse particles during milling and the

selection of grinding aids used in particle dispersion. In this study, we milled an aluminum

material exposing a very reactive native aluminum surface in the process. It was important to

choose a milling medium that would have a favorable reaction with the exposed native aluminum

surface. Experiments performed using water as a medium led to the rapid oxidation of the native

aluminum surface. This caused the material to become brittle and fracture. It also led to the

production of hydrogen gas and the generation of large amounts of heat. It was decided that this

was not a favorable condition to make high aspect ratio metallic flakes due to the production of a









large amount of alumina particles. Instead, nonpolar solvents such as isopropyl alcohol or

heptane were considered. After consulting with Dr. Kremer at Silberline Inc. and reviewing

literature on metallic flake production, the explosive hazards of high surface area aluminum

powders were brought to light [10]. Milling in heptane would produce particles of high specific

surface area with a minimum amount of passivation. This could potentially be extremely

explosive. Therefore, isopropyl alcohol, which would provide a limited amount of passivation

during the milling process, was used to reduce the threat of explosions.

Mill loading and grinding aids

One of most important parameters in a wet milling study is the viscosity of the slurry. The

slurry viscosity is likely to change as a function of milling time due to increased surface area of

the product from deformation and/or particle breakage. Many researchers have investigated the

effect of viscosity on mill performance [67-69]. Viscosity directly effects the energetic of

milling. The higher the viscosity the more energy lost to viscous friction. One method for

improving performance is by using grinding aids [70]. In order to minimize the effect of

viscosity on experiments milling was done at very low solids loading to reduce any changes in

viscosity as a function of milling time. The solids loading is the amount of material that is to be

milled, in this case, it is the amount of aluminum powder added to the mill. In this experiment 25

g of aluminum was used which amounts to a solids loading of 8.3%.

At high solids loading it becomes important to select the proper grinding aid that reduce

the viscosity of the system. As stated above, this study maintained a low solids loading in the

mill thus eliminating the need for a viscosity reducing grinding aid. Generally, grinding aids are

used to disperse particles broken during stirred media milling and to prevent particle

agglomeration.









Dry ball mills use grinding aids to prevent coldwelding of metallic particles. The most

common grinding aid used to prevent coldwelding is stearic acid. Figure 3-13 shows the

chemical structure of stearic acid. Stearic acid is a saturated fatty acid used in firework

production to coat metallic particles and prevent oxidation. In this study, the wet milling

environment prevented the coldwelding of the metallic particles. However, stearic acid was used

to prevent oxidation of the metal during milling and after drying the product. An added benefit of

the stearic acid was that it aided in the dry dispersion of the metallic particles when used for IR

obscurants.

Milling time and temperature

Milling time is one of the most significant factors when running a milling experiment.

Time affects the overall amount of energy consumption of the mill, which will ultimately define

milling efficiently. The final product and final product quality is highly dependent on the amount

of time it has been milled. On the laboratory scale, it is possible to sample the milling process at

varying milling times in order to evaluate the temporal effects on the product. In this study,

samples were taken though out the milling experiment to identify when particle deformation

and/or particle breakage occurs. Figure 3-14 represents a typical particle size distribution data

obtained from a milling experiment where particle deformation and breakage occur. The graph

shows that samples taken from the mill up to 240 minutes show an increase in particle size, this

is due to particle deformation and an increase in the major dimensions. After 240 minutes, the

particle size begins to decrease due to particle breakage. From these results it is possible to

identify at what time particle deformation is the dominate mechanism and at what time particle

breakage was dominate. It is possible to see this effect in SEM images of samples taken at

different milling times. Figure 3-15 a-c are SEM images taken at a) 0 minutes, b) 240 minutes,

and c) 600 minutes of milling. At 0 minutes, the aluminum particles appear to be roughly









spherical. As milling time increases to 240 minutes the spherical particles were deformed and are

now disk shaped "silver dollar" particles. It is important to note that at roughly 240 minutes the

edges of the particles are smooth and rounded indicating that particle breakage has not occurred,

it is also important to note that since particle breakage has not occurred that the volume of the

particles in image a) and b) are equal. This observation of constant volume is important for

calculations and models developed later in this study. After 600 minutes of milling the particles

are much thinner and nearly transparent to the electron beam. They have irregular edges

indicating that breakage has occurred and the particles can no longer be taken to have the same

volume as the "as received" aluminum powder. This study focuses on the particle deformation

process; therefore, it is important to understand when deformation stops or slows and when

breakage begins. Experiments performed in this study were milled for 10 hrs with sample taken

at times of 1, 3, 5, 8, 12, 15, 30, 45 minutes and 1, 1.5, 2, 4, 6, 8, 10 hours. The time of 10 hrs

was selected because it was the longest time required to cause particle breakage.

Temperature can affect mill performance by changing the properties of the environments

such as viscosity of the medium. This study found that temperature has an added effect from the

humid Florida environment where the experiments were preformed. Initial experiments were

performed at a constant temperature of 150C; however, condensation occurred and caused large

amounts of water to be added to the vessel. This problem was solved by increasing the milling

temperature to 250C.

Media properties and rotational rate

The preliminary study showed the effect of media loading on mill performance. It was

found that greater media loading improved mill performance. This was an intuitive result

because the greater media loading, the larger the frequency of contacts that result in deformation.

For this study, the media loading was fixed at 40% for all milling experiments. There are









additional properties of the grinding media that affect the milling process. Those properties are

hardness, density and media size. If the media hardness is not greater than the product material

then deformation and breakage will be difficult. This study uses yttria-stabilized zirconia

grinding media (mohs hardness 9.0) to mill aluminum powder (mohs hardness 2-3). Another

property of the media that can affect its performance is the media size. In previous studies,

Hamey showed that if the media is not of sufficient size it fails to impart enough energy on the

material to cause particle breakage. The author also showed that if the media is too large it does

not perform as optimally as smaller media [71]. This study will focus on grinding media of sizes,

0.5, 1.0, 1.5, and 2.0 mm, these sizes are several orders of magnitude larger than that of the

product material (mean particle size of as received aluminum powder 4.2 [tm) Media density

can significantly affect mill performance; higher density media possess more momentum. Kwade

showed the effect of media density on particle size, higher density media yielded better mill

performance for particle breakage [72]. This study focuses on one media density for the

aluminum experiments (zirconia 6 g/cm3).

Rotation rate of the stirrer directly affects the energy of the grinding media and in turn the

amount of energy the product material experiences. For this study, rotation rate was varied from

a high of 2000 rpm to 1000 rpm. Rotational rates higher than 2000 rpm also resulted in leakage

from the vessel, similar to the effect of very high media loadings. Many authors have

investigated the affect of rotational rate on breakage [61, 72]. They found that the greater the

rotational rate the more breakage that occurred. However, it is still unknown what effect rotation

rate will have on deformation. It is even difficult to postulate this effect due to the complex

nature of media-particle-media contacts. Chapter 4 will provide results of the effect of rotational

rate on deformation.













MWan. 4.19017
SU J1//1/
d10: 1.1771
dEO: 3.27005
dEO: 8L 6184


Particle diameter [pm)


Figure 3-1. Particle size distribution of as received H-2 aluminum, along with particle size data,
obtained from Coulter LS 11320.






































Figure 3-2. Image of Yttria-stabilized zirconia grinding media obtained form an optical
microscope, Imm in diameter.













































Figure 3-3. Union Process stirred media mill use in study, a) stirred media mill, and b) milling
vessel and agitator.
















61

































Figure 3-4. Phase diagram for carbon dioxide, indicating the region of supercritical fluid
formation.




































Figure 3-5. SPI pressure vessel used for supercritical drying of powder samples taken from
milling


Figure 3-6. Galia partial vacuum chamber used to disseminate powder a) chamber, b) glass slide,
c) silicone wafer chip, and d) DPT slide.








































Figure 3-7. Scanning electron microscope image of as received H-2 aluminum powder.











8

--Volume distribution
6 Number distribution


ED- 4


2

1
0
0.1 1 10 100
Particle diameter (pm)

Figure 3-8. Differential volume and number distributions for as received H-2 spherical aluminum
powder measured by light scattering.


1 10
Particle diameter (gim)


Figure 3-9. Volume particle size distribution of milled aluminum flake as measure by light
scattering.











10-
S- Volume percent image analysis -
9-

8 Cummulative volume percent
7 7 divided by 10 image analysis
C 6

a 5

E 4






0
> 3,,





1 10 100
Particle diameter (m)


Figure 3-10. Volume particle size distribution of milled aluminum flake as measured by the
Occhio image analysis.

Table 3-1. Statistical data comparison for light scattering and image analysis performed using the
Occhio particle counter.
Image analysis Light scattering
mean 6.6 11.2
d10 5.4 3.6
d50 10.1 9.6
d90 22.9 20.5









Table 3-2. Stirred media milling equipment variables.
Equipment variable
Mill geometry (horizontal or vertical)
Mill size
Vessel size
Vessel material
Stirrer geometry
Stirrer size
Stirrer material

Table 3-3. Stirred media milling operating variables.
Operating variables
Mill speed (rotational rate)
Media size
Media material (density, hardness)
Media loading (volume fraction)
Liquid medium (viscosity, density, polarity)
Material loading (volume fraction)
Surfactant, grinding aids, lubricants, and dispersants
Temperature
Time

Table 3-4. Experimental parameters used in the hexagonal statistical design of experiment.
Experiment number Media loading (%) Rotational rate (rpm)
1 0.0 2000
2 5.3 700
3 5.3 3300
4a 8.0 2000
4b 8.0 2000
4c 8.0 2000
5 16.0 700
6 16.0 3300
7 21.3 2000










Table 3-5. Milling conditions for preliminary study.
Milling parameter Level
Media size 0.5 mm
Meda type Zircona
Material Brass flake/90g
Mill temperature 20 C
Surfactant Steric Acid/5g
Medium 2-propanol/300ml


Figure 3-11. SEM image of US Bronze "as received" brass flake used in preliminary milling
experiments.









1.29

1.24

1.19

1.14

1.09


Media oadi 351
O /71]! 1 Media loading (%)


Figure 3-12. Surface response of figure of merit with respect to media loading and RPM.


Figure 3-13. Stearic acid a saturated fatty acid found in many animal fats and vegetable oils.









100
-- As recieved

80 -E 15 min.
-45 min
S60 -240min
| 600 min

40


20


0
0.1 1 10 100
Particle size (pgm)

Figure 3-14. Cumulative undersized versus particle size as at varying milling time for H-2
aluminum.






























Figure 3-15. Scanning electron microscope images of a) as received aluminum, b) milled for 240
minutes aluminum and c) milled for 600 minutes aluminum.









CHAPTER 4
RESULTS AND DISCUSSION

In order to understand and predict particle deformation during milling it is important to

study and thoroughly characterize the milling experiments. Prediction of deformation can be

obtained by modeling the milling process and the change in particle size. By combining these

two methods of studies, it is possible to establish a more comprehensive understanding of the

flaking process during milling. This chapter will discuss the empirical, semi-empirical and

theoretical results obtained from milling. These results will then be used to construct models that

explain and predict the effect of milling parameters on particle deformation during milling.

Experimental Results of Stirred Media Milling

Effect of Mill Parameters

The goal of this study was to develop a better understanding of particle deformation during

stirred media milling. An observation was made from preliminary milling studies that indicated

certain milling parameters resulted in larger diameter flakes and less particle breakage than other

parameters. This was an important observation because most processes that use metal flakes

prefer flakes with high aspect ratios that have not been broken. The particle size distributions

shown in Figure 4-1 are for 1.0 mm milling media at 1000 rpm. In this figure samples were taken

at different milling times. The bimodality seen in the figure is due to particle breakage. It can be

seen that bimodality was present at short milling times. This indicates that the experiment

resulted in breakage at short milling times. Figure 4-2 shows the particle size distributions for an

experiment, which used 1.5 mm milling media and 1000 rpm. In this figure, the distribution was

monomodal and shifted to larger particle sizes indicating that the particle deformation occurred

and there was only limited breakage. The differences in the effects of milling media on the

aluminum flakes were also seen visually. The samples milled with 1.0 mm media appeared dull









and non-reflective and showed limited pearlescence. Whereas, the 1.5 mm media produced

flakes that were highly reflective showed a high amount of pearlescence[57, 73]. Figure 4-3

shows a comparison of the maximum particle diameter achieved in each milling experiment. It

can be seen that for all rotational rates the 1.5 mm result in the largest particle size. This result

correlates well with the visual observations. In order to understand the deformation process

during milling it was necessary to study the temporal effects of media size on mean particle

diameter.

The grinding curves (mean particle size versus milling time) for this study are shown in

Figures 4-4 to 4-6. In all the experiments the mean particle size increases to some maximum

value due to the flattening of the particles. This maximum particle size was than followed by a

decrease in particle size at long milling times, which can be attributed to breakage of the

particles. It can be seen that the 1.0 mm grinding media results in breakage at the shortest

milling times and that the 1.5 mm results in the largest particle diameter. Figure 4-7 shows that

rotational rate does not have a strong influence on maximum particle size. However, rotation rate

does affect how rapidly the maximum particle size was reached. The influence of rotation rate

on deformation will be discussed in more detail later in the chapter. At first, it was not

completely understood why certain media sizes resulted in larger diameter particles and

ultimately a greater amount of deformation before breakage. However, this study developed a

theory and methodology to explain the relationship between particle deformation and milling

parameters such as media size and rotational rate. The next section compares and contrasts data

using existing models for particle breakage to the deformation data.

Kinetic Energy Model

For particle breakage, researchers have related kinetic energy to the fineness of the

product. Several researchers have shown that the size of grinding media can significantly affect









the particle breakage and particle deformation process during milling [14, 21, 63, 71]. Media size

is an important parameter in Kwade's equations of stress intensity (SI) (Equations 4-1). Stress

intensity is simply the kinetic energy (KE) of the grinding media in liquid slurry. In Equation 4-

1, Dp is the diameter of the milling media, v is the linear velocity of the media, and pf and pp are

the density of the fluid medium and milling media respectively. A more detailed discussion of

Kwade's equations can be found in chapter 2.

SI = Dp,- p ) 4-1

This equation has been the basis for the scaling up of many particle breakage processes.

However, there is no data or models relating grinding media size to particle deformation during

stirred media milling. As a first step, this study applied Kwade's models to particle deformation.

The hypothesis, equivalent to that for particle breakage, is that if higher stress intensity (KE)

resulted in a greater amount of particle breakage, than higher stress intensity (KE) may also

result in a greater amount of deformation. Selection of the milling parameters chosen for this

study was discussed in chapter 3. The effect of four grinding media sizes (0.5, 1.0, 1.5, and 2.0

mm), and three rotation rates (1000, 1500, and 2000 rpm) on the resulting deformation. Table 4-

1 gives the calculated stress intensities for all the experiments performed in this study. As media

size and rotation rate increase, the stress intensity (KE) also increases. This is due to the higher

kinetic energy resulting from the larger mass and greater velocity objects. Kwade verified the

results with experimental data for breakage, however, no one has shown whether the models are

verifiable with deformation data. Table 4-2 shows the maximum mean particle size for each

milling experiment. By comparing the data in Tables 4-1 and 4-2 it is possible to determine if

there is a relationship between kinetic energy (stress intensity) and deformation (mean particle

size). It is logical that kinetic energy is dependent on both media size and rotation rate. Thus,









equivalent kinetic energies could be obtained at different combinations of these milling

conditions. However, it can be seen in Figure 4-3 and Table 4-2 that deformation does not vary

with rotation rate. The energies for the [1.5 mm, 1500 rpm] and [2.0 mm, 1000 rpm] are

equivalent. Thus, according to this theory the particles sizes for these experiments should also be

equivalent. However, the maximum particle sizes differ indicating that kinetic energy can not be

used to predict deformation. This result led to develop of a new model that was not based on

kinetic energy of the grinding media, such as a contact mechanic model based on Hertz's theory

discussed in the next section.

Stress Model

Eskin et al. derived a model that relates milling parameters to particle breakage [20]. The

basis for Eskin's model was the use of Hertz theory to calculate a contact area, which was then

used to calculate milling efficiency. Eskin used a term called granular temperature as a

representation of velocity was used to calculate the force. In this investigation the linear velocity

is used so a more direct comparison between this model and stress intensity (KE) could be

established[74].

In order to apply Hertz theory, the contact of the colliding bodies must be elastic. In this

research, the colliding objects are spherical zirconia grinding media with an elastic modulus of

approximately 200GPa and a Poisson's ratio of 0.31. Thus, the assumption of elastic contact is

valid. The following equations were derived from Johnson's book "Contact mechanics", and can

also be found in Eskin [20, 74]. In Equation 4-2 the average normal force, Fb, of identical

elastic beads is related to the Young's modulus of the bead, Yb, the Poisson's ratio of the bead,

rib, multiplied by the density, pb, times the velocity, v, and radius of the bead, Rb.










Fb = 0.76 5b J(2pb ) 5 5vRb 4-2


The radius of the contact area, ab, was calculated from the force acting on the bead and the area

of a circle. Dividing the force by the contact area is the stress, ob, exerted at the center of the

contact area by the beads, as described by Equation 4-3 and 4-4.


ab = 1-17b Rb Fb 4-3


3 F
ab 4-4
2 (ab )2

Figure 4-8 is a schematic of the collision of elastic spheres. Table 4-3 gives the force,

contact radius and stress calculated from Equations 4-2, 4-3, and 4-4. It was determined that the

force increases with increasing media size and rotational rate. The media size and rotational rate

also increase the contact area. The stress for a specific rotational rate is equivalent for all media

size. This indicates that stress exerted by the grinding media is not sufficient to describe the

variations in deformation from varying media size. However, the important quantity to know is

the magnitude of the stress exerted on the particle and not the stress acting on the media.

The stress acting on the particle ultimately results in the deformation of the material. A

schematic of the particles trapped in the region between two colliding grinding media can be

seen in Figure 4-8. From the size of the contact area of colliding media and the concentration of

particles inside the mill, it was possible to calculate the number of particles in the contact area

and subsequently, the stress acting on each particle. The stress exerted on the particle, op, was

calculated by dividing the contact stress by the number of particles in the contact area for a

specific media size, as given by Equation 4-5 and Equation 4-6. The number of particle, Np, was









calculated by dividing the contact area by the projected area of the particle and then multiplying

by the solids volumetric concentration of aluminum particles in the mill, c.

2
b
N bc 4-5
P R2


Ca b 4-6
NP

Table 4-4 gives the values for number particles in the contact area and stress acting on the

particles as a function of media size and rotational rate. It can be seen that the stress acting on the

particle decreases with increasing media size due to the larger number of particles in the contact

area. Also, the number of particles in the contact area increases with increasing rotation rate due

to the increase in contact radius, which is attributed to the milling media colliding with greater

force. The yield stress and tensile strength of aluminum are 35 MPa and 90 MPa, respectively.

However, the aluminum in this study was heavily strained, therefore, it may be more appropriate

to consider the strain hardened values of aluminum for comparison purposes. The strained

harden values of yield strength and tensile strength of aluminum are 117 MPa and 124 MPa,

respectively [31].

Using the strained hardened values and Equations 4-2 to 4-6, data was calculated and the

following results were obtained. The 0.5 mm grinding media exerts the most stress on the

material, nearly two orders of magnitude greater than the strength of the material. Such high

stress could be responsible for low amount of deformation before breakage, this result can be

seen in Figure 4-7. In the figure the maximum particle size achieved for the 0.5 mm for any

rotation rate was approximately 13 jtm. It should also be noted that out of all the media size the

0.5 mm milling media result in the longest amount of time to reach the maximum amount of

deformation as can be seen by the figure, where, the 100 rpm experiment took almost 10 hrs to









reach the maximum particle size. This result will be explained in the next section, which will

discuss deformation rate.

Similarly, the stress of the 1.0 mm grinding media is higher than the tensile strength of

the material thus resulting in particle breakage and a minimum amount of deformation. The high

normal stress of the 1.0 mm media resulted in rapid particle breakage, which can also be seen in

Figure 4-7 and in more detail in Figure 4-1. Conversely, the 1.5 mm media exerts a stress that is

closest to the yield stress of a strain hardened aluminum material. Each compression event for

the 1.5 mm media could be considered a loading to the yield strength of the particle resulting in

finite amount of deformation. Media of this size can therefore apply the maximum amount of

deformation to the material without failure. The largest media, 2.0 mm, apply a stress to the

particle that is below the yield stress of the material and thus resulting in less deformation then

the 1.5 mm media. These results can also be seen in Figure 4-7 where the 1.5 mm results in the

most rapid deformation and the largest magnitude of deformation.

However, it can be seen in Table 4-4 that the stress acting on the particles for a specific

media size does not change significantly with rotation rate. This stability in stress values may

explain the data in Figure 4-7, which shows that the maximum deformation was not significantly

affected by rotation rate. These results are in contrast to the breakage models that state rotation

rate has a large effect on breakage. The models described above give a fundamental

interpretation of particle deformation as a function of mill parameters. However, this study has

so far neglected the influence of number frequency of media collisions on particle deformation.

The next section will discuss the effect of media size on the frequency of media collisions and

the resulting affect on deformation and deformation rate.









Deformation Rate

This study demonstrated that rotation rate affected the rate at which deformation was

achieved. Eskin and Kwade developed models that expressed the rate at which collisions

between grinding media take place. Kwade's Equation 4-7, was used to determine the rate at

which media collide. In Equation 4-7, stress frequency, SF, is equal to angular velocity, oa,, times

diameter of the stirrer, Ds, divided by the diameter of the bead, Db, multiplied by milling time, T.


SF = ),D T 4-7
Db

Kwade determined the necessary stress frequency for reaching a specific material fineness. The

author found a specific fineness could be obtained by using a high stress frequency and low

kinetic energy of grinding media, or vise versa. In addition, Kwade found this relationship held

up to a point of a minimum amount of kinetic energy. Kwade considered a constant volumetric

media loading, for comparisons of stress frequency. Eskin derived a similar model, but chose to

measure the number of times a particle encounters the media. Both authors found that the

frequency was dependent on the media size and rotational rate. Table 4-5 gives the values of

stress frequency for this study as calculated using Kwade's model. The model shows that stress

frequency increases with media size. The model also gives a linear relationship between

frequency of collisions and rotation rate. However, values in the table refer to the collision of the

media and fail to give the number of compressions the particles experiences.

The time at which the particle size begins to decrease for each experiment, was estimated

from the data in Figure 4-7, and was recorded in Table 4-6. Kwade's model indicates that smaller

media size had the largest collision frequency, as shown in Table 4-5. When this result is coupled

with the high stress acting on the particle associated with the 0.5 mm one would expect this

experiment to approach the maximum amount of particle deformation more rapidly than the









larger media sizes. However, it was shown in the previous section that the 0.5 mm reached the

maximum deformation slowly. It is challenging to compare the effect of media size on the onset

of particle breakage for all the media sizes due to the complex nature of breakage. However, it is

certain that rotational rate affects particle breakage. Higher rotation rates result in a more rapid

onset of particle breakage. In addition, Kwade's model would predict a more direct correlation

between media size and breakage, which is not the case for particle deformation as observed in

this study. In order to explain the discrepancies in these models with the experimental data it

was necessary to formulate a new model.

Therefore, in this investigation the stress frequency (Table 4-5 and Equation 4-7) was

multiplied by the number of particles in the contact area (Table 4-4 and Equation 4-5) to obtain

the total number of particles stressed as a function of milling time, Table 4-7 gives the total

number of particles stressed in 60 minutes. From Table 4-6, it can be seen that the total number

of stressing events does not change as a function of media size. However, the total number of

stressing events does change as a function of rotation rate. These results explain the increase in

the breakage as the rotation rate increases, as shown by the experimental data in Table 4-6. The

frequency model suggests that the number of collisions is only a function of rotation rate and not

of media size. This is way rotation rate only effects the time it take to achieve the maximum

amount of deformation and does not effect the actual maximum deformation achieved.

In summary, it has been found that milling media size determines the maximum amount

of particle deformation until failure. This was explained by calculating the stress acting on the

particles using Hertz's theory. Rotational rate was found to have only a minor effect on the

maximum achievable particle size, Table 4-2. The calculation of stress acting on the particle

showed that there was a minimal amount of change in stress as a function of rotational rate. Thus









confirming the result that rotation rate did not have a significant effect on maximum particle

deformation. The time it took a particle to break was modeled using stress frequency. Rotation

rate was shown to affect the rate of deformation. Existing models failed to explain how rotation

rate affects particle breakage. Therefore, a new model was derived that showed rotation rate

directly influenced the number of stressing events. The models give the fundamental relationship

between media size, and rotational rate on particle deformation.

However, the models above do not predict the energy efficiency of the milling process or

give any indication on how the milling parameters affect efficiency.

Strain Energy and Milling Efficiency

Strain Energy

One of the greatest challenges in any milling operation is the measurement and calculation

of milling efficiency. It is relatively easy to measure the amount of energy input to a milling

process, however, it is very difficult to determine how the energy is utilized. Many authors have

studied the amount of energy required to break either a single particle or an ensemble, a review

of this work was given in Chapter 2 [43, 47, 75]. However, the milling process consists of

numerous breakage events and a particle will likely fracture several times before it reaches the

desired particle size. It is then difficult to determine the energy required for a particle to break to

a specific size then break again to another size. This makes calculating the energy of breakage

difficult. Generally, milling efficiency is estimated by measuring the energy input and

subtracting the energy that is expended as heat. Most mills are cooled to maintain a constant

temperature, and the heat is usually measured from the coolant. This method does not directly

establish how much of the energy went into the material itself, only the energy dissipated as heat.

Efficiency calculated from this method does not represent the milling efficiency to cause particle

breakage, thus making it difficult to optimize the milling process.









However, the particle deformation process is unique in that the number of particles does

not change and the energy that goes into deformation is stored in the material as strain energy.

From measurements of the amount of deformation the material experienced, it was possible to

calculate the strain of the material. From the strain, it was then possible to determine the strain

energy. The strain energy was then used to calculate milling efficiency as a function of milling

parameters.

Usually strain is directly measured by studying the change in dimensions of the object. In

this study, it is difficult to measure all the dimensions of the particles being strained. It was

shown in Chapter 3 that light scattering could be used to characterize that major dimension of the

particles. However, there is no direct method for characterizing the minimum dimension, i.e.,

thickness of the flakes. Thickness was calculated by considering there was no change in the

volume of the material up to the point where the material fails. Failure is taken to be the point at

which the mean particle size begins to decrease. Equations 4-8, 4-9, and 4-10 were used to

calculate the thickness of the flake by calculating the volume of the "as received" particles. In

Equation 4-8, V, is the volume of the "as received" spherical aluminum particles and Dp,o is the

diameter of the "as received" aluminum particles. The volume of the "as received" sphere is

equal to the volume of the flake in Equation 4-9. The volume of the flake was calculated using

the equation for the volume of a cylinder, Vcy (a flake is essentially a cylinder with a small

height, ie., thickness), and the diameter of the flake, Dpf at some sampling time; tpf is the

thickness of the flake at the corresponding sampling time. The thickness of the flake was

calculated by equating the volume of the sphere to the volume of the flake using Equation 4-10.


V = 4 4-8
3 2










7 ) 2Kpf
V I= t,,f 4-9


43
4 Dpo
3 2
t ) 4-10
Pf 2



The strain, E, was then calculated form thickness of the flake by using the true stain,

Equation 4-11. A diagram for the change of the spherical particle to a flake can be seen in Figure

4-9. The true strain was calculated for every sampling time up to the point where the particle size

decreases.


E = In t 4-11
DP0

The work hardening equation, Equation 4-12, was used to calculate the stress. The coefficient

and exponent (K and n, 180 MPa and 0.2 respectively) for the work hardening equation were

obtained form literature for aluminum [31].

= Kcn 4-12

Figure 4-10 is the stress-strain curves at 1000 rpm for each media sizes. The 1.5 mm media

yields the highest strain for all the experiments at 1000 rpm rotational rate. Similar behavior can

be seen in the 1500 rpm and 2000 rpm stress-strain curves, as seen in Figure 4-11 and 4-12,

respectively. The number of data points on each set of data decreases as rotational rate increases

due to the more rapid particles breaking at the high rotational rates and thus shorter milling time.

Mill Efficiency

As discussed in Chapter 2, stirred media milling processes have been estimated to operate

at energy efficiencies of less than 2%. These estimates are for particle breakage processes and









are generally obtained from monitoring heat generated in a mill. There is no research, which

measures the amount of energy supplied to the material. The following section will demonstrate

that milling processes do not innately operate at low energy efficiencies, and will describe

methods of improving the milling efficiency for particle deformation. This study used

measurements of particle deformation to calculate the amount of energy stored in the material.

Strain energy was obtained by integrating the stress-strain curves and dividing by the

volume (volume of the particle remains constant at this point on the stress-strain curve) of the

particles. Table 4-8 is the strain energy as a function of media size and rotation rate. It is clear

that the particles milled with the 1.5 mm grinding media have the highest strain energy. It is

important to note that nearly the same strain energies were obtained at the higher rotational rates

in a much shorter amount of time.

The total strain energy divided by the energy input to the mill is milling efficiency. The

total strain energy was obtained by multiplying the strain energy by the total number of particles

in the mill. Table 4-9 gives the total strain energy from each milling experiment. The 1.5 mm

milling media resulted in the highest strain energy values, which was due to the large amount of

deformation produced by the media. The energy input into the mill was calculated using the

equations developed in Eskin [20]. In order to determine the energy input Eskin calculated the

energy for a stirred mixing tank modeled by Nagata [76]. Equation 4-13 describes the power

input, Pw, to the mill calculated from tank geometry, mill speed, and slurry density. The terms N,

H, D, and A, are the rotation rate, tank height, stirrer diameter and tank diameter, respectively.

N3D3
P. = Mp. D 4-13
A2H

Slurry density, pm, was calculated from the density of media, pb, liquid density, pL, and media

solids loading, c as shown in Equation 4-14.









P= Pbc + PL (1 ) 4-14

Mis a coefficient calculated using Equations 4-15 through 4-20. In Equation 4-15, Re is the

Reynolds number for the tank, which is calculated by Equation 4-16. The coefficients A, B, and

exponent h, are empirically determined constants [76].

A 103 +1.2Re066
M =-+B---- 4-15
Re 10 + 3.2Re066


Re = D 4-16
/m

In Equation 4-17, b is total stirrer diameter, and for this study b is equal to the diameter of the

stirrer multiplied by the number of stirrer, four (four stirrer arms on the agitator).


A =14+ 670 D-0.6 +185 4-17


13-4 (b ) 5 -114(-1
B=10 A A 4-18


h = 1.1+4 -2.5 -0.52 4-29
A IA

The dynamic slurry viscosity, /i,, used in Equation 4-16 was calculated by Equation 4-20.

Pm,, = L [1+2.5c+10c2 +0.0019exp(20c)] 4-20

Table 4-10 gives the values of the mill geometry and mill operating constants used in Equations

4-13 through 4-20. Mill power was then multiplied by the mill volume and milling time to obtain

the energy of the mill as a function of time. Equation 4-21 gives milling efficiency, 7, which is

equal to total strain energy, SET, divided by mill energy, E,.

SE
7 = *100 4-21
Em









The mill energy was estimated for a stirred tank, which may underestimate the true mill

energy. This is because a mill requires more power to move the high mass of milling media in

the chamber. This is not a major concern, because most milling processes can measure milling

energy directly from measurements of torque or power draw and therefore obtain a better

measurement of milling power and efficiency.

Figure 4-13 shows a plot milling efficiency versus milling time to breakage for the 0.5, 1.0,

1.5, and 2.0 mm media at 1000 rpm. At short milling times, the milling efficiency is very high

due to the rapid deformation of the material. It also can be seen that for the 1.5 mm media the

efficiency approaches 100 percent. This result is due to the method for estimating mill energy. It

was observed that the 1.5 mm milling process is much more efficient than the rest of the media

sizes, due to the more rapid and greater amount of deformation than the other media sizes. Figure

4-13 gives the efficiency for 1 hour of milling, however, the milling experiments were conducted

for 10 hours. It can be seen at 1 hour of milling that the efficiencies begin to approach the values

reported in literature of <2%. Further milling leads to an even more inefficient process. This

result would suggest that milling time is the leading cause of inefficiency. Therefore, it is

important to select parameters that will result in the most rapid deformation, such as, a high

rotation speeds.

Similar results are observed at rapid rotational rates as shown in Figures 4-14 and 4-15,

for 1500 and 2000 rpm respectively. The 1.5 mm media results in the most efficient milling

process, however it is difficult to determine from these plots whether efficiency was affected by

rotational rate. In order to understand the effect of rotational rate on mill efficiency, the results

for the 1.5 mm media was plotted in Figure 4-16 at varying rotational rates. By plotting the

milling efficiency at maximum strain, it is possible to obtain a relationship between rotational









rate and efficiency. Figure 4-17 is milling efficiency as a function of rotational rate. The plot

focuses on the data as it approaches particles breakage. It can be seen from the graph that the

2000 rpm speed resulted in a greater efficiency. This is due to the more rapid rate of particle

deformation. Ultimately, milling at rapid rotation rate would result in energy saving due to

increased efficiency. Rapid rotation rates would also increase time saving due to reduced

processing time. This study has shown that the benefits can be realized with no detrimental effect

to the product quality, i.e. strain. By investigating milling efficiency it was also possible to

determine the mechanisms by which energy is wasted, as discussed below.

As mentioned earlier milling processes are generally considered inefficient processes,

with efficiencies of approximately a fraction of a percent. Usually, most of the energy was

considered lost to viscous friction. This study found that milling could be a very efficient process

at short milling times, indicating that energy lost to viscous friction was grossly overestimated. It

can be assumed that the energy lost to friction is constant throughout the milling process due to

the low solids loading in the mill, and minimal change in viscosity during the milling process.

This indicates the low efficiency observed at longer milling times in this study must be due to

another mechanism. Table 4-11 shows the milling efficiency as a function of percent strain, and

table indicates that the efficiency decreases as percent strain increases. This could be explained

by the fact that it becomes increasingly difficult to cause plastic deformation in the material, due

to the increase in strain hardening with increased strain. The media contacts with the powder

results in increased elastic contacts that do not cause strain and result in a waste of energy or in

efficient milling.

In summary, this study has determined the most effective milling parameters for

maximizing particle deformation. The study demonstrated how stress acting on the particle









ultimately determines deformation. The strain energy stored in the material during processing

was also calculated, and the effect of milling parameters on efficiency was identified. However,

one of the goals of this study was to develop a model that predicts particle deformation based on

milling parameters. In order to achieve this goal, statistical design of experiments was used to

develop a milling model for predicting particle deformation.

Empirical Modeling of Particle Deformation through Statistical Design of Experiment

Statistical design of experiment (DOE) is a powerful tool used for, identification of

significant process parameters, development of semi-empirical models, and process optimization.

Many industries regularly employ the use of DOE in the development and implementation of

new process [77]. In Chapter 3, screening designs were used to identify the significant variables

affecting particle deformation. These designs successfully determined the effect of grinding

media loading on deformation. The results of these designs were used to plan future experiments

to study other milling parameters. The data presented in this chapter was the result of using DOE

to develop a model capable of predicting particle deformation. A central composite design

(CCD) was used for this purpose. The CCD was used to obtain the coefficients to a quadratic

model and to obtain a response surface. The surface response was used to obtain a relationship

between design variables (milling parameters) and the response variable (strain). A detailed

description of the CCD methodology can be found in Montgomery and in the design of

experiment software package Stat-Ease [78].

Two CCD designs were used to model the milling experiments. Both designs were two-

factor designs. The first design's data can be seen in Table 4-12. The media size for this design

was varied from 0.5 to 1.5 mm and the rotation rate was varied from 1000 to 2000 rpm.

Repetitions were used for the [0.5 mm, 1500 rpm], [1.0 mm, 1000], and [1.0 mm, 2000rpm] to

insure repeatability. The total design involved 12 experiments and the parameters modeled in this









study were rotation rate, media size, and the interaction between media size and rotation rate.

The analysis of variance (ANOVA) results are shown in Table 4-13. This study used a

significance level of 5 % or a confidence interval of 95 %, because this is a standard method of

determining statistical significance. It can be seen that the p-value for the model is 0.0165, which

means the model is statistically significant .When the p-value is less than 0.05 then it means the

terms in the model (media size and rotation rate) have a significant effect on the response

variable (strain). Equation 4-22 is developed based on the DOE and is in terms of coded factors.

The coded factors allow for a direct comparison of the coefficients in the equation. If the terms

were uncoded or the actual terms it would be necessary to normalize the equations by the

magnitude of the factors in order to evaluated there affects. For example, rotational rate has a

numerical value three orders greater than that of the media size 1000 rpm compared to 1.0 mm.

Which means the coefficient of the rotational rate in actual terms would need to be three orders

of magnitude less for the media and rotation rate to have a similar effect on strain.

The coefficients of the coded equation represent the significance of that factor effect on the

response, higher magnitudes of the coefficients indicate greater significance. In Equation 4-22, A

is the coefficient for rotation rate, B is the media size, and is the strain at failure.

E = 2.18 0.12A + 0.2B 0.067AB 0.062A2 + 0.54B2 4-22

From the equation, it can be seen that the media size (B) has the greatest effect on strain

and that the interaction between media size and rotation rate (A) has the smallest effect on strain.

To accurately model the mill performance it is important to model the process with respect to the

actual milling parameters.









Equation 4-23, models the design in terms of the mill parameters. The model was used to

predict the strain with respect to rotation rate and media size, within the framework of this design

(i.e. between 1000-2000 rpm and 0.5-1.5 mm media).

S= 3.38+7.65*10 4A-3.56B-2.70*10 4AB+2.48*10 7A2 +2.18B2 4-23

The contour plots and surface response can be seen in Figures 4-18 and 4-19, respectively.

Figure 4-20 is a plot of standard error associated with the experiments. It can be seen that the

error was low at the center points in the study but increased towards the limits. From the surface

response, Figure 4-19, a minimum strain at failure was obtained from the 1.0 mm media. The

existence of the minimum correlates with the analysis performed previously, that also showed

that 1.0 mm media resulted in the least amount of deformation. The surface response further

confirms the result that rotational rate does not significantly influence strain. This result can be

seen by the effect on strain at different rotational rates and constant media sizes. One of the

problems with this design is that extrapolation beyond the limits of the design may provide

misleading information. For example, the surface response and model indicate larger media size

will result in higher strains. However, this study has shown from empirical and theoretical data

that this does not occur. In order to develop a stronger model a second central composite design

was constructed that would investigate the effect of milling parameters on strain, beyond the

limits of the first design.

The second composite design investigated the effect of a larger media size on strain. Table

4-14 gives a list the experiments and response used in the second design. There were 11

experiments modeled in this design with repetitions of [1.0 mm, 1000 rpm], and [1.0 mm, 2000

rpm]. Table 4-15 gives the analysis of variance results for the second design. The p-value for this

design was 0.008, which means the model is statistically significant. Coded variable were used to









determine the significance of the milling parameters on strain. Equation 4-24 is the second

design's model in terms of coded variables.

E = 2.85 0.12A + 0.18B 0.045AB 0.05A2 0.58B2 4-24

The coefficients of the coded equation indicate the magnitude of the effect on the strain.

The rotation rate has a negative affect on strain meaning increasing rotational rate will decrease

strain. However, the media size has a greater effect than rotational rate, with a strong negative

second order effect. The second order effect of media size ultimately leads to a maximum in

strain at intermediate media sizes. The interaction term AB has a small coefficient, which means

that there is only a minor effect on strain from the combination of media size and rotational rate.

Equation 4-25 is the model for the second central composite design in actual terms. This model

can be used to predict maximum deformation within limits of the design.

S= -1.68 -1.11*103A+ 7.04B -1.80*104 AB + 2.01 *107 A2 -2.32B2 4-25

Figures 4-21 and 4-22 are the contour and surface response plots for the second model. Both

graphs indicate that an optimum amount of stress may be obtainable with grinding media of

approximately 1.6 mm. The plots also show that there is little effect on the deformation with

respect to rotational rate, which further confirms the qualitative result observed earlier in Figure

4-7. Figure 4-23 give a plot of the interaction between media size and rotational rate. If there was

a significant interaction the curves would intersect, but as they are, there is only a slight

interaction at the 1000 rpm rotation rate.

The small interaction terms seen in the first and second designs shown here strengthens the

hypothesis that stress determines the magnitude of particle deformation. This is because stress is

a function of media size and only moderately effected by rotational rate. It also strengthens the

theory that stress frequency is constant as a function of media size. The stress frequency only









varies as a function of rotational rate and constant as a function of media size. Both the proposed

models for stress and frequency indicate only a small interaction between the rotational rate and

media size. The standard error for the second central composite design can be seen in Figure 4-

24. In this model, the error increase at the limits of the rotation rate and at the highest media size.

In summary, design of experiments is a powerful tool for process optimization. In this

study, this experimental methodology enabled the development of models capable of predicting

particle deformation. This was proven by Equation 4-25 that relates media size and rotation rate

to strain. Design of experiments also allowed for the validation of the proposed milling theory

discussed previously by quantifying the effect and interaction of milling parameters. One more

method of validating the above results was to look at the changes in the microstructure of the

aluminum flake. If the number of dislocation is quantified as a function of milling time, it may

be possible to recalculate strain energy, which would offer another source of validation of the

milling models. However, this study only was able to perform a qualitative analysis with respect

to dislocation density as shown below.

Microstructure Analysis

The study of the microstructure and change to microstructure is an extensive part of

material science research. If the change in the number of dislocations could be characterized, it

may be possible to determine the energy per dislocation and ultimately the strain energy. An

attempt was made in this study to characterize the dislocations in the material by using

transmission electron microscopy (TEM). However, only information of crystallinity and flake

thickness were obtained, and the quantification of dislocation density was unsuccessful. It was

determined that the study of dislocation density was beyond the scope of this research.

Many of the processes that deform metallic flakes plan on the material to coldwelding to

refine the microstructure[48]. This study did not observe coldwelding taking place, and that the









wet milling process used in this study prevented the coldwelding of the metal powder. In order to

confirm these results transmission microscope images and diffraction patterns of the powder

before and after milling were taken. Figure 4-25 is a TEM image of an "as received" aluminum

particle and Figure 4-26 is the diffraction pattern for the particle. It can be seen from the

diffraction pattern that the aluminum particle in this study is single or nearly a single crystal.

Many more pictures would be needed to obtain a statistical representation of crystallinity,

however, the particles sampled here all gave similar results for crystallinity. Upon initial

inspection of the images, the dark regions in the image were thought to be grains; however,

Figure 4-26 shows a single crystal diffraction pattern which means they cannot be grains. The

discrete points indicate a single crystal pattern; a polycrystalline material would have filled or

nearly filled rings. Samples from the 1.5 mm, 1000 rpm milling experiments after 4 hr of milling

were sectioned and then imaged using a transmission electron microscope. The flakes in Figure

4-27 are approximately 100-200 nm thick after 4 hours of milling and this compares closely with

the estimated thickness from calculation of 200 nm. This indicates that the assumption that only

a small amount of failure has occurred and that the flakes maintain a constant volume was

correct. Figure 4-28 gives the diffraction pattern of a milled flake, which appears to be that of a

single crystal material as described above. This result confirms that coldwelding has not taken

place. If coldwelding was present the diffraction pattern of the milled flake would be

polycrystalline due to the welding of multiple particles together. It is possible to obtain

information about lattice strain and dislocation density from the TEM. However, this work was

beyond the scope of this dissertation.

Summary

This study has developed a fundamental relationship between milling parameters and

particle deformation. Through investigations of milling mechanics, it was possible to determine









the stress exerted on a particle as a function of rotational rate and grinding media size. The study

determined, for the first time, the frequency at which a particle was stressed during milling. It

also established a relationship between milling parameters and frequency. A novel approach was

used to calculate milling efficiency based on strain measurements and strain energy calculations.

Conformation of the frequency and stress models were obtained from statistical design of

experiment. A predictive model for particle deformation as function of milling parameters was

also developed using statistical design. Finally, TEM studies were performed that confirmed the

thickness calculations used in the strain measurements and the crystallinity of the material. The

above work is a significant step in the understanding of the deformation process during milling.










0

7_-a- "as received"
-x- 15 minutes
6 -- 2 hours
e-- 4 hours
5 A 10 hours
-
w4
C" Particle Particle
E breakage deformation








0.1 1 10 100
Particle diameter (pm)


Figure 4-1. Volume percent particle size distributions for experiments using 1.0 mm grinding
media at 2000 rpm. Bimodality in the particle size distribution is indicative of particle
deformation and particle breakage occurring simultaneously.




























Particle diameter (plm)


Figure 4-2. Volume percent particle size distributions for the experiments with 1.5 mm grinding
media at 1000 rpm, a visible shift in the particle size distribution to larger sizes is
seen due to particle deformation.


-a- "as received"
-- 15 minutes Particle
--2 hours r / deformation
-e- 4 hours
10 hours
._________oo.____ ,____ ^ \ yA __ ^_______










18

16-

o 14

o 12

a 10 Mill Speed
a| 1000 rpm
8 1500 rpm
2000 rpm
6
0 0.5 1 1.5 2 2.5
Media Size (mm)


Figure 4-3.Maximum mean particle size achieved for each milling experiment versus media size,
the 1.5 mm milling media produced the largest amount of deformation.












A-- --
1 -A - "
_m' 14 --~-0" I




12 1.0 mm zirconia media
a 0






8 0- 1.5 mm zirconia media
H f) 1.0 mm zirconia media
|[B 1.5 mm zirconia media
6 -- 2.0 mm zirconia media



4
0 5000 10000 15000 20000 25000 30000 35000 40000
Milling time (sec)


Figure 4-4. Particle deformation versus milling time at a milling speed of 1000 rpm for varying
grinding media sizes.





















E Ei
14 "4







s 81

.2 I 0.5 mm zirconia media
S* [ 1.0 mm zirconia media
2L 6
c- *B [A 1.5 mm zirconia media
S-* < 2.0 mm zirconia media
4 ,






0 5000 10000 15000 20000 25000 30000 35000 40000
Milling time (sec)


Figure 4-5. Particle deformation versus milling time at a milling speed of 1500 rpm for varying
grinding media sizes.












-0-.5 mm zirconia media
- I3 1.0 mm zirconia media
-A 1.5 mm zirconia media
- 2.0 mm zirconia media


0 5000 10000 15000 20000 25000 30000 35000 40000
Milling time (sec)


Figure 4-6. Particle deformation versus milling time at a milling speed of 2000 rpm for varying
grinding media sizes.












16 a) 0.5 mm media
14
:12
1-0,----


10 -

6 -0 1000 rpm
S-A 1500 rpm
-- 2000 rpm
2

0 5000 10000 15000 20000 25000 30000 35000 400
Milling time (sec)

16-
14 A
12
10
8
6
4 1000 rpm
-A 1500 rpm
22 2000rpm c) 1.5 mm media
0-


0 5000 10000


15000 20000 25000 30000 35000 4000(
Milling time (sec)


E
12
t; 10


6 O
c 41
2


b) 1.0 mm media






-0 1000 rpm
A 1500 rpm -
I--'-- 2000 rpm
3X 2000 rpm


0 5000 10000 15000 20000 25000 30000
Milling time (sec)


35000 4000(


0 5000 10000 15000 20000 25000 30000 35000 40000
Milling time (sec)


Figure 4-7. Mean particle size versus milling time as a function of rotation rate, a) 0.5 mm
grinding media, b) 1.0 mm grinding media, c) 1.5 mm grinding media, and d) 2.0
grinding media.

Table 4-1. Calculated stress intensity for experiments performed in this study. The table indicates
that it is possible to achieve equivalent energies and different milling conditions.


Rotation rate (rpm)


1000
1500
2000


+


0.5 mm media


Stress intensity *10^-5 (joule)
1.0 mm media 1.5 mm media


69
160
280


2.0 mm media


160
370
660


Table 4-2. Maximum mean particle size of milling experiments. Even though the kinetic energy
of some of the milling experiments are equivalent the maximum particles sizes differ.


Rotation rate (rpm)
1000
1500
2000


Maximum mean particle diameter at failure (pm)
0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
11 11.0 +/-0.6 16 12
13.0 +/-0.3 9 15 12
11 9.4 +/-0.8 14 11


V--c 0








Fb


\ I////
Aluminum


particles
Fb

Figure 4-8. Contact between colliding elastic spheres. Aluminum particles are caught in the
contact area between the media.










Table 4-3. Force, radius of contact area, and stress acting on the milling media as a function of
media size and rotation rate (calculated using Hertz theory).
Force of milling media (N)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 3.5 14 32 56
1500 5.7 23 52 92
2000 8.1 32 73 130
Radius of contact area of milling media *10^-5 (m)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 1.4 2.9 4.3 5.7
1500 1.7 3.4 5.0 6.7
2000 1.9 3.8 5.7 7.5
Stress acting on milling media (MPa)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 8300 8300 8300 8300
1500 9700 9700 9700 9700
2000 11000 11000 11000 11000


Table 4-4. Number of particles stressed in a single collision between grinding media and the
stress exerted on an individual particles.
Number of 4.2 pm diameter "as received" aluminum particles in contact area
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 7 30 67 119
1500 10 41 92 164
2000 13 52 116 206
Stress acting on an individual particle (MPa)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 1115 279 124 70
1500 948 237 105 59
2000 845 211 94 53


Table 4-5. Stress frequency as calculated from Kwade's model. The frequency is the number of
times milling media collide with each other.
Stress frequency (Hz) *10A8 in 60 minutes of milling
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 54 14 6 3
1500 81 20 9 5
2000 110 27 12 7









Table 4-6. The time required to reach the maximum particle diameter. As milling speed increases
the time it takes to obtain the maximum particle diameter decreases.
Milling time at breakage (second)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 36000 14400 36000 28000
1500 14400 1800 14400 5400
2000 14400 3600 3600 2700

Table 4-7. Total number of particle compressions in 60 minutes, as calculated using stress
frequency and number of particles, Equation 4-2 and 4-6.
Number of particle compressions in 60 minutes *10^10 (Hz)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 4 4 4 4
1500 8 8 8 8
2000 14 14 14 14


tp,f


Dp,f


Figure 4-9. Diagram of the change in dimensions of particle and the measurements used to
calculate stress and strain.


r- 12






























0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5


Strain


250


200


" 150


100


50


Strain


0 1 2 3


0 1 2


Strain


Strain


Figure 4-10. Stress-strain behavior for experiments performed at 1000 rpm a) 0.5 mm media, b)
1.0 mm media, c) 1.5 mm media, and d) 2.0 mm media.





















105
















200


a 150

(n
100

50

50


0 1 2


Strain


250


200


a 150


100


50


150


100

-50- 1.0 mm media
50-


o[]-------------
3 0 0.5 1 1.5 2 2.5

Strain
250


200


a 150


100

-B-2.0 mm media
50


HIllH


0 1 2 3


Strain Strain


Figure 4-11. Stress-strain behavior for experiments performed at 1500 rpm a) 0.5 mm media, b)
1.0 mm media, c) 1.5 mm media, and d) 2.0 mm media.























106











250 250

Cl b)



150 150


100 100-

SB- 0.5 mm media 1.0 mm media
50- 50


o[]------------- o[]---------
0 0.5 1 1.5 2 2.5 0 0.5 1 1.5 2 2.5

Strain Strain
250 250


200 200


150- 150-
0) I )
) I II) I
= 100- 100

S1.5 mm media --2.0 mm media
50- 50



0 1 2 3 0 1 2 3

Strain Strain


Figure 4-12. Stress-strain behavior for experiments performed at 2000 rpm a) 0.5 mm media, b)
1.0 mm media, c) 1.5 mm media, and d) 2.0 mm media.

Table 4-8. Strain energy per particle for deformation study.
Strain energy per particle *10"-8 (joule)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 1.6 1.5 2.2 1.7
1500 1.9 1.3 2.1 1.8
2000 1.6 1.2 2 1.7









Table 4-9. Total strain energy for deformation study. The 1.5 mm media milling experiments
resulted in the largest amount of stain energy.
Total strain energy for each milling experiment (joule)
Rotation rate (rpm) 0.5 mm media 1.0 mm media 1.5 mm media 2.0 mm media
1000 3900 3700 5300 4200
1500 4400 3100 5000 4300
2000 3800 2970 4700 4000


Table 4-10. List of constants for equations used to calculate milling power


Constants


Values


N 1000, 1500, 2000 rpm

H 0.05 m

D 0.06 m

A 0.09 m

Pb 6000 kg/m^3


785 kg/m^3

0.028 m

0.00286 kg/(m*s)

0.533










120


-R-0.5 mm media, 1000 rpm
100 x -0- 1.0 mm media, 1000 rpm

-X- 1.5 mm media, 1000 rpm
80, -o 2.0 mm media, 1000 rpm
-80 ,


60

40
= 40 -L



0 \
20 -^^

^~-e-- '* --

00 -------------------------' -"
0 500 1000 1500 2000
Milling time (sec)


Figure 4-13. A plot of milling efficiency as a function of milling time at 1000 rpm, for all media
sizes.














-x 0.5 mm media, 1500 rpm
-0- 1.0 mm media, 1500 rpm
S- -X- 1.5 mm media, 1500 rpm
-- 2.0 mm media, 1500 rpm


500 1000 1500


2000


Milling time (sec)


Figure 4-14. A plot of milling efficiency as a function of milling time at 1500 rpm, for all media
sizes.


40


35


30


25
>-
" 20

- 15


10


5


0













-R-0.5 mm media, 2000 rpm

20 -0-- 1.0 mm media, 2000 rpm
x -X- 1.5 mm media, 2000 rpm
-o 2.0 mm media, 2000 rpm
-0
15 -


t-





5 _=X\



00 ----------------- ~i LJ -
01
0 500 1000 1500 2000
Milling time (sec)


Figure 4-15. A plot of milling efficiency as a function of milling time at 2000 rpm, for all media
sizes.
















--1.5 mm media, 2000 rpm
-0- 1.5 mm media, 1500 rpm
-X- 1.5 mm media, 1000 rpm





Efficiency at failure

....-------------- -- ------ -


0 5000 10000


15000 20000 2
Milling time (sec)


5000 30000 35000


Figure 4-16. A plot of milling efficiency as a function of milling time for the 1.5 mm at varying
rotational rates.


40000













0.9

0.8

0.7

F 0.6

4--
A) 0.5

S0.4 \

0.3 \
-0-1.5 mm media, 2000 rpm X
0.2
-0- 1.5 mm media, 1500 rpm
0.1 -X- 1.5 mm media, 1000 rpm

0
1000 10000 100000
Milling time (sec)


Figure 4-17.A plot of milling efficiency at failure, right side of Figure 4-13 used for comparison
of efficiencies as a function of rotational rate.

Table 4-11 .Milling efficiency as a function of percent amount of strain the material has
experienced at 1000 rpm and varying media size. The 1.5 mm media reaches over 60
% of the maximum strain while milling at 37% efficiency. The last few percent of the
maximum strain result in the most inefficient milling.
0.5 mm media 1.0 mm media 1.5 mm media
Percent of total strain Mill efficiency Percent of total strain Mill efficiency Percent of total strain Mill efficiency
23 8 37 14 64 37
50 2 57 9 76 17
69 0.8 76 3 83 4
91 0.4 92 1 91 2


I










Table 4-12. Experimental design for central composite design 1.

Factor 1 Factor 2 Response

Run A: Rotation rate (rpm) B: Media size (mm) Strain at failure

1 1500 1.5 2.89

2 1000 1 2.41

3 2000 1.5 2.72

4 1500 1 1.95

5 1000 1 2.26

6 1500 0.5 2.70

7 2000 0.5 2.31

8 2000 1 2.17

9 2000 1 1.88

10 1000 1.5 3.04

11 1000 0.5 2.36

12 1500 0.5 2.63


Table 4-13. Analysis of


Source

Model

A: Rotation

B: Media S

AB

A\^2

B^2


variance for the
Sum of
Squares
1.222

n rate 0.123

Size 0.261

0.018

0.009

0.812


central composite design 1 described in Table 4-12.
Degrees of Mean F p-value
Freedom Square Value Prob>F

5 0.244 7.141 0.0165

1 0.123 3.579 0.1074

1 0.261 7.613 0.0329

1 0.018 0.532 0.4931

1 0.009 0.276 0.6182

1 0.812 23.715 0.0028























N


1.00



0.75


Strain at failure


1500 1750 2M0
Rotation rate (rpm)


Figure 4-18.Contour plot for central composite design 1, a minimum strain of 2.17 can be seen at
approximately 0.9 mm milling media.


































1.25

Media size (mm)


Rotation rate (rpm)

Figure 4-19. Surface response for central composite design 1, a trough exists at the 1.0 mm
media size.


2000












0.161


0.14425


-o
0.1275

0.11075
-o
0.094


1.50




Media 1.00 20
size (mm) 1750
0.75 1500

0.50 1000 Rotation
rate (rpm)

Figure 4-20. Standard error associated with central composite design 1, error increases at the
limits of the study.










Table 4-14. Experimental design for central composite design 2.
Factor 1 Factor 2 Response
Run A: Rotation rate (rpm) B: Media size (mm) Strain at failure

1 1000 1 2.41

2 1000 1 2.26

3 1500 1.5 2.98

4 1500 2 2.55

5 2000 2 2.39

6 2000 1.5 2.72

7 1000 2 2.50

8 1500 1 1.95

9 2000 1 1.88

10 2000 1 2.17

11 1000 1.5 3.04











Table 4-15. Analysis of variance for the central composite design 2 described in Table 4-10.
Sum of Degrees of Mean F p-value
Source Squares Freedom Square Value Prob>F
Model 1.216 5 0.243 11.921 0.008
A: Rotation rate 0.106 1 0.106 5.175 0.072
B: Media Size 0.230 1 0.230 11.254 0.020
AB 0.011 1 0.011 0.546 0.493
A^2 0.005 1 0.005 0.265 0.629
B^2 0.717 1 0.717 35.129 0.002











Strain at failure
200 2.50




1 75 -



Sf2.85

S1 50 -



2.68
1 25 -





1I I I00
1000 1250 1500 1750 2000

Rotation rate (rpm)


Figure 4-21. Contour plot for central composite design 2, shows a maximum at approximately
the 1.6 mm media.


























2.WX 1750
-7-5

Media size (mm) 10 Rotation rate (rpm)


Figure 4-22. Surface response for central composite design 2, a peak can be seen at intermediate
media size.










Interaction


3.10 -




2.75 -




S2.40 -




S2.05 -




1.70 -


A 2.0 mm milling media
* 1.0 mm milling media


I I I I I
1000 1250 1500 1750 2000


Rotation rate (rpm)

Figure 4-23. A plot of the interaction between the media size and rotation rate, it can be seen that
there is only a slight interaction at low.

















11425


0.1005

0
b 0.08675


S 0.073



2.00 2000
1.75 1750
1.50 1500

Media size (mm) 1.25 1250 Rotation rate (rpm)
1.00 1000


Figure 4-24. Standard error associated with central composite design 2.
































122
















































Figure 4-25. Transmission electron microscope image of "as received" aluminum particle.

















































Figure 4-26. Diffraction pattern obtained from particle in Figure 4-20.














124








9%


0.5 pm

Figure 4-27. Transmission electron microscope image of sectioned flake after 4 hours of milling.













































Figure 4-28. Single crystalline diffraction pattern of sectioned flake after 4 hour sof milling.

















126









CHAPTER 5
CASE STUDY: MATERIAL SELECTION AND DEVELOPMENT FOR INFRARED
OBSCURANTS

Introduction

U.S. Army has used obscurant smokes and clouds as a defense against spectral targeting

for many years. Recent advancements in technology have led to communication and detection

systems functioning in the infrared and microwave region of the electromagnetic spectrum and

new obscurant materials are needed to meet these wavelengths. According to

"globalsecurity.org", virtually all nations have access to sensors and guided munitions utilizing

the infrared portion of the electromagnetic spectrum. The widespread availability of these

devices has led the United States Military to search for the best countermeasures to mitigate their

effectiveness.

Janon Embury measured the mass extinction coefficients of many materials and later

theoretically calculated the extinction coefficients for materials depending on their shape and

size [7, 8, 79-82]. The calculations indicated that electrically conductive high aspect ratio

particles with the minor dimension in the nanometer size range and the major dimension of about

1/3rd the wavelength of light would exhibit many fold improvement over the materials currently

available. The method of choice for this purpose is to disperse fine metal flakes or rods into the

atmosphere between friendly forces and the enemy. With higher obscuration efficiency, less material

will be required, thus reducing the weight or volume of obscurant munitions carried by soldiers

or combat vehicles. Attenuation by a particulate cloud follows Beer-Lambert law, Equation 5-1,

where I, and I are incident and attenuated light intensities, a is the extinction coefficient, C is the

concentration of the particles in the light path of length L.










In )= -a-C-L 5-1


In Equation 5-1, a is dependant on size and optical properties of the particles and its units may

be expressed on either mass or volume basis with the use of corresponding concentration and

path length units.

The goal of this study was to obtain a four fold improvement over the existing IR

obscurant material. In order to achieve this goal the army uses Equation 5-2 to gauge the

performance of the obscurant.

Figure of Merit = FoM = a p Y P 5-2

In Equation 5-2, a is the extinction coefficient on the mass basis, p is the material density, Y is

the yield factor that is a measure of the ease of a material to be disseminated and is the ratio of

the mass of particulates successfully disseminated and the total mass of the material in a grenade.

The factor P is the packing factor that measures the mass fraction of material in the fill volume.

From these terms it can be seen that the figure of merit is a function of the type of material,

material properties, loading of the grenade and dissemination of the powder in the grenade. The

Particle Engineering Research Center applied a systems approach to achieve this goal by

improving every factor in the figure of merit equation (FOM). However, this study will only

focus on the selection and development of obscurant materials, for example, thinner flakes will

be used that will increase the area of coverage at a cost of less material. The other aspects of the

FOM equation can be found in a thesis by Tedeschi [64].

Development of New Obscurant Materials

In order to achieve the improvement goal, the FOM has to increase four fold. The right

hand terms in the Equation 5-2 are not mutually exclusive. First, the a is dependant on the

material. Higher a values are achieved by using high aspect ratio and electrically conductive









particles. Based on previous work, metal flakes and rods were expected to be ideal candidates [7,

8, 79-82]. The best shape and size of materials are flakes and rods with a major dimension

greater then 3 |tm and a minimum dimension of less than 100nm and lower the better. These

high aspect ratio materials (major/minor dimension) are difficult to manufacture. The cost of the

material and processing into desired shape and dimensions must be minimized to reduce the

overall cost of the grenade. Metal nanorods with these dimensional requirements are difficult and

expensive to produce. Entanglement of fibers also results in poor dispersion and dissemination.

On the other hand, flakes nearing these dimensions have been manufactured in large quantities

for decades by milling. Metal ductility plays a significant roll in the aspect ratio of the flake that

can be produced. It is possible to obtain a higher aspect ratio with softer materials, which means

that softer materials can be made thinner, thus increasing the aspect ratio. Higher density

materials have been shown to have greater dissemination efficiencies, resulting in higher

performance. It was possible then to construct a chart of relevant material properties ductilityy,

electrical conductivity, and density) to aid in the selection of candidate materials.

The relative magnitudes of density, ductility, and conductivity for several materials can be

seen in Figure 5-1. Theoretically, a material with high values in each of these areas would be and

ideal infrared obscurant. The figure shows that one such material is copper, which has a high

ductility, conductivity and density. Another possible candidate would be silver; however, silver

is a precious metal and is expensive. Aluminum is also a good candidate with respect to

conductivity and ductility; however, its low density will hurt its overall performance on a volume

basis. It is also possible to use alloys of some of these materials to increase performance.

As shown above the data presented in Figure 5-1 is a useful tool for selecting possible

materials, however, density, ductility, and conductivity are not the only material properties that









are important to material selection. As mentioned earlier, cost of the raw and processed material

is also important. Aluminum has been found to be pyrophoric, which means that it is possible to

ignite the aluminum powder. Copper powder does not form a passivation later, which means it

will eventually oxidize to copper oxide which will reduce its conductivity. There are also

environmental concerns for all the materials. Additional work in coating the material also being

performed in this project, however, this work is beyond the scope of this specific study.

Several materials were selected for processing from Figure 5-1. A comparison of the

experimentally measured performance of different materials is shown in Figure 5-2. This figure

shows that copper gives the highest performance followed by silver-coated copper and

aluminum. The reason for the high performance of copper is due to its high ductility, high

density and high conductivity. The performance of aluminum is relatively low because of its low

density (roughly a third that of copper). Even though tin has a high ductility and high density, it

has a low performance due to it low electrical conductivity.

A wet stirred media milling process was used to produce high aspect ratio flakes from the

raw material. This is the same process that was the focus of this dissertation, however, it was

found that milling just beyond the point of particle breakage gave the best obscurant material,

this result will be discussed later in this chapter. The milling equipment and procedures for this

study are the same that were used in the dissertation work with the exception that most of the

experiments in the case study involved longer milling times. Several of the characterization

techniques discussed in Chapter 3 were developed specifically for the measurements of the

figure of merit of the materials. However, the next section will give a brief review of these

techniques and how they were used to calculate FOM.









Material Characterization and Obscurant Performance Measurement Method

Particle size characterization was performed on the samples using the Coulter LS Model

13320 laser diffraction particle sizer. The specific surface area measurements are obtained from a

Quantachrome NOVA 1200 gas sorption instrument. Images of the particles were taken using a

JEOL 6330 field emission scanning electron microscope.

The FOM values were calculated from transmission data obtained from the FTIR using a

dual pass transmission (DPT) cell and from measurement of the amount of material per area. In

this measurement the powder was dispersed onto DPT slides and the transmission of IR light

through the slide was measured in a reflection mode. By this method the degree of dispersion

and material performance could be evaluated. The amount of powder used was within the

measurement error of lab balance so for a higher degree of accuracy Inductively Coupled Plasma

Spectroscopy (ICP) was used and is described in detail below. This allowed the accurate

measurement of masses well below those seen for these experiments. The material performance

was determined based on the material FOM, or the improvement of the new materials compared

to a standard material (brass flakes) supplied by the US Army.

The standard procedure used for FOM measurements is as follows. The dry powders were

dispersed in a Galai vacuum dispersion chamber onto a glass microscope slide and a DPT

slide. The powder was allowed to settle for forty five minutes after which the DPT slide was

placed in the FTIR for measurement. The glass microscope slide was placed in a Nalgene

bottle with a 1M HCL acid solution to dissolve the metal in order to accurately measure the

concentration using the ICP. Mass was calculated by measuring the concentration of Al in

solution, and normalized by the area of the glass slide.

The Perkin-Elmer Plasma 3200 Inductively Coupled Plasma Spectroscopy (ICP) system is

equipped with two monochromators covering the spectral range of 165-785 nm with a grated









ruling of 3600 lines/mm. The ICP operates on the principle of atomic emission by atoms ionized

in the argon plasma. Light of specific wavelengths is emitted as electrons return to the ground

state of the ionized elements, quantitatively identifying the species present. The system is

capable of analyzing materials in both organic and aqueous matrices with a detection limit range

of less than one part per million. This detection limit is well below the concentrations used in

these experiments and therefore allowed accurate measurement of the material mass

disseminated.

Production of Infrared Obscurant

In this project several metallic materials were measured, their FOM was compared with the

current obscurant material brass which has a FOM of 1. This study evolved over a several years,

the first material chosen to work with was the current brass material. Further milling was used to

improve the current material. The starting material was brass flake obtained from US Bronze.

The flake was produced by dry ball milling with the addition of stearic acid to prevent oxidation.

The stearic acid coating also aids in the dispersion of the brass by making the brass surface

hydrophobic.

A screening design of experiment (DOE) was followed to determine the effects of media

loading and rotational rate on the massed based FOM of brass. The milling media used in all

experiments was high-density yttria stabilized zirconia with a mean particle size of 500 microns.

The rotational rate was varied from 700, 2000, and 3300 rpm. The experimental parameters for

this study can be found in Table 5-1.

The results of the initial screening design can be seen in Figure 5-3. It was determined that

rotational rate had very little effect on FOM value and showed that as media loading increased

FOM increased. It should be noted that the effect of rotational rate on FOM is similar to that

observed in the theoretical study of deformation of aluminum. An increase of 25% in the FOM









was obtained from the screening design. The DOE determined that larger media loading

increased the FOM. Using the information obtained form the DOE, Figure 5-4 gives FOM for

the "as received" material, and that milled with a media volume of 480 ml of milling media. It

can be seen in Figure 5-4 that a much larger FOM is obtained when the media volume was

increased. However, it can also be seen that at high wavelengths the FOM value for high media

loading decreases more rapidly. The decrease at high wavelengths is due to the reduction in

particle size form milling. When the particle sizes become less than 1/3 the size of that

wavelength no longer scatters light at the wavelength [7, 8]. This design enabled future

experiments to set the media loading at a high concentration.

The second design used in this study was a full factorial (22) design; this involves choosing

two factors (rotational rate and media size) and two levels per factor for a total of 4 experiments.

Table 5-2 is a list of the milling parameters used in the 22 statistical design. The media loading in

this experiment was fixed at 300 ml, which is the maximum capacity of the mill. This design

measured responses of FOM and particle size at milling times of 6,12, 18, and 24 hrs. The

surface response of figure merit versus the design factors can be seen in Figure 5-5. At

increasing media sizes the figure merit increased. It was determined that the 1.5 mm media

resulted in the highest figure of merit. The increase in the FOM is due to the increased

deformation of particles.

It is important to note that particle sizes roughly 1/3 the wavelength of IR light do not

effectively scatter light at that wavelength [8]. Particle breakage in the mill can reduce the

effectiveness of the material at longer wavelengths. Figure 5-6 gives the effect of milling time on

the particle size distribution of brass. As received brass has a mean particle size of 10 |tm with a

nearly normal distribution and as milling time increases the particle size decreases. The particle









size distribution becomes bi-modal, which is indicative of a breakage mechanism. After 24 hr the

mean particle size was approximately 2.3 itm. This means that the flakes should no longer

effectively scatter light above 7 rtm. Figure 5-7 is a plot of FOM versus wavelength at varying

milling times. As milling time increased the FOM increased, this is due to increase in particle

deformation. At a milling time of 18 hrs the FOM begins to decrease at the higher wavelengths,

due to particle breakage, the affect of particle breakage on the decrease in more dramatic at 24

hrs of milling time. At 24 hrs the particle size is 2.3 |tm and the wavelength where the FOM

begins to decrease is 7 |tm, this agrees with the calculations made by Embury. The combination

of these designs and the theoretical studies established a basis for empirically relating milling

parameters to FOM. The designs determined that high media loadings and that the 1.5 mm

grinding media resulted in the highest performing flake. Rotation rate was found to have only a

minimal affect on FOM. These results correlate with the theoretical study of particle

deformation.

Media density was not investigated in the study of deformation of aluminum particles,

however, density is a component of all the models. Theoretically media density should affect the

stress acting on the particle and in turn affect particle deformation. Experiments were performed

to mill copper with stainless steel, zirconia, and tungsten carbide grinding media (densities of

9.8, 6.0, and 16 g/cc, respectively), Figure 5-8 gives these results. Tungsten carbide media

results in the largest amount of deformation before breakage, followed by stainless steel, and

zirconia. That indicates the highest density media, tungsten carbide density (16 g/cc), performed

the best and, zirconia, the lowest density media (6 g/cc) was the worse.









Summary

The above studies proved that is possible to improve the obscuration properties of existing

materials and to produce new and improved obscurants by stirred media milling. It was found

that high aspect ratio flakes could be produced by milling soft metals, that the softer the metal

the thinner the flake that could be produced. A relationship based on fundamental material

properties was derived for selecting materials that would produce the best infrared obscurant.

The relationship suggested that copper, aluminum, and silver would yield the best flakes from

milling. The study found that larger media produced a larger thinner flake; the results indicate

that the smaller media fractured the materials causing the materials to have a lower aspect ratio

that was detrimental to their performance. Denser media yield higher aspect ratio in shorter

milling time. The shorter milling times would ultimately result in economic saving if the process

was scaled up. Milling speed did not appear to have any effect on the quality of the product;

however, greater rotational rate resulted in much faster deformation. Greater rotational rates

would result in energy and time saving if the process was scaled up. During this study particle

size was closely monitor by sample from the milling during the experiments. All the experiments

showed and initial increase in the particle size at short milling times due to deformation followed

by a decrease due to particle breakage. This study found that a final optimum particle size of

approximately 5 |tm yielded the highest performing obscurant.































C E C 0, 0 Q E
c c o, c !>F

EE
C 3 0
E
C-


Figure 5-1.A chart of relevant material properties of materials for use as an infrared obscurant.
High values in each property are preferred for candidate materials.
















L.





UL


40


30


20


10


3 6 9 12 15 18 21 24
Wavelength (micron)

Figure 5-2. The performance of various milled metals, the copper has the highest performance to
due to the properties being the highest in all the desired area.


Table 5-1. Experimental parameters used in the statistical design of experiment used to


determine milling parameters.
Experiment
1
2


Media volume (ml)
0
40
40
60
60
60
120
120
160


Rotation rate (rpm)
2000
700
3300
2000
2000
2000
700
3300
2000









1.29


1 .24



1.14
0
109







3299
2 50
S 2000 120 160

"e 1351 480

S0 edia volume (ml)

Figure 5-3. The results of the initial design of experiment, indicating that higher media loading
increase the performance of the material.













1.9 As received brass

1.7 Media volume of 480 ml

1.5

S1.3

1.1

0.9

0.7
3 5 8 10 13 15
wavelength (urn)
Figure 5-4. Plot of FOM for as received and ground brass at specific milling condition of 480ml
media volume and 700 rpm rotational speed

Table 5-2. Experimental parameters used in the full 2 factorial (22) statistical design of
experiment.
Experiment Media size (mm) Rotation rate (rpm)
1 0.5 500
2 0.5 1500
3a 1.0 1000
3b 1.0 1000
4 1.5 500
5 1.5 1000












3.13


2.78

2.40 0

0 2.04 /
S1.67






1.50

1.25 1500
44 1111 125-n



5 500

Figure 5-5. The results of the affect of varying media size and rotation rate on the performance of
a milled brass flake.











-- As Received (10.3um)
-X- 6hr (7.5um)
G--- 12h (5.4um)r
S-- 18hr (3.4um)
-- 24hr (2.3um)









-I




J~oP -^^ V U


0.1


Particle diameter (pLm)

Figure 5-6. Differential volume particle size distribution at increasing milling times.


5.00

4.50

4.00

3.50

3.00

2.50

2.00

1.50

1.00


3 5 7 9 11 13

Wavelength( m)


Figure 5-7. Figure of merit versus IR wavelength for increasing milling times.


7


6


5


4

E
. 3
0

2


1


0

























-A- Tungsten Carbide


-- Stainless Steel


25

20

15

10


Milling time (hr)

Figure 5-8. Mean particle size as a function of milling time for different density milling media,
zirconia (6 g/cc), stainless steel (9.8 g/cc), and tungsten carbide (16 g/cc).


- Zirconia


B,
B









CHAPTER 6
SUMMARY, CONCLUSIONS, AND FUTURE WORK

Summary and Conclusions

The goal of the study was to develop a fundamental and predictable understanding of

particle deformation during stirred media milling. This study confirmed that particle deformation

during stirred media milling is a complex phenomenon. It is difficult to understand and predict

the interaction of milling parameters with each other and the material being processed. Novel

characterization techniques and methods were used to understand the deformation process.

Statistical analysis and design of experiments were used to interpret data, develop predictive

models, and develop screening experiments for determining the most significant milling

parameters and their effect on deformation. Relationships between milling parameters, particle

deformation, and milling efficiency were established based on the fundamental theories of

contact mechanics and mechanical behavior of materials. Finally, knowledge gained from this

study was then used to develop a material for the US Army to obscure infrared light.

In order to achieve the goals of this study a systematic approach was used to study the

process including:

* Development of characterization techniques and sample preparation methods capable of
measuring high aspect ratio particles

* Determine the milling parameters that have the most significant affect on particle
deformation

* Establish a relationship, based on fundamental theories between, milling parameters and
particle deformation

* Calculate milling efficiency with respect to milling parameters

* Develop milling models capable of predicting particle deformation

To achieve these objectives it was necessary to develop particle sizing techniques capable

of measuring high aspect particles. The predominate characterization technique used in this study









to measure particle size was light scattering, which was designed for measuring spherical

particles. The results proved that light scattering was capable of measuring the major dimension

of high aspect ratio particles.

The case study required the ability to measure the capability of metallic flakes to obscure

IR light. The measurement needed to be representative of how the metallic flakes would perform

in an aerosolized cloud in a combat environment. In order to obtain this measurement three

challenges had to be addressed:

* Obtain a dry dispersible powder
* Disperse powder
* Characterize the dispersed powder

A dry powder was obtained through supercritical drying. This technique involved the

exchange of the solvent (alcohol) to liquid carbon dioxide. Following the solvent exchanges, the

sample was then dried using the supercritical drying technique described in Chapter 3, this

method yielded an easily dispersible powder. The sample was then disseminated onto a dual pass

transmission slide and placed into and FTIR for characterization. These procedures allowed for a

measurement of grenade performance that closely relates to that observed in the field.

Several statistical designs of experiments were performed in this study. Preliminary

designs were used to determine which milling parameters were significant and required a more

rigorous study and which variables could be held constant. The preliminary designs determined

that a maximum media loading was important for causing the maximum amount of particle

deformation. The early designs also showed that rotation rate and media size could affect media

loading and required further investigation to determine their exact behavior. Milling studies at

varying media size and rotation rate showed that the 1.5 mm milling particles resulted in the

highest amount of deformation. Rotation rate was found to affect the rate of particle deformation









and had little to no influence over the maximum attainable deformation. Hertz's theory was used

to explain the relationship between rotation rate, media size and particle deformation.

Hertz's theory was used to determine the stress acting on the media, the theory was then

expanded to determine the stress acting on the particle by calculating the number of particles in

the contact area between colliding milling media. The data indicated that as media sizes

decreased the stress acting on the particle increased. The 1.5 mm media resulted in the highest

amount of deformation, due to the stress being close to the yield stress of the aluminum material.

The media sizes smaller than 1.5 mm caused stresses greater than the yield stress and thus

resulted in an early onset of particle breakage. Models that describe media frequency of contacts

failed to describe the rotation rate effects on deformation observed in this study. New models

were derived that demonstrated that media size had no effect on the number of times that

particles were stressed during the experiment. The new models also determined that rotation rate

effect the number of times a particle was stressed during the milling experiments. The empirical

data agreed with the results of the rotational rate models, however, further experimentation

would be required to obtain a strong validation of the rotation rate models.

The strain energy was calculated from assumptions of the conservation of volume of the

particles during deformation. For the first time it was possible to calculate milling efficiency

based on the strain energy. The study found that milling efficiency was the greatest for the 1.5

mm media and that efficiency decreased with milling time. A rotation rate of 2000 rpm resulted

in the highest efficiency at the point of particle breakage. The study also determined the reasons

for inefficiency during the milling processes. Previous studies had proposed that viscous friction

and heat generated in the milling process were the leading cause of inefficiency. This study

found that elastic collision which resulted in no deformation led to the inefficiency.









Statistical design of experiment was used to develop an empirical model capable of

predicting particle deformation during milling based on rotation rate and media size. The model

also validated the fundamental theories used to explain the behavior of milling parameters on

deformation. The statistical design showed that the interactions between media size and rotation

rate had a minimal effect on deformation. This result was also obtained by the fundamental

theory based on contact mechanics which showed the stress acting on the particle only varied

slightly with rotation rate. The statistical design predicted that a media size of 1.5 mm would

yield the most strain in the particle.

For the first time relationships between particle deformation and milling parameters based

on fundamental theories were established using Hertz's theory and collision frequency. The

relationship between milling parameters and the stress acting on the particle could be used as a

scaling parameter to improve and design milling processes. Results of this investigation showed

that it was possible to predict milling efficiency, which could be directly applied to many milling

operations currently in use. The case study proved that the understanding developed in this study

could be directly applied to developing products used in real world applications.

Future Work

This study developed a knowledge base for milling processes where particle deformation is

desired. However, most milling processes involve particle breakage, a suggestion for future work

would be to determine correlation between the models developed in this study and the particle

breakage models. An example of this is that many researches have recognized the non linear

breakage rate kinetics in a milling process [83, 84]. A similar relation would exist between stress

acting on a particle and particle size as a function of milling time. As particle size decreases in a

breakage process, the number of particles in the contact area would increase, thus decreasing the

stress acting on the particle. It is also known that as particle size decrease the number of defects









in the particle also decreases, requiring higher stress to fracture the particle. These results could

be the reason that a so called "grinding limit" is observed in many milling process. One method

of testing the applicability of the models developed in this study to breakage processes would be

to reduce the number concentration of particles during the milling process, thus maintaining a

constant level of stress acting on the particles.

Another area of interest for future work would be to investigate the effect of media density

on particle deformation. Density has a direct effect on the stress acting on the particle, it would

be interesting to determine if the effect on deformation is as predicted by the models. One area,

which requires more study, is that of the frequency of media and particle contacts. The results

obtained in this study were not sufficient and conclusive enough to make up a significant theory.

More empirical data and short sampling intervals would be required to obtain a more robust

model capable of describing the effect of media size on frequency of particle contact.

One of the most novel aspects of this study was the ability to obtain an actual value for

milling efficiency. It is possible to test this method of measuring milling efficiency by

manipulating other milling variables such as viscosity and determine the effect on energy

utilization. It would also be interesting to further investigate the mechanisms of milling

efficiency such as the result of elastic collisions and their effect on inefficiency. Further work is

needed to construct milling models that are capable of predicting deformation of materials based

on milling parameters and material properties.

Finally, the goal of this study was to develop a fundamental understanding of particle

deformation during milling and develop a predictive model. In most respects this goal was

accomplished, however, it is important to develop a model that includes a time parameter. It may









also be possible to construct a model that includes a material property parameter and properties

of milling media, such as, media density, and hardness.









LIST OF REFERENCES

1. American Society for Metals International, Powder Metal Technologies and
Applications, Materials Park, Ohio, 1998.

2. F. Muller, R. F. Polke, Powder Technology 105 (1999) 2-13.

3. M. Senna, H. Kuno, Journal of the American Ceramic Society 54 (1971) 259-&.

4. P. S. Gilman, J. S. Benjamin, Annual Review of Materials Science 13 (1983) 279-300.

5. S. H. Hong, B. K. Kim, Materials Letters 51 (2001) 139-143.

6. P. Wissling, dmg Business Media Ltd (1999).

7. J. Embury, Emissive versus attenuating smokes, in: R. a. T. Directorate (Ed.) Edgewood
Chemical and Biological Center, 2002.

8. J. Embury, Maximizing infrared extinction coefficients for metal discs, rods, and spheres,
in: R. a. T. Directorate (Ed.) Edgewood Chemical and Biological Center, 2002.

9. D. M. Bigg, D. E. Stutz, Polymer Composites 4 (1983) 40-46.

10. C. B. Keemer, in: Communication (Ed.) Siblerline Manufacturing Co., Inc., 2005.

11. H. R. Zoz, R. Reichardt, H.U. Benz, A. Nadkami, G. Wagner, Advances in Powder
Metallurgy and Particulate Materials (1999).

12. J. Y. Huang, Y. K. Wu, H. Q. Ye, Materials Science and Engineering a-Structural
Materials Properties Microstructure and Processing 199 (1995) 165-172.

13. S. Hwang, C. Nishimura, P. G. McCormick, Scripta Materialia 44 (2001) 1507-1511.

14. R. Watanabe, H. Hashimoto, G. G. Lee, Materials Transactions Japan Institute of Metals
36 (1995) 102-109.

15. G. G. Liversidge, K. C. Cundy, International Journal of Pharmaceutics 125 (1995) 91-97.

16. C. Frances, C. Laguerie, B. Mazzarotta, T. Veccia, Chemical Engineering Journal 63
(1996) 141-147.

17. S. Mende, F. Stenger, W. Peukert, J. Schwedes, Powder Technology 132 (2003) 64-73.

18. S. Mende, F. Stenger, W. Peukert, J. Schwedes, Journal of Materials Science 39 (2004)
5223-5226.

19. L. G. Austin, P. Bagga, Powder Technology 28 (1981) 83-90.









20. D. Eskin, O. Zhupanska, R. Hamey, B. Moudgil, B. Scarlett, Aiche Journal 51 (2005)
1346-1358.

21. A. Jankovic, Minerals Engineering 16 (2003) 337-345.

22. A. S. Kwade, J., KONA 52 (1997) 91-101.

23. E. Bilgili, Scarlett, B., Estimation of the selection and breakage parameters from batch
grinding: a novel full numerical scheme, in: Proceedings of the AIChE Annual Meeting,
San Francisco, 2003.

24. J. A. Herbst, Y. C. Lo, International Journal of Mineral Processing 44-5 (1996) 209-221.

25. A. Kwade, Powder Technology 105 (1999) 14-20.

26. H. Rumpf, Particle Technology, chapman and hall, 1990.

27. W. S. McCabe, J Harriott, P., Unit Operations of Chemical Engineering, McGraw-Hill
inc, New York, 1993.

28. C. L. Prasher, Crushing and Grinding Process Handbook, wiley, 1987.

29. R. Pond, Getting metals into usable form: Cold working, in: Metallurgy for non-
metallurgist, ASM International, Metals Park.

30. R. A. R. Reed-Hill, Physical Metallurgy Principles, pws publishing company, 1994.

31. W. Callister, Materials Science and Engineering an Introduction, John Wiley and Son,
New York, 2000.

32. A. J. R. Inigoraj, R. Narayanasamy, K. S. Pandey, Journal Of Materials Processing
Technology 84 (1998) 143-148.

33. S. Lou, D. O. Northwood, Journal Of Materials Engineering And Performance 3 (1994)
344-349.

34. H. J. Kleemola, M. A. Nieminen, Metallurgical Transactions 5 (1974) 1863-1866.

35. R. P. Reed, Cryogenics 12 (1972) 259-&.

36. T. G. Nieh, P. S. Gilman, J. Wadsworth, Scripta Metallurgica 19 (1985) 1375-1378.

37. C. Tome, G. R. Canova, U. F. Kocks, N. Christodoulou, J. J. Jonas, Acta Metallurgica 32
(1984) 1637-1653.

38. M. A. Meyers, A. Mishra, D. J. Benson, Progress in Materials Science 51 (2006) 427-
556.

39. J. A. Hines, K. S. Vecchio, Acta Materialia 45 (1997) 635-649.









40. C. R. Bemrose, J. Bridgwater, Powder Technology 49 (1987) 97-126.

41. Y. S. Cheong, A. D. Salman, M. J. Hounslow, Powder Technology 138 (2003) 189-200.

42. H. Cho, M. A. Waters, R. Hogg, International Journal of Mineral Processing 44-5 (1996)
607-615.

43. K. Schonert, International Journal of Mineral Processing 44-5 (1996) 1-16.

44. K. Schonert, in: Summer school in winter, Particle Engineering Research Center,
Gainesville, Florida, 2003.

45. K. Kendall, Nature 272 (1978) 710-711.

46. K. U. Schonert, H. Kleman, W., in: Proceedings of the 2nd International Conference of
Fracture, Brighton, England, 1969, pp. 474-482.

47. K. Schonert, About the particle breakage and the possibities and limits for improving
size-reduction processes, in: The International Symposium on Powder Technology,
Powder Technology, Kyoto, Japan, 1981.

48. J. S. Benjamin, Metallurgical Transactions 1 (1970) 2943-&.

49. J. B. Fogagnolo, F. Velasco, M. H. Robert, J. M. Torralba, Materials Science And
Engineering A-Structural Materials Properties Microstructure And Processing 342 (2003)
131-143.

50. S. X. Mao, N. A. McMinn, N. Q. Wu, Materials Science And Engineering A-Structural
Materials Properties Microstructure And Processing 363 (2003) 275-289.

51. S. K. Moothedath, K. V. S. Sastry, Powder Technology 75 (1993) 89-96.

52. A. W. Weeber, A. J. H. Wester, W. J. Haag, H. Bakker, Physica B & C 145 (1987) 349-
352.

53. J. Guerrero-Paz, D. Jaramillo-Vigueras, Nanostructured Materials 11 (1999) 1123-1132.

54. Booz, Metal Flake Production, in: US (Ed.) vol 4065060, US, 1976.

55. M. M. Moshksar, S. M. Zebarjad, Iranian Journal of Science and Technology 23 (1999)
239-250.

56. H. R. H. Zoz, R. Reichardt, H.U. Benz, A. Nadkami, G. Wagner, Advances in Powder
Metallurgy and Particulate Materials (1999).

57. L. P. Sung, M. E. Nadal, M. E. McKnight, E. Marx, B. Laurenti, Journal of Coatings
Technology 74 (2002) 55-63.

58. S. H. Hong, B. K. Kim, Materials Letters 57 (2002) 275-279.









59. M. Riffel, J. Schilz, Journal of Materials Science 33 (1998) 3427-3431.

60. J. A. Rodriguez, J. M. Gallardo, E. J. Herrera, Journal of Materials Science 32 (1997)
3535-3539.

61. A. Jankovic, Minerals Engineering 14 (2001) 1177-1186.

62. T. Tanaka, KONA 13 (1995).

63. L. Blecher, A. Kwade, J. Schwedes, Powder Technology 86 (1996) 59-68.

64. S. Tedeschi, Improving aerosol dispersion through a fundamental understanding of
interparticle forces, in: Materials Science and Engineering, vol Ph. D., University of
Florida, Gainesville, 2007.

65. S. Brunauer, P. H. Emmett, E. Teller, Journal Of The American Chemical Society 60
(1938) 309-319.

66. T. Allen, Particle Size Measurements, Chapman & Hall, London, 1997.

67. D. W. Fuerstenau, A. Z. M. Abouzeid, International Journal of Mineral Processing 31
(1991) 151-162.

68. M. Z. He, Y. M. Wang, E. Forssberg, Powder Technology 147 (2004) 94-112.

69. J. Zheng, C. C. Harris, P. Somasundaran, Powder Technology 86 (1996) 171-178.

70. M. Z. He, Y. M. Wang, E. Forssberg, Powder Technology 161 (2006) 10-21.

71. R. Hamey, Production of organic pigment nanoparticles by stirred media milling, in:
Materials Science and Engineering, vol Masters of Science, Florida, Gainesville, 2003.

72. A. Kwade, L. Blecher, J. Schwedes, Powder Technology 86 (1996) 69-76.

73. A. H. Tullo, Chemical & Engineering News 82 (2004).

74. K. L. Johnson, Contact mechanics, Cambridge University Press, 2003.

75. R. A. Broadbent, Callcott, T.G., Journal of Institute of Fuel 29 (1956) 528-539.

76. S. Nagata, Mixing: Principles and Applications, Wiley, New York, 1975.

77. C. Alamprese, L. Datei, Q. Semeraro, Journal Of Food Engineering 83 (2007) 629-636.

78. D. C. Montgomery, Design and Analysis of Experiment, John Wiley and Son Inc, New
York, 2004.

79. J. Embury, Technical Report US Army Edgewood Chemical and Biological Center
(1993).









80. J. Embury, Technical Report US. Chemical Systems Laboratory (1994).

81. J. Embury, Technical Report US Army Edgewood Chemical and Biological Center
(1994).

82. J. Embury, Technical Report US Army Edgewood Chemical and Biological Center
(1998).

83. L. G. Austin, Powder Technology 5 (1971)

84. E. Bilgili, Jepes, J., Scarlett, B., A non-linear framework for population balance
modeling of batch grinding: beyond first-order kinetics., in: Proceedings of the 2nd
International Conference on Population Balance Modeling, Valencia, Spain, 2004b.









BIOGRAPHICAL SKETCH

Rhye Garrett Hamey was born in Medina, Ohio. He attended Cloverleaf High

School. After high school he obtained a degree in chemical engineering at The Ohio State

University. While at Ohio State, Rhye attended a class on particle technology instructed

by visiting Professors Brian Scarlett and Dr. Brij Moudgil. Through continued contact

with Professor Scarlett, Rhye became interested in particle technology and chose to

pursue a graduate degree from the Department Materials Science and Engineering at the

University of Florida. He performed his study while working at the Particle Engineering

Research Center. In 2005, Rhye obtained a Master's degree in Materials Science and

Engineering. After receiving his Master's Rhye decided to continue his Ph.D. study under

the supervision of Dr. Hassan El-Shall.


154





PAGE 1

1 PARTICLE DEFORMATION DU RING STIRRED MEDIA MILLING By RHYE GARRETT HAMEY A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Rhye Garrett Hamey

PAGE 3

3 My family and fianc Yun Mi Kim, for their encouragement

PAGE 4

4 ACKNOWLEDGMENTS I would like to acknowledge the friendship and the wisdom of the late Professor Brian Scarlett, whose instruction and support were the reasons for the pursuit of a Ph. D. Dr. Brian Scarlett, taught me how to be a better engineer and a better person. I am grateful for the help of Dr. Hassan El-Shall for his count less hours of guidance through out this study. Dr. El-Shall was a constant inspiration. I would al so like to thank Dr. Mecholsky, Dr. Fuchs, Dr. Whitney, and Dr Svor onos for serving as committee members and for their guidance and discussions during this Ph..D. study. A special thanks goes out to all the frie nds I have made during this study, Maria Palazeulos, Scott Brown, Milorad Djomlija, Vija y, Krishna, Stephen Tedeschi, Nate Stevens, Kerri-Ann Hue, and Dauntel Specht. I would lik e to acknowledge Marco Verwijs for being a great friend and roommate for so many years. I would also like to th ank Yun Mi Kim for her love and support. I would like to acknowledge the financial support of the Par ticle Engineering Research Center (PERC) at the University of Florid a, The National Science Foundation (NSF) and the Industrial Partners of the PERC for support of this research. Th anks are extended to the US Army for their support of the project.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........8 LIST OF FIGURES.......................................................................................................................10 ABSTRACT...................................................................................................................................14 CHAP TER 1 INTRODUCTION..................................................................................................................16 Impact of Milling Metallic Powders....................................................................................... 16 Research of Particle De form ation during Milling.................................................................. 18 Method for Understanding Particle Defor mation during Milling........................................... 18 2 BACKGROUND.................................................................................................................... 21 Comminution Equipment........................................................................................................ 22 Crushers...........................................................................................................................23 Grinders....................................................................................................................... ....23 Ultrafine Grinders............................................................................................................ 24 Cutting Machines.............................................................................................................24 Particle Deformation and Particle Breakage........................................................................... 24 Particle Deformation....................................................................................................... 24 Particle Breakage.............................................................................................................27 Cracks and Defects.......................................................................................................... 28 Particle Deformation during Milling............................................................................... 29 Effect of Milling Parameters on Particle Deformation.................................................... 31 Particle Breakage during Milling....................................................................................32 Effect of Milling Parameters on Particle Breakage......................................................... 33 Summary..........................................................................................................................35 3 MATERIALS, CHARACTERI ZATION AND EXPERIMENTAL PROCEDURE ............. 41 Materials.................................................................................................................................41 Metal Powders.................................................................................................................41 Milling Media..................................................................................................................41 Experimental Procedures........................................................................................................ 42 Media Mill.......................................................................................................................42 Milling Procedure............................................................................................................42 Drying of Aluminum Slurry and Milling Samples.......................................................... 43 Sample Preparation by Dispersion of Dry Powder..........................................................44 Statistical Design of Experiment............................................................................................ 45

PAGE 6

6 Characterization......................................................................................................................45 Light Scattering...............................................................................................................46 Microscopy and Image Analysis..................................................................................... 47 Electron microscopy................................................................................................. 47 Optical microscopy.................................................................................................. 48 Occhio optical particle sizer..................................................................................... 48 Surface Area Analysis (BET).......................................................................................... 48 Inductively Couple Plasma Spectroscopy (ICP)............................................................. 49 Fourier Transform Infrared Spectroscopy (FTIR)...........................................................49 Light Scattering Results..........................................................................................................50 Results of Aluminum Powder......................................................................................... 50 Results for Alum inum Flake........................................................................................... 51 Summary of Characterizati on and L ight Scattering........................................................ 52 Milling Condition Selection...................................................................................................52 Preliminary Milling Study............................................................................................... 52 Preliminary Milling Results............................................................................................ 53 Establishing Milling Conditions......................................................................................54 Milling medium........................................................................................................ 54 Mill loading and grinding aids................................................................................. 55 Milling time and temperature................................................................................... 56 Media properties a nd rotational rate .........................................................................57 4 RESULTS AND DISCUSSION............................................................................................. 72 Experimental Results of Stirred Media Milling...................................................................... 72 Effect of Mill Parameters................................................................................................ 72 Kinetic Energy Model.....................................................................................................73 Stress Model....................................................................................................................75 Deformation Rate............................................................................................................79 Strain Energy and Milling Efficiency.....................................................................................81 Strain Energy...................................................................................................................81 Mill Efficiency................................................................................................................ .83 Empirical Modeling of Particle Deformati on through Statistical Design of Experim ent...... 88 Microstructure Analysis..........................................................................................................92 Summary.................................................................................................................................93 5 CASE S TUDY: MA T E RIAL SEL E CT I ON AND DEVELOPMENT FOR INFRARED OBSCURANTS ....................................................................................................................127 Introduction................................................................................................................... ........127 Development of New Ob scurant Materials .......................................................................... 128 Material Characterization and Obscur ant Perform ance Measurement Method................... 131 Production of Infrared Obscurant......................................................................................... 132 Summary...............................................................................................................................135

PAGE 7

7 6 SUMMARY, CONCLUSIONS, AND FUTURE W ORK................................................... 143 Summary and Conclusions...................................................................................................143 Future Work..........................................................................................................................146 LIST OF REFERENCES.............................................................................................................149 BIOGRAPHICAL SKETCH.......................................................................................................154

PAGE 8

8 LIST OF TABLES Table page 3-1 Statistical data comparison for light s cattering and im age analysis performed using the Occhio particle counter................................................................................................ 66 3-2 Stirred media milling eq uipment variables........................................................................ 67 3-3 Stirred media milling operating variables.......................................................................... 67 3-4 Experimental parameters used in the hexagonal statistical design of experim ent.............67 3-5 Milling conditions for preliminary study........................................................................... 68 4-1 Calculated stress intensity for experi m ents performed in this study. The table indicates that it is possible to achieve equivalent energies and different milling conditions.........................................................................................................................101 4-2 Maximum mean particle size of milling experim ents. Even though the kinetic energy of some of the milling experiments are equivalent the maximum particles sizes differ.................................................................................................................................101 4-3 Force, radius of contact area, and stre ss acting o n the milling media as a function of media size and rotation rate (cal culated using Hertz theory)...........................................103 4-4 Number of particles stressed in a sing le collis ion between grinding media and the stress exerted on an individual particles.......................................................................... 103 4-5 Stress frequency as calculated from Kwade s m odel. The frequency is the number of times milling media collid e with each other.................................................................... 103 4-6 The time required to reach the maximu m particle diameter. As milling speed increases the time it takes to obtain the maximum particle diameter decreases.............. 104 4-7 Total number of particle compressions in 60 m inutes, as calculated using stress frequency and number of particles, Equation 4-2 and 4-6............................................... 104 4-8 Strain energy per par ticle for deformation study. ............................................................107 4-9 Total strain energy for deformation s tudy. The 1.5 mm media milling experiments resulted in the largest amount of stain energy................................................................. 108 4-10 List of constants for equations us ed to calculate milling power...................................... 108 4-11.Milling efficiency as a function of pe rcent amount of strain the m aterial has experienced at 1000 rpm and varying media size. The 1.5 mm media reaches over 60 % of the maximum strain while milling at 37% efficiency. The last few percent of the maximum strain result in the most inefficient milling............................................... 113

PAGE 9

9 4-12 Experimental design for cen tral com posite design 1....................................................... 114 4-13. Analysis of variance for the central co m posite design 1 described in table 4-12.............. 114 4-14 Experimental design for cen tral com posite design 2....................................................... 118 4-15 Analysis of variance for the central com posite design 2 described in table 4-10............ 118 5-1 Experimental parameters used in the statistical design of experim ent used to determine milling parameters.......................................................................................... 137 5-2 Experimental parameters used in the full 2 factorial (22) statistical design of experim ent..................................................................................................................... ...139

PAGE 10

10 LIST OF FIGURES Figure page 1-1 Methodology for understanding particle de form ation during stirred media milling......... 20 2-1 1) Particle stressed by compression or sh ear, 2) particle stressed by im paction on rigid surface, and 3) particle stressed in shear flow...........................................................36 2-2 Milling efficiency vari ations by m achine type..................................................................36 2-3 Types of grinding mills.................................................................................................... ..37 2-4 Diagram of a stirred media mill, energy is supplied to the grinding media by the rotation of the agitator. ...................................................................................................... .37 2-5 Face centered unit cell for aluminum................................................................................. 38 2-6 Slip planes and directions for FCC aluminum................................................................... 38 2-7 True stresstrue strain curve for a mate rial undergoing plastic de form ation and strain hardening............................................................................................................................39 2-9 Effect of impact..................................................................................................................40 3-1 Particle size distribution of as receiv ed H-2 aluminum, along with particle size data, obtained from Coulter LS 11320.......................................................................................59 3-2 Image of Yttria-stabilized zirconia grinding m edia obtained form an optical microscope, 1mm in diameter............................................................................................ 60 3-3 Union Process stirred m edia mill us e in study................................................................... 61 3-4 Phase diagram for carbon dioxide, indicat ing the region of supercritical fluid for mation............................................................................................................................62 3-5 SPI pressure vessel used for supercri tical drying of powder samples taken from milling........................................................................................................................ ........63 3-6 Galia partial vacuum chamber used to disseminate powder.............................................. 63 3-7 Scanning electron microscope image of as receiv ed H-2 aluminum powder....................64 3-8 Differential volume and nu m ber distributions for as received H-2 spherical aluminum powder measured by light scattering................................................................ 65 3-9 Volume particle size distribution of m illed aluminum flake as measure by light scattering............................................................................................................................65

PAGE 11

11 3-10 Volume particle size distribution of m illed alum inum flake as measured by the Occhio image analysis....................................................................................................... 66 3-11 SEM image of US Bronze as received brass flak e used in preliminary milling experiments........................................................................................................................68 3-12 Surface response of figur e of m erit with respect to media loading and RPM.................. 69 3-13 Stearic acid a saturated fatty acid found in m any animal fats and vegetable oils..............69 3-14 Cumulative undersized versus particle size as at v arying milling time for H-2 aluminum....................................................................................................................... ....70 3-15 Scanning electron microscope images of a) as received alum inum, b) milled for 240 minutes aluminum and c) milled for 600 minutes aluminum............................................ 71 4-1 Volume percent particle size distribu tions for experim ents using 1.0 mm grinding media at 2000 rpm. Bimodality in the part icle size distribution is indicative of particle deformation and particle breakage occurring simultaneously.............................. 95 4-2 Volume percent particle size distributi ons for the experim ents with 1.5 mm grinding media at 1000 rpm, a visible shift in the par ticle size distribution to larger sizes is seen due to particle deformation........................................................................................ 96 4-3 Maximum mean particle size achieved for each m illing experiment versus media size, the 1.5 mm milling media produced th e largest amount of deformation................... 97 4-4 Particle deformation versus milling time at a m illing speed of 1000 rpm for varying grinding media sizes.......................................................................................................... 98 4-5 Particle deformation versus milling time at a m illing speed of 1500 rpm for varying grinding media sizes.......................................................................................................... 99 4-6 Particle deformation versus milling time at a m illing speed of 2000 rpm for varying grinding media sizes........................................................................................................ 100 4-7 Mean particle size versus milling tim e as a function of rotation rate.............................. 101 4-8 Contact between colliding elastic spheres. Aluminum particles are caught in the contact area between the media....................................................................................... 102 4-9 Diagram of the change in dimensions of particle and the m easurements used to calculate stress and strain................................................................................................. 104 4-10 Stress-strain behavior for e xperim ents performed at 1000 rpm....................................... 105 4-11 Stress-strain behavior for e xperim ents performed at 1500 rpm....................................... 106 4-12 Stress-strain behavior for e xperim ents performed at 2000 rpm....................................... 107

PAGE 12

12 4-13 A plot of milling efficiency as a functi on of m illing time at 1000 rpm, for all media sizes..................................................................................................................................109 4-14 A plot of milling efficiency as a functi on of m illing time at 1500 rpm, for all media sizes..................................................................................................................................110 4-15 A plot of milling efficiency as a functi on of m illing time at 2000 rpm, for all media sizes..................................................................................................................................111 4-16 A plot of milling efficiency as a func tion of m illing time for the 1.5 mm at varying rotational rates............................................................................................................... ...112 4-17 A plot of milling efficiency at failure, righ t side of figure 4-13 used for comparison of efficiencies as a function of rotational rate.................................................................. 113 4-18 Contour plot for central composite design 1, a m inimum strain of 2.17 can be seen at approximately 0.9 mm milling media.............................................................................. 115 4-19 Surface response for central composite design 1, a trough ex ists at the 1.0 mm media size........................................................................................................................... ........116 4-20 Standard error associated with central com posite design 1, error increases at the limits of the study............................................................................................................ .117 4-21 Contour plot for central composite design 2, shows a m aximum at approximately the 1.6 mm media................................................................................................................... 119 4-22 Surface response for central composite design 2, a peak can be seen at intermediate m edia size..................................................................................................................... ....120 4-23 A plot of the interaction between the medi a size and rotation rate, it can be seen that there is only a slight interaction at low. ........................................................................... 121 4-24 Standard error associated with central composite design 2............................................. 122 4-25 Transmission electron microscope image of as received alum inum particle............... 123 4-26 Diffraction pattern obtained fr om particle in Figure 4-20...............................................124 4-27 Transmission electron microscope image of sectioned flake after 4 hours of milling.... 125 4-28 Single crystalline diffraction pattern of sectioned flake after 4 hour sof milling............ 126 5-1 A chart of relevant material properties of m aterials for use as an infrared obscurant. High values in each property are preferred for candidate materials................................ 136 5-2 The performance of various milled metals, the copper has the highest performance to due to the properties being the hi ghest in all the desired area. ........................................ 137

PAGE 13

13 5-3 The results of the initial design of expe rim ent, indicating that higher media loading increase the performance of the material......................................................................... 138 5-4 Plot of FOM for as received and ground brass at specific m ill ing condition of 480ml media volume and 700 rp m rotational speed................................................................... 139 5-5 The results of the affect of varying me dia size and rotation rate on the perform ance of a milled brass flake...................................................................................................... 140 5-6 Differential volume particle size di stribution at in creasing milling times....................... 141 5-7 Figure of merit versus IR wave length for increas ing milling times................................ 141 5-8 Mean particle size as a function of milling time for different density milling media, zirconia (6 g/cc), stainless steel ( 9.8 g/cc), and tungsten carbide (16 g/cc). ................... 142

PAGE 14

14 Abstract of Dissertation Pres ented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy PARTICLE DEFORMATION DU RING STIRRED MEDIA MILLING By Rhye Garrett Hamey August 2008 Chair: Hassan El-Shall Major: Materials Science and Engineering Production of high aspect ratio metal flakes is an important part of the paint and coating industry. The United States Army also uses high as pect ratio metal flakes of a specific dimension in obscurant clouds to attenuate infrared radiation. The most common method for their production is by milling a metal powder. Ductile metal particles are initia lly flattened in the process increasing the aspect ratio. As the proces s continues, coldwelding of metal flakes can take place increasing the particle size and decr easing the aspect ratio. Extended milling times may also result in fracture leading to a further d ecrease in the particle si ze and aspect ratio. Both the coldwelding of the particles a nd the breakage of the particles are ultimately detrimental to the materials performance. This study utilized characteriza tion techniques, such as, light scattering and image analysis to determine the change in particle size as a function of milling time and parameters. This study proved that a fundamental relati onship between the milling parameters and particle deformation could be established by using Hertzs theory to calculate the stress acting on the aluminum particles. The study also demonstr ated a method by which milling efficiency could be calculated, based on the amount of energy required to cause particle deformation. The study found that the particle deformation process could be an energy efficient process at short milling times with milling efficiency as high as 80%. Finally, statistical design of experiment was used

PAGE 15

15 to obtain a model that related particle deformation to milling parameters, such as, rotation rate and milling media size.

PAGE 16

16 CHAPTER 1 INTRODUCTION In todays market there is a need to develop new and im proved products at an increasing rate. Particle deformation by millin g is important to many areas of product development, such as, powder metallurgy (P/M), paints and coatings, and military applications. Understanding and predicting particle deformation during milling, will lead to more rapid development of products in which these materials are used. Impact of Milling Metallic Powders Generally, m illing is used to reduce the size of a material or for the processing of brittle materials. However, large quant ities of ductile metals used in the powder metallurgy industry (P/M) are processed through milling [1]. Powder metallurgy is estimated to be a $1.8 billion industry annually with an estimated 70 % spent in the automotive industry [2]. Below is a list of reasons for the milling of materials in the P/M industry [1]: Particle size reduction (comminuti on or grinding) for sintering Particle size growth Shape change (flaking) Agglomeration Solid-state alloying (m echanical alloying) Solid-state blending (incomplete alloying) Modifying, changing, or altering properties of a material (density, flowability, or work hardening) Mixing or blending of two or mo re materials or mixed phases Nonequilibrium processing of metastable phase s such as amorphous alloys, extended solid solutions, and nanocrystalline structures

PAGE 17

17 During the processing and subsequent stressi ng of metallic powders, several phenomena may occur; the metal can fracture, deform, co ldwork, coldweld, or transform into other polymorphs [3]. Coldwelding is used to mechani cally alloy (MA) and refine the microstructure of immiscible metals [4]. Coldworking is used to strengthen metallic pa rticles. This study will focus on deformation of metallic particles during milling. This technique is used to make metallic flakes for paints, coating, and obscurants [5-9 ]. The metallic flakes in paints give what is termed a metallic look which means the randomly oriented flak es in the paint will give a shiny appearance. Radiation shielding in devices such as cell phones will use metallic particles to form a conductive coating. The US Army uses metallic flakes in smoke bombs to obscure infrared light (IR). The conductive metallic par ticles are capable of a ttenuating the IR signal which provides a shield against mi ssiles and IR imaging devices. The most common metallic pigment is aluminum, which when milled deforms into flakes with highly reflective surfaces. In order for the metallic flakes to be effective at scattering and reflecting light, the particle size, size distribution, particle sh ape, and particle morphology must be controlled. The particle size and size distribution affect the optical reflectance and light scattering properties of the flake. If the particle size distribution of these paints is broad the reflectiveness will be diminished and the pain t will appear dull. The particle morphology and surface roughness also affect the optical refl ectance i.e. the smoothe r the surface the more reflective the pigment. An additional difficulty to this processing method is that most of the aluminum pigments are produced in batch milling processes and there exists large batch to batch variations due to lack of process knowledge [10] In the end, these variations lead to additional processing costs from classification.

PAGE 18

18 Research of Particle Deformation during Milling Most m illing studies focus on particle br eakage with limited research on particle deformation. The particle deformation studies ha ve focused on the mechanical alloying process and do not investigate optimizing the flake dime nsions, which are necessary for paints and coatings. Zoz showed how the dimensions of fl akes are affected by the material hardness[11]. Several researchers have investigat ed the change in the microstructu re and material properties of metals during milling [12-14]. None of these studies has shown the effect of milling parameters on particle deformation. The research to date does not provide any methodology for predicting the magnitude or extent of particle deform ation during milling. In order to improve the production and quality of the products it is im portant to understand the effects of milling parameters on the materials being processed. Even though particle breakage during milling has been studied extensively, ther e is very little knowledge of th e particle deformation process during milling. This study will attempt to apply some of the knowledge gained from particle breakage during milling to the deformation process and to fill the gaps in the theories between the two processes. Breakage during milling has been studied at n early every length scale [15-19]. Numerous researchers have modeled milling kinetics and energetics for particle breakage [20-22]. Others have used population balance modeling to pred ict particle size dur ing milling [19, 23]. The response of the product material to nearly ever y milling parameter has been investigated for particle breakage. Wear on equipment and process optimization studies have also been researched [24, 25]. Method for Understanding Partic le Deformation during Milling The goal of this study is to understand and predict particle deformation during stirred media milling. Figure 1-1 shows the methodology that was used to accomplish this goal. The

PAGE 19

19 study investigated the affects of mill speed and media size on the particle size and morphology of an aluminum powder. The study shows how milling breakage models fail to describe particle deformation during milling. New models were de rived from the stress-strain behavior of aluminum during milling to describe particle deformation. Hertz theory was used to calculate the stress acting on a particle during. The stress acting on a pa rticle and the affect it had on particle deformation was then determined. Statistical analysis was used to determine the magnitude of the interaction between milling parameters. Design of experiment was used to construct a model that relates milling parameters to particle deformation. From these studies a more comprehensive understanding of particle deformation during milli ng was achieved. This may ultimately lead to a method to better predict and optimize milling proce sses where the deformation of metallic flakes is desired.

PAGE 20

20 Figure 1-1. Methodology for understanding particle deformation during stirred media milling.

PAGE 21

21 CHAPTER 2 BACKGROUND Comm inution is arguably the ol dest material processing tech nique. Earlier civilizations employed simple mortars and pestles to reduce th e size of agricultural and medicinal products. Even though there have been sign ificant changes in technology, th ese simple devices that were employed thousands of years ago can still be foun d in nearly every lab today. A simple definition of comminution is the reduction in particle size through the application of an applied load. In order to study and improve a milling process it is important to understand how the load is applied and to analyze the response of the material to that load. Rumpf suggested four modes by which stress can cause particle breakage [26]: 1. Compression-shear stressing 2. Impact stressing 3. Stressing in shear flow 4. Stressing by other methods of energy transmission (electrical, chemical, or heat) Figure 2.1 is a representation of the modes of stressing a mate rial as indicated by Rumpf. The first two modes are caused by contact for ces and are important for milling. Mode 3 is important for the dispersion of materials under sh ear, and only exerts enough force to break apart weak agglomerates [26]. The fourth mode is not applicable to milling where mechanical stressing is important. There are several types of machines designed to apply stress to a material by modes 1 and 2. The beginning of this chapter will describe the types of milling equipment used today. A review of literature on single and multi particle breakage will be presented. Followed by a discussion of the response of a material to the load applied during milling. This information will be used as a basis for this study and it will sh ed light on the deformation process during milling, which has not been studied as thoroughly. This chapter will conclude by discussing material deformation and reviewing deformation during milling processes. Ultimately, this study will

PAGE 22

22 compile what has already been done to study pa rticle deformation during milling and use that knowledge, combined with particle breakage and further experimentation, to more accurately develop the understanding of particle deformation during milling. Comminution Equipment Comm inution is a diverse process, found in nearly every industry. Therefore there are numerous types of milling equipment. The equipm ent must be able to operate in rigorous, high throughput environments of the mineral processing industry and in the ultra pure, contamination free pharmaceutical industry. McCabe, Smith and Harriott identified four classes of milling equipment each with subclasses [27]: A. Crushers (coarse and fine) a. Jaw crushers b. Gyratory crushers c. Crushing roll B. Grinders (intermediate and fine) a. Hammer mills, impactors b. Rolling-compression mills i. Bowl mills ii. Roller mills c. Attrition mills d. Tumbling mills i. Rod mills ii. Ball mills; pebble mills iii. Tube mills; compartment mills C. Ultrafine grinders a. Hammer mills with internal classification b. Fluid-energy mills c. Agitated mills D. Cutting machines

PAGE 23

23 a. Knife cutters; dicers; slitters The differences in equipment are due to the mode in which stress is applied to the material and the final achievable particle size. Some of the equipment may work in clean environments and process small quantities of products while other may process large amounts of material in wet and dry environments. Schonert has shown that some of these machines can operate at high energy efficiencies >90% while others operations are a fraction of a percent efficient. Most of this inefficiency is associated with the increas e in energy required to achieve smaller particle size. Examples of the efficiency of select mills can be seen in Figure 2-2, obtained from tabulated data given by Prasher for the energy consumption associated with milling to a specific size [28]. Crushers Jaw crusher s, gyratory crushers and crushing roll mills are used to process large quantities of materials. They are often used in the mine ral processing and cement industries to reduce the size of rock and ore for further processing. Grinders Grinders a re one of the most diverse sets of milling equipment. They are most often used to reduce large aggregates to a powder. Gri nding mills can supply a load to a material by impaction, attrition, or compression. Hammer m ills apply a load via impaction by dropping a weight repeatedly on a bed of material. Attritio n mills wear away a material by the breakage of small fragments from the surface. Roll mills and ball mills apply a load through compression by trapping material between colliding media or in between two heavy rollers. Figure 2-3 is a diagram of a ball mill and a roller mill. Ball mills have been the subject of a considerable amount of research because they can be used in wet and dry environments and their large capacities. Ball

PAGE 24

24 mills operate by rotating a drum containing hard dense balls and product material. The balls fall from a height equal to the diameter of the drum and compress the material. Ultrafine Grinders Ultrafine g rinders have recently become ve ry popular due to there ability to produce nanoparticles. The most popular u ltrafine grinders are fluid ener gy mills such as jet mills and agitated mills such as stirred media mills. Jet mills stress a material by entraining particles in a gas stream and impacting the particles on a hard surface or against each other. Jet mills are used extensively in the pharmaceutical industry due to their ability to produce fine particles without the wear of mechanical parts. Another class of mills used to produ ce fine particles, are the stirred media mills, which are similar to ball mills exce pt they contain an agit ator, which supplies the necessary energy instead of rotati on of the vessel. The agitator allo ws the media to collide with a much higher force than is possible in the ball m ill. Due to their importance in this study, stirred media mills and particle breakage during stirred media mills will be discussed in more detail later. Figure 2-4, is a diagram of a stirred media mill. Products in a stirred media mill are often ground in a wet environment to allow for easier stabilization of the product. Cutting Machines Cutting m achines are used to reduce the size of resilient materials that are not easily fractured [27]. Some examples of materials that are processed us ing cutting machines are tire rubber for recycling, paper products, and plant mate rials. Cutting machines apply a shear stress to materials, which causes them to fail or break. Particle Deformation and Particle Breakage Particle Deformation Plastic deform ation in a metal occurs when th e material is stressed beyond its yield stress; the atomic bonds inside a material are broken and then reformed as planes of atoms are moved

PAGE 25

25 past each other. Plastic deforma tion results in a perment change in the shape of the object being stressed. Plastic deformation is affected by the stress ( strain rate ( d/dt) temperature (T), and the microstructure of the material [29, 30]. This study focuses on the deformation of aluminum particles when subjected to a compressive stress. Aluminum is a crystalline material with a face center cubic (FCC) structure with atoms at each corner and in the center of each face of a cube. Figure 2-5 is a sketch of a FCC unit cell. For such a structure, deformation often occurs by slip, the process of plastic deformation by dislocation motion, along the pref erred crystal planes. The {111} plane for aluminum is the most atomically dense plane and it is in this plane, along the <110> di rection, that slip occurs [31]. There are 4 slip planes for alumi num and 3 slip directions this m eans that there are 12 total slip systems for dislocation motion to occur in. Figu re 2-6 show the highlighted {111} plane and the <110> direction is denoted by the arrows beneath the highlighted areas. When dislocations encounter each other and a strengthening of the material will occur resulting in a loss of ductility. This strengthening is called work hardening, cold working, or stain hardening; and is due to the increase in dislocation density and de crease in the allowable dislocation motion. Equation 2-1 is the work hardening equation, y is the true yield stress, is the true strain of the material, Kw is the strengthening or work hardening coefficient, and n is the work hardening exponent. In the equation the stress and strain are true stre ss and true strain. The exponent and coefficient for the work hardening equation are generally ob tained experimentally. Many researchers have calculated the expone nt for numerous materials under numerous conditions [32-34]. n wyK 2-1

PAGE 26

26 According to this relationship, higher stresse s are required to obtai n greater amounts of deformation. The material will eventually work harden to the extent that no additional deformation may be achieved and the material will ultimately fail. In a milling process, a material will under go multiple loading and unloading cycles. This behavior can be characterized by a stress strain curve. Figure 2-7 is a generic representation of a stress-strain curve for plastic deformation follo wed by failure. When the material is unloaded after experiencing plastic deformation, it will rema in in the deformed state. The area under the elastic region of the curve is calle d resilience. It is the strain en ergy due to elastic deformation of the material. The total area under the curve until fr acture is called toughness. This is the same toughness discussed earlier for particle breakage. The area under the curve represents the amount of energy the material can absorb before it will fracture, or the strain energy due to plastic deformation. It should be noted that the area under the linear part of the curve is the energy due to elastic deformation. Strain rate, d/dt, is the rate at which a material is deformed. Schnert referred to strain rate as stress velocity and theorized that it could affect material breakage The strain rate also affects deformation. Several researchers have in vestigated the effects of strain rate on the deformation of materials [35-37] Tome and Canova investigated the stress strain response of aluminum and copper at different strain rates[37]. They showed that over the strain rate range tested, copper did not appear to have strain rate dependence. Howe ver, aluminum did have strain rate dependence; lower strain rates resulted in less deformation before failure. Nieh observed superplastic-like behavior, also known as extend ed ductility, of aluminum composites at very high strain rates. Meyers has also shown that at very high strain ra tes it is possible for a material to even recrystallize and form new grains [38]. Hines also observed this result in copper [39]. In

PAGE 27

27 order to achieve a greater understanding of partic le deformation during milling it is important to review particle breakage and to corr elate these results to deformation. Particle Breakage Num erous authors have studied single partic le and multi particle breakage [40-43]. They have found that particles can break in different ways depending on the method in which the load is applied. In a stirred media mill, load is app lied to particles trapped between colliding grinding media, media and the wall of the vessel, and betw een the stirrer and media. Therefore, particle breakage is a complex phenomenon with numerous st ages to study. Schnert extensively studied breakage, crack propagation, and crack velocity fo r many particulate materials and found that the strain rate or stressing velocity can influence material failure [44]. Polymer materials, for example, are strongly affected by st rain rate. Metals also have b een shown to have some strain rate dependence, however, the effect is much less than that of polymers [44]. There are three modes by which a material can fracture, which are a function of the material as well as the stress applied to the material. (1) Opening mode that corresponds to an applied tensile load (2) Sliding mode where the force is applie d in the plane of th e defect or crack (3) Tearing mode where a force is applied in a di rection out of the plane of the crack or defect For most particles, breakage is considered to take place in the first mode of failure. This may be due to artifacts in the te chniques developed to measure par ticle breakage. For this reason, most techniques for measuring particle failure invol ve stressing the particle in tension. If a normal load is applied to a partic le a load equal in magnitude is acting internally in the direction perpendicular to the load and can then be considered to be a tensile load. Tensile failure is more likely than compressive failure for many materials. Particle breakage can take place through many different mechanisms of loading. These mechanisms of loading as well as the material being stressed can affect how the particle will

PAGE 28

28 break. The mechanisms of breakage were allude d to earlier when discussing the types of milling machines and now will be discussed in more de tail. Rumpf suggested th ere are three ways by which a particle can break, which are fracture, cleava ge or attrition [26]. During fracture a mother particle will break into daughter particles or fragments that have a wider size distribution. During cleavage a mother particle will break into smaller daughter particles of roughly the same sizes. Attrition takes place when a mother partic le has surface asperities removed from the particle. This leads to much smaller daughter par ticles. Figure 2-8 repres ent the three different mechanisms of particle breakag e. If a brittle material experiences a normal load such as impaction in a milling process it would undergo a fr acture or attrition mechanism of breakage. If a ductile material experiences a shear force, it is most likely to fail by a cleavage mechanism [45]. Cracks and Defects Understanding crack form ation and propagation is of fundamental importance to the study of particle breakage. Particle breakage is preceded by crack gr owth, which is preceded by crack formation. Crack formation is enhanced by local defects in a material. Defects are separated into four classes, which correspond to their dimensions. (1) Point defects (vacancies, inters titials, or substitution) (2) Line dislocations (screw or edge) (3) Interfacial (4) Volume The strength of a material is related to the number, size and types of defects present. Most materials are only a fraction as strong as the theoretical strength would predict due to the presence of defects. A method by which cracks ar e initiated is when th e applied stress exceeds the strength of the material. Following crack ini tiation is crack growth, wh ich leads to particle

PAGE 29

29 breakage. Schnert extensively studied crack gr owth and found that crack velocity can vary widely depending on the material, environment, and mode of stressi ng [46]. Schnert also studied particle breakage during milling, however few researchers have studied deformation during milling [47]. Particle Deformation during Milling A significant am ount of work has gone into investigating particle deformation during milling, however, that work has focused on the mech anical alloying process[ 1]. This process is important because it is used to make alloys that can be made in no other way. Beside the formation of novel alloys, mechani cal alloying can be used to refi ne the microstructure of the materials to increase their stre ngth. Mechanical alloying and micr ostructure refinement involves adding a metallic powder(s), fatty acid/dispersant, and grinding media to a ball mill. The mixture is then milled until an equilibrium particle size is reached [4]. The particles generally produced in this process are semi-spherical agglomerates. Benjamin is the pioneering researcher of mech anical alloying. The author was the first to form an indium alloy [48]. Since then numerous authors have studied mechanical alloying (MA) [49-53]. Most of the studies on MA have focu sed either on alloying or on refinement of microstructure. Weeber studied the effect of di fferent ball milling operations on the final product and found that the milling equipment can affect the final product [52]. The author found microstructure can vary using different types of ball mills. Gilman investigated the effect of additives on the alloying process and found that if too little quantity of dispersant was added that the metal would coldweld together and not effectively alloy[4]. Even though mechanical alloying and the pr oduction of metallic flak es involve similar equipment and similar physics, the final produc t for flake production is very different. The desired product of the flake making process is a powder with a large major dimension and a

PAGE 30

30 narrow particle size distribution. It is not possible to run a flak ing process to an equilibrium particle size, as is the case with MA. Further milling will cause the flakes to break, no longer scatter light effectively, and perf orm poorly as metallic effect pigments. The major differences between the MA and flaking process are the us e of dispersant and the grinding time. Many flaking processes mill in a solvent to prevent th e metal particle from coldwelding together and the processes are run for shorter le ngths of time to limit breakage. The production of metallic flakes for use as pigments has been ongoing for over 50 years. Some of the first available works are patents for devises made to produce metallic flakes [54]. More recently metallic flakes have found increasi ng use in other applications such as conductive and electromagnetic resistant coatings and infrared obscurants [7, 8]. It is important for flakes to have a large major dimension in order to function in the specific part of the electromagnetic spectrum desired for these applications. Breakage during milling of these metallic particles ultimately leads to poorly perfor ming obscurants and paints. It is necessary to understand the relationship between particle deformation and milling. Moshksar has shown that when milling aluminum powder that the aspect ratio of the material [major dimension divided by minimum dimension (d/t)] increases to some maximum, then begins to decrease at longer milling times [55]. The decrease is due to particle breakage and a subsequent reduction in aspect ratio. The results suggest that an ideal product may be produced by monitoring the aspect ratio as a function of m illing time. The author also found that the cold welding and agglomeration of the powder particles was hindered by the fracture of the flakes at longer milling times. Zoz showed that the maximum value of the d/t ratio varies with the hardness of the metal being milled [56]. The lesser the hardness of the metal and the greater the d/t ratio, the more likely the material will form flakes. Material selection and strain rate can

PAGE 31

31 significantly affect whether the material breaks or deforms. It may also be possible to relate material properties such as har dness and strain rate to the asp ect ratio. This relationship would enable the development of a model that relates material properties, and milling parameters to particle deformation. Effect of Milling Parameters on Particle Deformation The deformation process has yet to be studied as extensively as the breakage process; the studies that have been perform ed are mainly focused on a specific application. Much of the research that has been performed is directed at understanding the micr ostructure development during mechanical alloying (MA) [4]. For the pa int industry, changes in particle morphology are important and not the change in the microstructure of the materi al [57]. Few researchers have investigated the milling of metals for use in the paint industry [5, 56, 58]. Zozs work was discussed above [11]. Sung studi ed the affect of particle morphology on the reflectance and quality of the paint. Hashimoto studied the effect of milling media diameter on the total energy consumption as well as the energy per impact on a metal powder in a simple vibratory mill [14]. From this, the average energy per collision was calculated. The author found that the total energy consumption was higher for larger media. The author also fo und that as media loading increased, the average energy per collision is reduced due to ball-ball interactions. Kwades models of stress frequency and stress intensity neglect ball-ball interactio ns. Hashimoto measured the change in the hardness of the metal with milling time, the work hardening of the metal, and found that the increase in the hardness of the powder was grea ter with larger media. This increase was attributed to cold working of the material during the milling process. Riffel determined that an optimum media size existed when mechanically alloying (MA) Mg2Si compounds. The author determined that if the media was too small it did not provide enough kinetic energy to alloy or

PAGE 32

32 reduce the size of the material [59]. Riffels results for MA also agree with the Kwades predictions of an optimum stress value (milling conditions) for particle breakage Kwade and Riffels result suggest that it may be possible to use particle breakage equations to describe the particle deformation process. Rodriguez milled aluminum in a stirred ball mill, and measured the hardness as a function of milling time [60]. Th e researcher found an initial increase in Vickers hardness which reached a maximum at a value of 130 kg/mm2. The increase in hardness measured by Rodriguez agrees with the increase also observed by Hashimoto. Hua ng also measures an increase in hardness with milling time in copper [12]. Huang used the change in the diffraction pattern measured from the high-resolution transmission electron microscope (HRTEM) to determine the lattice strain and measured a 0.2% increase in la ttice strain. Hwang also used powder x-ray diffraction data (XRD) to investigate the effect of milling tim e on lattice strain and determined that it increased[13]. Huang and Hwangs results agree and provide other methods of characterizing the work hardening of a material during milling. Particle Breakage during Milling Prasher gives a chronology of proposed laws relating energy f or breakage to simple parameters such as changes in surface area or change in particle size [28]. Two of the most notable relationships are Rittingers and Kicks Law. Rittinger states that, energy E, to break particles is equal to the ne w surface area times a constant as described by Equation 2.1. sK xx E 1 1211 2-1 The terms x1 and x2 describe the particle diameter of the feed and product material respectively, K1 is a constant and s is the change in surface area. Kicks Law, Equation 2.2,

PAGE 33

33 finds the energy to break a part icle is equal to a constant, K2, times the log of the starting particle size divided by the endi ng particle size. 2 1 2log x x KE 2-2 These relationships apply to a limited number of cases with precise particle sizes and specific materials and have not been shown to apply to processes where particle deformation is desired [28]. Figure 2-9 shows a comparison of brittle mate rials and ductile materials trapped between impinging grinding media. Brittle particles are co mpressed and fractured by the media, whereas ductile particles are plastically deformed and fl attened. Both processes re present an increase in overall surface area, however, during particle deformation th e volume of the particle is conserved. Effect of Milling Parameters on Particle Breakage Choosing the proper m illing parameters can have a significant effect on the breakage rate, energy consumption, and efficiency of the milling process. Grinding media density and size, for example, directly affects the magnitude of the load on the particles, and th us can be expected to alter milling. Many authors have studied the effects of milling parameters on particle breakage in a stirred media mill [21, 25, 61, 62]. Kwade form ulated a simple equation (eq. 2-3) to mathematically relate media size, media densit y, and rotational rate to the stress intensity. Kwades stress intensity term is merely a represen tation of the kinetic energy of the media and is not a representation of the actual stress act ing on the particles. Stress intensity ( SI ) is directly proportional to circumferential velocity v media size Dp, and media density p less the fluid density f through Equation 2-3 [63]:

PAGE 34

34 2 3fppDSI 2-3 Kwade then showed that at a specific energy input an optimum in stress intensity existed that provided minimum particle s size for a minimum energy input. The optimum energy input for producing a maximum fineness exists betwee n a stress intensity of 0.01 Nm and 0.1 Nm. Kwade also derived an expression th at relates the stress frequency ( SF ) to media size [63]. The stress frequency is a representation of the number of media-particles interactions, i.e. the number of times a particle is contacted by the media, and described by Equation 2-4. T D D SFp s s 2 2-4 Here T is milling time, Dp is the diameter of grinding media, Ds is diameter of the stirrer, and s is angular velocity. Kwade showed that the grea test stress frequency results from the smallest media size. Large stress frequency also results in the smallest particle size. The author was able to provide a method for optimizing and even scaling a milling process. However, these results have only been shown to apply to a system where particle size reduction is desired. The effect of stressing velocity on the deformation of materi als in a milling system has not been studied. Stressing velocity and strain rate in a mill can be assumed to be equivalent terms as previously discussed by Schnert [44]. Using Kwades model of SI it is now possible to relate specific milling parameters (media size, media density, and rotational rate) to the strain rate. This would be the first step in obtaining a realistic model to describe the deformation process in a stirred media mill. Eskin et al. developed a more sophisticated approach to modeling the energetics of a milling process. The authors calculated granular temperature, and used it to represent the velocity of the grinding media in their process [20]. The authors then calculated the force, and

PAGE 35

35 stress of the grinding media based on Hertz s theory and rigid body impact. Equation 2-5 demonstrates how force, F is related to media size, R, granular temperature(velocity), and materials properties, such as, Poissons ratio, density, and modulus of elasticity, Y [20]. 5 3 2 5 3 5 2 21 96.1 R Y F 2-5 Eskin derived a relationship determining the number of compressions a particle would experience as a function of media size. This rela tionship is similar to Kwa des stress frequency term. Eskin concluded by calculating a milling effi ciency term based on the energy input into the mill. The work by Eskin was found to be important to this study and the equation and derivations of his work will be revisited in more detail in Chapter 4. Summary In summ ary, Kwade has provided models that de scribe the stress intens ity and stress frequency in a stirred media milling process for particle breakage. These models may also be valid for describing the stress intensity in a milling process where deformation is desired. It is known that some metals have a strain rate dependent behavi or and that greater defo rmation may be achieved from a lower strain rate. Zoz showed that the am ount of deformation a ma terial can undergo in a milling experiment is also related to the hardne ss of the material. Researchers have also shown that the extent of work hardening varies with media size, concentration and milling time. The researchers have provided three methods by wh ich work hardening can be characterized, including indentation, XRD, and HRTEM. From these studies, it is now possible to begin to construct models that can pred ict material deformation based on the hardness of the material, deformation, milling time, media size, media density and rotational rate of the stirred media mill.

PAGE 36

36 Figure 2-1. 1) Particle stressed by compression or shear, 2) part icle stressed by impaction on rigid surface, and 3) particle stressed in shear flow. 0 10 20 30 40 50 60 70 80 90 100 Fluid Energy 2% Pin Mill 5% Ball Mill 8% Ball RoceMill 13% Swing Hammer Mill 22% Roll Crusher 80% Free Crusher 100% 0 10 20 30 40 50 60 70 80 90 100 Fluid Energy 2% Pin Mill 5% Ball Mill 8% Ball RoceMill 13% Swing Hammer Mill 22% Roll Crusher 80% Free Crusher 100% Figure 2-2. Milling efficiency variations by machine type. 1) 2) 3)

PAGE 37

37 Ball Mill Roll millA) B) Ball Mill Roll mill Ball Mill Roll millA) B) Figure 2-3. Types of grinding mills, A) ball mill and B) roll mill. Figure 2-4. Diagram of a stirred media mill, energy is supplied to the grinding media by the rotation of the agitator. Stirred media

PAGE 38

38 Figure 2-5. Face centered unit cell for aluminum. Figure 2-6. Slip planes and di rections for FCC aluminum.

PAGE 39

39 yUTS Elastic region Plastic region Unloading Stress Strain yUTS Elastic region Plastic region Unloading Stress Strain Figure 2-7. True stresstr ue strain curve for a material under going plastic deformation and strain hardening.

PAGE 40

40 Increasing relative particle size Increasing relative particle size Figure 2-8. Relative particle size distributions for attrition, cleav age, and fracture, respectively. Figure 2-9. Effect of impact (a ) brittle single particle, and (b) ductile sing le spherical particle. a) b)

PAGE 41

41 CHAPTER 3 MATERIALS, CHARACTE RIZATION AND EXPERIME NTAL PROCEDURE The goal of the study was to understand and pred ict particle deformation during milling. In order to achieve this goal it was necessary to develop experimental procedures and characterization techniques for performing and assessing particle deformation. This chapter describes the milling equipment and procedures used in this study. Characterization techniques and validation of the results are discussed in this chapter. In addition, the chapter describes the statistical analysis techniques used for predic ting particle deformation during stirred media milling. The chapter concludes with identification of the milling parameters used in this study. Materials Metal Powders The alum inum powder (product name H-2) used in this study was purchased from Valimet Inc, which is generally used in the production of metal effect paints and aluminum based explosives. Valimet manufactures the aluminum by atomization. Molten aluminum is sprayed through a high pressure nozzle and solidifies when it is exposed to argon atmosphere forming semi-spherical aluminum particles. The aluminum is then classified to obtain the desired size distribution. The powder has a pur ity of 99.7 weight percent aluminum and a median particle size of 3.2 m as measured by the manufacture. In house measurements of the particle size using light scattering determined the median size to be 3.8 m, and a mean of 4.2 m. Figure 3-1 shows the size distribution of the as received H-2 aluminum as measured by the Coulter LS13320 light scattering particle sizer. Milling Media The m illing media used in this study was yttria-s tabilized zirconia, purch ased from Advanced Materials Inc. in sizes of 0.5, 1.0, 1.5, and 2.0 mm. The density of the media is 6.0 g/cm3. Figure

PAGE 42

42 3-2 is an image of the 1.0 mm gr inding media used in this study. In order to maintain a consistent size distribution the media was sieved after each experiment to remove fines and fragments caused by fracture and wear [1]. Experimental Procedures Media Mill The m ill used in this study was a bench top Union Process attrition mill as seen in Figure 3-3. The mills operational speed is between 500-5500 rp m. Numerous types of milling vessels and stirrers can be equipped on this model. This study used a 750 ml stainless steel milling vessel and a 4 bar stainless steel agitator, the bars of the agitator are approximately 60 mm in length. It is important to note that the length of the agitator arm and the rotational rate was used to calculate the tip speed of the mill and ultimately the kinetic energy of the grinding media. The vessel is 90 mm in diameter and 90 mm in he ight. Figure 3-3 is a picture of the milling vessel and the agitator used in the study. The mill is attached to a recirculation reservoir that maintains the temperature of the mill contents at 25oC. The mill is attached to a digital controller, which provides speed and power control over the mill. Milling Procedure The m ill was assembled according the specificati on provided in the mill manual. To assure experimental continuity, the volume of material used and the distance between the bottom of the agitator and the bottom of the milling vessel (1/4 in.) was maintained in all experiments. The contents and milling sequence were performed in the following order: 1. 25 grams of H-2 aluminum powder 2. 5 grams of stearic acid 3. 300 ml of isopropyl alcohol 4. Initial dispersion 5. 400 ml of grinding media 6. Start milling experiment (sample at appropriate times)

PAGE 43

43 The milling medium used in this study was is opropanol with 5g of stearic acid dispersant. A more detailed description of these materials will co me later in the chapter. Prior to the loading of milling media, the mill was run for 5 min at 500 rpm to disperse the metal powder in the medium (liquid suspension). Following the 5 minutes of dispersion the mill was ramped up to the desired milling rotational rate. It was then shut off. The milling experiment was then restarted when the milling media was added. Samples were taken from the mill at times of 1, 3, 5, 8, 12, 15, 30 45, and 60 minutes then at 1.5, 2, 4, 6, 8, and 10 hours. In order to reduce the eff ect of the change in volume on milling kinetics the samples removed fr om the mill were less than 1.0 ml. Samples were diluted with 20 ml of is opropyl alcohol, sealed and then stored for either drying or characterization. At the end of each experiment, the mill was stopped and dissembled. The contents of the mill were pored through two sieves, one of suffi cient size (i.e. 1.0 mm mesh for 1.5 mm media) to catch grinding medi a and another of 90 m to catch fragments of grinding media. After the slurry was separated form grinding media it was placed in a Nalgen bottle for storage and later drying. Every part of the mill that was exposed to the aluminum slurry was rigorously cleaned with soap and water and allowed to dry in an oven before performing another experiment. The grinding media caught in the sieve was placed in a 1000 ml Erlenmeyer flask and washed with a 3 M hydrochloric acid (HCl) solution. The contents of the flask was then emptied into a sieve and placed in a sonic water bath. The sonicator was flushed repeatedly with water to remove HCl. The media was then placed in an oven and allowed to dry. Drying of Aluminum Slurry and Milling Samples There has been a considerable am ount of effort put into establishing the best way to dry a liquid suspension. A review of the work can be found in a dissertation on dry dispersion by

PAGE 44

44 Stephen Tedeschi [64]. This study will use a supercritical drying technique to dry samples obtained from the milling process. When liquid CO2 is heated to above 31.2 oC in a confined volume, CO2 becomes a meta stable fluid termed a supercritical fluid Figure 3-4. The supercritical drying method involve s solvent exchanging the suspensi ons fluid with that of liquid CO2, then heating the pressure vessel above 31.2oC and releasing the meta stable fluid. The importance of drying across each phase boundary wa s discussed in Tedeschi work; it was found that supercritical drying reduced capillary for ces that are present in conventional drying techniques [64]. Figure 3-5 is th e pressure vessel used in this study, it was purchased from SPI Inc. The vessel is capable of sustaining a pressure of 2000 psi and a temperature of 50oC. The method we used is as follows. The pa rticle suspension samples removed during milling was placed in a 25,000 Dalton dialysis ba g and diluted with isopropyl alcohol. The dialysis bag was used to allow diffusion of solven ts, while capturing the part icles. The bags were then placed in the pressure vessel, which was then filled with liquid CO2 and allowed to sit for 2 hours. After two hours, approximately 70% of the fluid inside the vessel was released and the vessel was again filled with CO2. The step of emptying and refilling the vessel was continued until all the isopropyl alcohol wa s removed, which took approximately 10 washes. After the final solvent exchange, the vesse l was heated to above 31.2 oC and the meta stable fluid was slowly released as a gas. The samples were then remove d from the vessel and placed in Nalgen bottles for storage and later analysis and testing. Sample Preparation by Dispersion of Dry Powder For m any of the characterization and analytical techniques used in this study the milled product must be dispersed in the dry form. A par ticle vacuum chamber was used to disperse the powder. Samples prepared in this way were used in image analysis, obscuration tests, and mass measurements. They were also used for calcula ting IR obscurant performance. Obscurant results

PAGE 45

45 will be discussed later in the case study. The de vice used for this study was a Galai partial vacuum chamber. Approximately 2 mg of superc ritically dried sample was dispersed in the chamber using 5 bars of vacuum. The amount of sample used was found by trail and error, and was determined to be enough to give a near monolayer of particles. Figure 3-6, shows a) the Galia chamber, as well as, b) a glass slide used for mass and image analysis measurements, c) a silicone wafer chip used for scanning electron microscope measurements, and d) a double pass transmission (DPT) slide used for Fourier tr ansform infrared spectroscopy measurements (FTIR). The exact use and importance of the samp les prepared by this method will be discussed in the appropriate characterization section. Statistical Design of Experiment Statistical design of experim ent (DoE) is a powerful tool used to optimize and analyze many industrial and commercial processes. This study first used statistic al design to understand and choose the parameters that are important to particle deformation during stirred media milling. The variables studied were rotation ra te, media loading, and media size. Additional designs and analysis were used to determine the interaction between milling parameters and define their significance. A central composite design was used to analyze the measured material response to milling (strain) as a function of milli ng parameters. The DoE software used in this study was purchased from Stat-Ease, and is calle d Design-Expert 7. The software is capable of performing statistical analysis of the data, based on analysis of variance (ANOVA). It is also capable of developing and experimental design that can then be modeled by the software. Characterization Characterization is a vital part of any study, it is important to understand and relate process variable to product qualities. By characterizing and interpreting the produ ct m aterials, it is possible to determine and predict the behavior of a material or process. This study used

PAGE 46

46 characterization techniques to measure the respon se of aluminum to process parameters before, during, and after mill. The charac terization techniques used in th is study and the sample there importance are discussed in the following sections. Light Scattering Light scattering has been used to rapidly and accurately measure particle size and size distributions. The main principle of light scattering for pa rticle size analysis is that light scatters at different angles based on the size of the partic le. A detector measures the angle and intensity of the scattered light and is able to reconstruc t a particle size distribution from that data. One limitation of light scattering is that it assumes all particles are spherical fo r size calculations. This limitation will be addressed in more detail in chapter 4. This study uses the Coulter LS11320 which is capable of measuring particle sizes from 40 nm to 2,000 m. The Coulter measures the particles size distribution of the samples taken from the mill. In order to achieve accurate measurements, the samples from the mill are fu rther diluted with isopropyl alcohol to a concentration fit for size measurement. To achieve an accurate measurement from the LS11320, light must be able to pass through the sample, th e approximate concentration must be <1 %. The data from the Coulter is used for calculations of stress, strain energy, st rain, milling efficiency and strain rate, these results will be shown in chapter 4. The LS11320 is capable of providing data output in number, surface area, and volume distributions. The differences in these measurements are due to how they are calculated. If the particle size distribution were narrow, these three methods would provide equivalent. However, if the distribution is bi-model or broad a number w ould appear to have smaller particles size and a volume distribution would appear larger.

PAGE 47

47 Microscopy and Image Analysis This study produced high aspect ratio m etallic flakes with major dimensions as large as 10 m and minimum dimension sizes as small as 10 nm This led to challenges in obtaining clear images from one specific device. Some of thes e challenges were getting both quantitative and qualitative flake thickness measurements due to difficulty in getting the flakes to stand perpendicular to the surface. There were also ch allenges in getting quantitative particles size measurements of the flake diameter to verify light scattering data. This challenge will be addressed later in this chapter. Electron microscopy Most of the im ages used in this study we re taken using a scanni ng electron microscope (SEM). The SEM functions by focusing a beam of electrons onto a surface. The beam of electrons then excites and emits electrons from the surface that are then detected and used to reconstruct the surface topography. The microscope used in this study was a JEOL 6330 cold field emission microscope. Samples for the SE M image were prepared by disseminating the powder onto a silicone wafer and attaching the wa fer to an aluminum stub with carbon tape. This procedure allows for a conductive contact betwee n the flakes and the instrument. SEM images were used to verify light scattering data usi ng image analysis software called ImagePro. The software was capable of thresholding the images and counting and measuring the particles. SEM was used to study particle morphology and dete rmine when particle breakage occurred. Transmission electron microscope (TEM) was used in this study to measure particles size, and crystallinity. TEM image materials by passin g an electron beam through a thin slice of sample. TEM can also determine crystallinity a nd microstructure of materials. Two TEMs were used in this study, the JEOL TEM 200CX was capa ble of imaging at high magnification, and the JEOL TEM 2010F was capable of measure crystall inity over a broad size range. Samples were

PAGE 48

48 prepared for the TEM analysis by mixing the dry powder with an epoxy. The epoxy was then allowed to cure. After curing, the epoxy was sectioned using a microtome to cut slices approximately 100 nm thick. Optical microscopy Im aging was also done on an Olympus BX60 op tical microscope equipped with a SPOT Insight Digital CCD camera. The microscope was capable of magnifying and image between 2 to 100 times. This microscope requires very little sample preparation to obtain images. Optical microscopy provided a quick way to estimate particle size and mor phology. However, it was challenging to measure particles of less than 5 m due to the low magnification and low resolution at that length scale. Occhio optical particle sizer The Occhio optical particle sizer was also used in this study. The instrument passes a sam ple below a fixed high magnification CCD camera capable of imaging thousands of particles in seconds. It then analyzes th e images with Calisto software. The Occhio instrument specifies that it is capable of measuring particles as small as 0.5 microns and as large as 3 mm. A comparison of data between the Occhio and light scattering will be shown later. Surface Area Analysis (BET) Surface area analys is by gas adsorption was first achieved by Stephen Brunauer, Paul Hugh Emmett and Edward Teller in 1938 and was later termed the BET technique [65]. BET involves measuring the amount of gas condensed on the surface of a sample at low temperatures and pressures. This study used the Quantachrome Autosorb 1C-MS to measure the specific surface area of samples taken before, during, and after milling. Samples for BET analysis were allowed to outgas under vacuum at 60oC for 12 hours before being placed in the BET for analysis. The particle size of the before milli ng samples were verified by image analysis and

PAGE 49

49 light scattering techniques. The after and during milling samples were used to study the change in particle size as a function of milling time and were used to determine IR obscurant performance. Inductively Couple Plasma Spectroscopy (ICP) Inductively coupled plas ma spectroscopy is us ed to measure the elemental composition of an organic or aqueous liquid. The ICP ionizes at oms is an argon plasma, the ionized atoms emit light when electrons return to the ground state. The light is detected and correlated to specific atoms. The instrument used in this study was the Perkin-Elmer 3200 Inductively Coupled Plasma Spectroscope. It has the capability to measure cont ents to less than 1 part per million. ICP was used to verify the purity of the starting material and to determine the ma ss used in the case study. As mentioned previously approximately 2 mg of sample was disseminated into the particle vacuum chamber. However, only a fraction of th e would end up on the DPT slide, this lead to difficulty in accurately determining the masstransmission relation necessary for measuring material performance as an IR obscurant. This wa s resolved by also disseminating on to a glass slide of known dimensions then measuring the c oncentration of metal dissolved off the slide. From the mass an accurate measurement of mass per area was obtained, which was then equated to the mass per area of the DPT slide. The meas urement of mass and obscuration values were then used to determine the performance of the material. Fourier Transform Infrared Spectroscopy (FTIR) Fourier transfor m infrared spectroscopy (FTIR) is most often associated with measurements used to identify compounds in liqu ids and on solid surfaces. The equipment used was the Thermo Electron Magna 760. The FTIR is capable of measuring the transmission, and reflectance of light at wa velengths from 400-4000 cm-1. This study used the FTIR to determine the percent obscuration of light in the infrared (IR) spectru m for the development of IR

PAGE 50

50 obscurants. The development of IR will be discussed in detail in chapters 7 & 8. As discussed previously, sample was disseminated onto DPT slid es, which were then place in the FTIR. Light was then passed through the sample and a percen t transmission was measured. This value was then used to calculate the perf ormance of the material, the calculation will be shown later. Light Scattering Results In this study, it was im portant to utilize techniques to char acterize the response (particle size) to varying milling conditions. This study used light scattering to measure particle size and used image analysis to validate the techniques. Results of Aluminum Powder Experim ental studies were performed on an aluminum powder (H-2) that deforms as a function of milling time. It was also important to c onfirm light scattering particle size results for the powders used in this study. Figure 3-7 shows an SEM image of the as received spherical aluminum powder. The powder is roughly spherical with few non-spherical agglomerates. The image also indicates that the par ticle size distribution for this part icular powder may be broad as can be seen by the large number of particles that are less than a micron and the large number that are greater than a micron. The particle size of this powder was measured by light scattering. Figure 3-8 gives the differential volume and number distribution data from this measurement. Figure 3-8 also shows the error bars that repres ent the standard deviation obtained from three separate measurements of the as received alum inum powder. The error bars are barely visible due to the very low sample-to-sample variations in the measurement technique. A difference in the volume and number distributions is apparent in Figure 3-8. If the particle size distribution was narrow and mono dispersed the number and vol ume distributions would overlay each other. However, the distribution is broad, as indicated by the SEM image. In addition, when there are many fine particles present the number distributi on tends to shift to the smaller particle size

PAGE 51

51 range. The mean particle size diameter as meas ured from the number and volume distributions are 1.0 m and 4.2 m, respectively. Surface area measurements were performed on the as received H-2 aluminum powder using a BET. This measurement was compared with the calculations of the specific surface area based on the mean particle size obtained from the light scattering data. The mass of an individual aluminum particle was calculat ed from the density (2700 kg/m3) for particles of mean diameters of 1.0 m and 4.2 m. The calculated specific surface area s for the mean number and volume particle sizes of 1.0 m and 4.2 m are 2.22 m2/g and 0.53 m2/g respectively. The measured surface area from the BET was determined to be 1.55 m2/g. The measured and calculated values are close only for the 4.2 micron part icle indicating that the BET a nd light scattering techniques are in close agreement. Results for Aluminum Flake However, this study focused on disk-shaped part icles obtained from the deformation of the spherical particles. It was therefor e necessary to verify the light sca ttering techniques used in this study with image analysis. To obtain a more accu rate measurement of the aluminum flake we employed the use of the Occhio particle si zing system (specifications and functionality discussed in chapter 3). Figures 3-9 and 3-10 are the particle size distribution data obtained from light scattering and image analysis using the Occhio. The data is for aluminum flakes that were milled for 10 hrs. The figures show an agreement between the distributions obtained from both measurement techniques. The mean particle size data obtained from the Occhio and light scattering measurements were 4.6 and 6.9 m, respectively. The Occhio counted over 16000 particles; this makes the measure more stat istically significant [66]. Table 3-1 gives a

PAGE 52

52 comparison between the light scat tering and image analysis data. It can be seen that the d10, d50 and d90 have a close values. Summary of Characterization and Light Scattering In any application where particle size is a cri tica l parameter, it is important to determine what the desired form of the particle size distribu tion is for a process. If the number of particles in the process is a key variable for qualit y control then a number distribution might be considered. However, if the mass of the materi al being processed is an important process condition the volume distribution may be important. This study will show later how the volume of the particle is important for modeling the m illing process. In this study the volume distribution obtained from light scattering data was de termined to be the most applicable. Even though there are ch allenges with using light scatteri ng to measure the particle size and size distribution of disk shaped particles, th e above work validating th e technique shows that can be used in this work. Particle size dist ribution of aluminum flak es obtained from light scattering and image analysis techniques were comp arable. This indicates that light scattering is a valid technique for characterizing the majo r dimension of disc shaped materials. Milling Condition Selection Preliminary Milling Study Stirr ed media milling is a complex proce ss with numerous operating and equipment variables affecting the performa nce of the mill and the final product. Some of the equipment variables that can be manipulated are listed in Table 3-2. The vessel size, geometry, stirrer size, and geometry were fixed. Table 3-3 gives a lis t of operating variables. Those variables were media size and density, solvent, material, and surfactants. However, many other operating variables need to be manipulated or determin ed in order to conduct a milling experiment.

PAGE 53

53 Preliminary Milling Results The best way to unders tand the effect of milli ng parameters is to vary conditions during a series of experiments. A preliminary study of the effect of media loading on mill performance was carried out using brass flakes obtained from US Bronze. The flake is the current material used in infrared obscurant. Scanning electron micr oscope images of the starting material can be seen in Figure 3-11. Statistical design of experi ments was used to determine the effect of media loading on deformation. A two fact or design was used in this e xperiment. This design involves varying the factor A (media load ing) between five diffe rent concentrations (levels) and a second factor B (rotational rate) between 3 different speeds (levels), then measuring a response (particle size). Table 3-4 gives the experiment al details used in this study. Experiments 4a-4c were used to confirm the repeatability of the experiments. The other parameters such as milling medium and medium volume, milling media size, material co ncentration and surfactant usage were held constant; Table 3-5 gives the specific milling conditions used in this study. The response variable used to measure mill performance was the figure of merit (FOM). Figure of merit is a measure of how efficiently a material obscures infrare d light (IR) and is an indication of the specific su rface area of the material. Equa tion 3-1 describes FOM where transmission, T is obtain from FTIR measurements, mass, m, is measured using the ICP and SA, is the surface area of a microscope slide. The material constant, Cm, is considered proprietary. m slideC SA m T LN FOM *2 100 3-1 A surface response plot of figure of merit vers us the factors listed above can be seen in Figure 3-12. The figure shows that as media load ing is increased the figure of merit is also increased. Higher media loading re sults in a larger number of gr inding media that can impart

PAGE 54

54 energy on the particles. The highe st figure of merit was obtaine d at 21% media loading. These results suggest that even higher media loading may further improve mill performance. However, it was found that if media loading was increased to greater that 70% that the vessel contents would overflow at velocities hi gher than 1000 rpm due to the tu rbulent environment inside the mill. It was then possible to maximize mill performance by maximizing media loading, this also allowed for the study of addition milling parameters by fixing media loading at a constant value. Establishing Milling Conditions This s tudy also focused on using a specific m ill and milling accessories described in Table 3-2. All milling experiments were conducted usi ng the same mill geometry, vessel size and stirrer geometries. Therefore, all the equipment va riables used in this study were held constant. We have established the importance of some of the operating variables on mill performance from the preliminary study. However, numerous othe r operating variables can be manipulated and may have a significant effect on mill performance. Milling medium The m illing medium is also an important parameter for a wet stirred media milling process. The medium affects the viscosity of the system, which in turn affects the energetic of the milling process. The medium can also affect the abi lity to disperse particles during milling and the selection of grinding aids used in particle di spersion. In this study, we milled an aluminum material exposing a very reactive native aluminum surface in the process. It was important to choose a milling medium that would have a favorab le reaction with the e xposed native aluminum surface. Experiments performed using water as a medium led to the rapid oxidation of the native aluminum surface. This caused the material to b ecome brittle and fracture. It also led to the production of hydrogen gas and the ge neration of large amounts of heat It was decided that this was not a favorable condition to make high aspect ratio metallic flakes due to the production of a

PAGE 55

55 large amount of alumina particles. Instead, n onpolar solvents such as isopropyl alcohol or heptane were considered. After consulting with Dr. Kremer at Silberline Inc. and reviewing literature on metallic flake production, the expl osive hazards of high surface area aluminum powders were brought to light [ 10]. Milling in heptane would pr oduce particles of high specific surface area with a minimum amount of pass ivation. This could potentially be extremely explosive. Therefore, isopropyl alcohol, whic h would provide a limited amount of passivation during the milling process, was used to reduce the threat of explosions. Mill loading and grinding aids One of m ost important parameters in a wet milling study is the viscosity of the slurry. The slurry viscosity is likely to change as a functi on of milling time due to increased surface area of the product from deformation and/or particle br eakage. Many researchers have investigated the effect of viscosity on mill performance [67-69]. Viscosity directly effects the energetic of milling. The higher the viscosity the more ener gy lost to viscous friction. One method for improving performance is by using grinding aids [70]. In order to minimize the effect of viscosity on experiments milling wa s done at very low solids loading to reduce any changes in viscosity as a function of milling time. The solids lo ading is the amount of material that is to be milled, in this case, it is the am ount of aluminum powder added to the mill. In this experiment 25 g of aluminum was used which amou nts to a solids loading of 8.3%. At high solids loading it becomes important to select the proper gr inding aid that reduce the viscosity of the system. As stated above, th is study maintained a low solids loading in the mill thus eliminating the need for a viscosity re ducing grinding aid. Generally, grinding aids are used to disperse particles broken during stirred media milling and to prevent particle agglomeration.

PAGE 56

56 Dry ball mills use grinding aids to prevent coldwelding of metallic particles. The most common grinding aid used to prevent coldweld ing is stearic acid. Figure 3-13 shows the chemical structure of stearic acid. Stearic acid is a saturated fatty acid used in firework production to coat metallic particles and pr event oxidation. In this study, the wet milling environment prevented the coldwelding of the meta llic particles. However, stearic acid was used to prevent oxidation of the metal during milling and after drying the product. An added benefit of the stearic acid was that it aided in the dry disper sion of the metallic particles when used for IR obscurants. Milling time and temperature Milling tim e is one of the most significant factors when running a milling experiment. Time affects the overall amount of energy consump tion of the mill, which will ultimately define milling efficiently. The final product and final pr oduct quality is highly dependent on the amount of time it has been milled. On the laboratory scale, it is possible to sample the milling process at varying milling times in order to evaluate the temporal effect s on the product. In this study, samples were taken though out the milling experi ment to identify when particle deformation and/or particle breakage occurs. Figure 3-14 represents a typica l particle size distribution data obtained from a milling experiment where particle deformation and breakage occur. The graph shows that samples taken from the mill up to 240 minutes show an increase in particle size, this is due to particle deformation and an increase in the major di mensions. After 240 minutes, the particle size begins to decrease due to particle breakage. From these results it is possible to identify at what time particle deformation is th e dominate mechanism and at what time particle breakage was dominate. It is possible to see th is effect in SEM images of samples taken at different milling times. Figure 3-15 a-c are SEM images taken at a) 0 minutes, b) 240 minutes, and c) 600 minutes of milling. At 0 minutes, the aluminum particles appear to be roughly

PAGE 57

57 spherical. As milling time increases to 240 minutes the spherical particles were deformed and are now disk shaped silver dollar particles. It is important to note that at roughly 240 minutes the edges of the particles are smooth and rounded indi cating that particle breakage has not occurred, it is also important to note that since particle breakage has not occurred that the volume of the particles in image a) and b) are equal. This observation of constant vo lume is important for calculations and models developed later in this study. After 600 minutes of milling the particles are much thinner and nearly transparent to the electron beam. They have irregular edges indicating that breakage has occu rred and the particle s can no longer be taken to have the same volume as the as received aluminum powder. This study focuses on the particle deformation process; therefore, it is important to understa nd when deformation stops or slows and when breakage begins. Experiments performed in this study were milled for 10 hrs with sample taken at times of 1, 3, 5, 8, 12, 15, 30, 45 minutes a nd 1, 1.5, 2, 4, 6, 8, 10 hours. The time of 10 hrs was selected because it was the longest ti me required to cause particle breakage. Temperature can affect mill performance by changing the properties of the environments such as viscosity of the medium. This study found that temperature has an added effect from the humid Florida environment where the experiment s were preformed. Initial experiments were performed at a constant temperature of 150C; however, condensation oc curred and caused large amounts of water to be added to the vessel. This problem was solved by increasing the milling temperature to 250C. Media properties and rotational rate The prelim inary study showed the effect of media loading on mill performance. It was found that greater media loading improved mill performance. This was an intuitive result because the greater media loading, the larger the fr equency of contacts that result in deformation. For this study, the media loading was fixed at 40% for all milling e xperiments. There are

PAGE 58

58 additional properties of the grinding media that a ffect the milling process. Those properties are hardness, density and media size. If the media hardness is not greater th an the product material then deformation and breakage will be difficult. This study uses yttria-stabilized zirconia grinding media (mohs hardness 9.0) to mill alum inum powder (mohs hardness 2-3). Another property of the media that can affect its perfor mance is the media size. In previous studies, Hamey showed that if the media is not of su fficient size it fails to impart enough energy on the material to cause particle breakag e. The author also showed that if the media is too large it does not perform as optimally as smaller media [71]. This study will focus on grinding media of sizes, 0.5, 1.0, 1.5, and 2.0 mm, these sizes are several orders of magnitude larger than that of the product material (mean particle size of as received aluminum powder 4.2 m) Media density can significantly affect mill performance; higher density media possess more momentum. Kwade showed the effect of media density on particle size, higher density media yielded better mill performance for particle breakage [72]. This study focuses on one media density for the aluminum experiments (zirconia 6 g/cm3). Rotation rate of the stirrer directly affects th e energy of the grinding media and in turn the amount of energy the product material experiences. For this study, rotation rate was varied from a high of 2000 rpm to 1000 rpm. Rotational rates higher than 2000 rpm also resulted in leakage from the vessel, similar to the effect of very high media loadings. Many authors have investigated the affect of rota tional rate on breakage [61, 72]. They found that the greater the rotational rate the more breakage that occurred. However, it is st ill unknown what effect rotation rate will have on deformation. It is even difficult to postulate this effect due to the complex nature of media-particle-media contacts. Chapter 4 will provide results of the effect of rotational rate on deformation.

PAGE 59

59 Figure 3-1. Particle size distribut ion of as received H-2 aluminum along with particle size data, obtained from Coulter LS 11320.

PAGE 60

60 Figure 3-2. Image of Yttria-sta bilized zirconia gr inding media obtained form an optical microscope, 1mm in diameter.

PAGE 61

61 a) b)Union Process Attrition Mill Vessel Stirrer a) b) a) b)Union Process Attrition Mill Vessel Stirrer Figure 3-3. Union Process stirred media mill use in study, a) stirred media mill, and b) milling vessel and agitator.

PAGE 62

62 Figure 3-4. Phase diagram for carbon dioxide, indicating the re gion of supercritical fluid formation.

PAGE 63

63 Figure 3-5. SPI pressure vessel used for superc ritical drying of powder samples taken from milling Figure 3-6. Galia partial vacuum chamber used to disseminate powder a) chamber, b) glass slide, c) silicone wafer chip, and d) DPT slide.

PAGE 64

64 Figure 3-7. Scanning electron microscope imag e of as received H-2 aluminum powder.

PAGE 65

65 0 1 2 3 4 5 6 7 8 9 0.1 1 10 100 Particle diameter ( m)Differential % Volume distribution Number distribution Figure 3-8. Differential volume and number distributions for as re ceived H-2 spherical aluminum powder measured by light scattering. 0 1 2 3 4 5 6 7 8 9 10 1 10 100 Particle diameter ( m)Volume percent (%) Volume percent light scattering Cummulative volume percent divided by 10 light scattering Figure 3-9. Volume particle si ze distribution of milled aluminum flake as measure by light scattering.

PAGE 66

66 0 1 2 3 4 5 6 7 8 9 10 11 01 0 0 Particle diameter ( m)Volume percent (%) Volume percent image analysis Cummulative volume percent divided by 10 image analysis Figure 3-10. Volume particle size distribution of milled aluminum flake as measured by the Occhio image analysis. Table 3-1. Statistical data comparison for light sc attering and image analysis performed using the Occhio particle counter. Image analysisLight scattering mean 6.6 11.2 d10 5.4 3.6 d50 10.1 9.6 d90 22.9 20.5

PAGE 67

67 Table 3-2. Stirred media milling equipment variables. Mill geometry (horizontal or vertical) Mill size Vessel size Equipment variable Vessel material Stirrer geometry Stirrer size Stirrer material Table 3-3. Stirred media milling operating variables. Operating variables Mill speed (rotational rate) Media size Media material (density, hardness) Temperature Time Media loading (volume fraction) Liquid medium (viscosity, density, polarity) Material loading (volume fraction) Surfactant, grinding aids, lubricants, and dispersants Table 3-4. Experimental parameters used in th e hexagonal statistical design of experiment. Experiment numberMedia loading (%)Rotational rate (rpm) 1 0.0 2000 25 37 0 0 3 5.3 3300 4a 8.0 2000 4b 8.0 2000 4c 8.0 2000 51 6 07 0 0 6 16.0 3300 7 21.3 2000

PAGE 68

68 Table 3-5. Milling conditions for preliminary study. Milling paramete r Level Media size0.5 mm Meda typeZircona Material Brass flake/90g Mill temperature20 C SurfactantSteric Acid/5g Medium 2-propanol/300ml Figure 3-11. SEM image of US Bronze as received brass flake used in preliminary milling experiments.

PAGE 69

69 R o t a t i o n a l r a t e ( R P M ) Figure of meritM e d i a l o a d i n g ( % ) R o t a t i o n a l r a t e ( R P M ) Figure of meritM e d i a l o a d i n g ( % ) Figure 3-12. Surface response of figure of mer it with respect to media loading and RPM. O O H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H H Figure 3-13. Stearic acid a satura ted fatty acid found in many animal fats and vegetable oils. 5.3 8.0 16.0 21.3

PAGE 70

70 0 20 40 60 80 100 0.1 1 10 100 Particle size ( m)Volume percent (%) As recieved 15 min. 45 min 240min 600 min Figure 3-14. Cumulative undersized versus particle size as at varying milling time for H-2 aluminum.

PAGE 71

71 a) b) c) a) b) c) Figure 3-15. Scanning electron microscope images of a) as received aluminum, b) milled for 240 minutes aluminum and c) milled for 600 minutes aluminum.

PAGE 72

72 CHAPTER 4 RESULTS AND DISCUSSION In order to u nderstand and predict particle deformation during milling it is important to study and thoroughly characterize the milling experi ments. Prediction of deformation can be obtained by modeling the milling process and the ch ange in particle size. By combining these two methods of studies, it is possible to establ ish a more comprehensiv e understanding of the flaking process during milling. This chapter wi ll discuss the empirical, semi-empirical and theoretical results obtained from milling. These results will then be used to construct models that explain and predict the effect of milling parameters on particle deformation during milling. Experimental Results of Stirred Media Milling Effect of Mill Parameters The goal of this study was to develop a better understanding of particle defor mation during stirred media milling. An observation was made fr om preliminary milling studies that indicated certain milling parameters resulted in larger diameter flakes and less particle breakage than other parameters. This was an important observation b ecause most processes that use metal flakes prefer flakes with high aspect ratios that have not been broken. The part icle size distributions shown in Figure 4-1 are for 1.0 mm milling media at 1000 rpm. In this figure samples were taken at different milling times. The bimodality seen in th e figure is due to particle breakage. It can be seen that bimodality was present at short millin g times. This indicates that the experiment resulted in breakage at short milling times. Figure 4-2 shows the particle size distributions for an experiment, which used 1.5 mm milling media and 1000 rpm. In this figure, the distribution was monomodal and shifted to larger particle sizes in dicating that the particle deformation occurred and there was only limited breakage. The differe nces in the effects of milling media on the aluminum flakes were also seen visually. Th e samples milled with 1.0 mm media appeared dull

PAGE 73

73 and non-reflective and showed limited pearlescence. Whereas, the 1.5 mm media produced flakes that were highly reflective showed a high amount of pearlescence[57, 73]. Figure 4-3 shows a comparison of the maximum particle diameter achieved in each milling experiment. It can be seen that for all rotational rates the 1.5 mm result in the largest pa rticle size. This result correlates well with the visual observations. In order to under stand the deformation process during milling it was necessary to study the tempor al effects of media size on mean particle diameter. The grinding curves (mean particle size versus milling time) for this study are shown in Figures 4-4 to 4-6. In all the e xperiments the mean particle si ze increases to some maximum value due to the flattening of the particles. This maximum part icle size was than followed by a decrease in particle size at long milling times, which can be at tributed to breakage of the particles. It can be seen th at the 1.0 mm grinding media results in breakage at the shortest milling times and that the 1.5 mm results in the la rgest particle diameter. Figure 4-7 shows that rotational rate does not have a st rong influence on maximum particle size. However, rotation rate does affect how rapidly the maximum particle si ze was reached. The influence of rotation rate on deformation will be discussed in more detail later in the chapter. At first, it was not completely understood why certain media sizes re sulted in larger diameter particles and ultimately a greater amount of deformation befo re breakage. However, this study developed a theory and methodology to explain the relations hip between particle deformation and milling parameters such as media size and rotational rate The next section compares and contrasts data using existing models for particle br eakage to the deformation data. Kinetic Energy Model For particle breakage, research ers have rela ted kinetic energy to the fineness of the product. Several researchers have shown that th e size of grinding media can significantly affect

PAGE 74

74 the particle breakage and particle deformation process during milling [14, 21, 63, 71]. Media size is an important parameter in Kwade s equations of stress intensity ( SI) (Equations 4-1). Stress intensity is simply the kinetic energy (KE) of the grinding media in liqui d slurry. In Equation 41, Dp is the diameter of the milling media, is the linear velocity of the media, and f and p are the density of the fluid medium and milling media respectively. A more detailed discussion of Kwades equations can be found in chapter 2. 2 3fppDSI 4-1 This equation has been the basis for the sca ling up of many particle breakage processes. However, there is no data or models relating gr inding media size to particle deformation during stirred media milling. As a first step, this study applied Kwades models to particle deformation. The hypothesis, equivalent to that for particle br eakage, is that if highe r stress inte nsity (KE) resulted in a greater amount of particle breakage, than higher st ress intensity (KE) may also result in a greater amount of deformation. Selection of the milling parameters chosen for this study was discussed in chapter 3. The effect of four grinding media sizes (0.5, 1.0, 1.5, and 2.0 mm), and three rotation rates (1000, 1500, and 2000 rpm) on the resulting deformation. Table 41 gives the calculated stress intensities for all th e experiments performed in this study. As media size and rotation rate increase, th e stress intensity (KE) also incr eases. This is due to the higher kinetic energy resulting from the larger mass and greater velocity objects. Kwade verified the results with experimental data for breakage, however, no one has shown whether the models are verifiable with deformation data. Table 4-2 s hows the maximum mean particle size for each milling experiment. By comparing the data in Tabl es 4-1 and 4-2 it is possible to determine if there is a relationship between ki netic energy (stress intensity) and deformation (mean particle size). It is logical that kinetic energy is dependent on both me dia size and rotation rate. Thus,

PAGE 75

75 equivalent kinetic energies could be obtaine d at different combinations of these milling conditions. However, it can be s een in Figure 4-3 and Table 4-2 that deformation does not vary with rotation rate. The energies for the [1.5 mm, 1500 rpm] and [2.0 mm, 1000 rpm] are equivalent. Thus, according to this theory the part icles sizes for these experiments should also be equivalent. However, the maximum particle sizes differ indicating th at kinetic energy can not be used to predict deformation. This result led to develop of a new model that was not based on kinetic energy of the grinding media, such as a contact mechanic model based on Hertzs theory discussed in the next section. Stress Model Eskin et a l. derived a model that relates milli ng parameters to particle breakage [20]. The basis for Eskins model was the use of Hertz theory to calculate a contact area, which was then used to calculate milling efficiency. Eskin used a term called granular temperature as a representation of velocity was used to calculate the force. In this investigation the linear velocity is used so a more direct comparison between this model and stress in tensity (KE) could be established[74]. In order to apply Hertz theory, the contact of the colliding bodi es must be elastic. In this research, the colliding objects ar e spherical zirconia grinding medi a with an elastic modulus of approximately 200GPa and a Poissons ratio of 0.31. Thus, the assumption of elastic contact is valid. The following equations were derived from Johnsons book Contact mechanics, and can also be found in Eskin [20, 74]. In E quation 4-2 the average normal force, Fb, of identical elastic beads is related to th e Youngs modulus of the bead, Yb, the Poissons ratio of the bead, b, multiplied by the density, b, times the velocity, and radius of the bead, Rb.

PAGE 76

76 b b b b bR Y F5 6 5 3 5 2 22 15 76.0 4-2 The radius of the contact area, b, was calculated from the force acting on the bead and the area of a circle. Dividing the force by the contact area is the stress, b, exerted at the center of the contact area by the beads, as described by Equation 4-3 and 4-4. 3 1 21 4 3 bb b b bFR Y 4-3 2)( 2 3b b bF 4-4 Figure 4-8 is a schematic of th e collision of elastic spheres. Table 4-3 gives the force, contact radius and stress calculated from Equati ons 4-2, 4-3, and 4-4. It was determined that the force increases with increasing media size and rotational rate. The media size and rotational rate also increase the contact area. The stress for a speci fic rotational rate is equivalent for all media size. This indicates that stress exerted by the grinding media is not sufficient to describe the variations in deformation from varying media size. However, the important quantity to know is the magnitude of the stress exerted on the par ticle and not the stress acting on the media. The stress acting on the particle ultimately results in the de formation of the material. A schematic of the particles trapped in the region between two colliding grinding media can be seen in Figure 4-8. From the size of the contact area of colliding media and the concentration of particles inside the mill, it was possible to calculate the number of particles in the contact area and subsequently, the stress acting on each particle. The stress ex erted on the particle, p, was calculated by dividing the contac t stress by the number of partic les in the contact area for a specific media size, as given by Equation 45 and Equation 4-6. The number of particle, Np, was

PAGE 77

77 calculated by dividing the contact area by the projected area of the particle and then multiplying by the solids volumetric concentration of aluminum particles in the mill, c c R Np b p2 2 4-5 p b pN 4-6 Table 4-4 gives the values for number particles in the contact area and stress acting on the particles as a function of media si ze and rotational rate. It can be seen that the stress acting on the particle decreases with increasing media size due to the larger number of particles in the contact area. Also, the number of particle s in the contact area increases with increasing rotation rate due to the increase in contact radius, which is attri buted to the milling media colliding with greater force. The yield stress and tensile strength of aluminum are 35 MPa and 90 MPa, respectively. However, the aluminum in this study was heavily strained, therefore, it may be more appropriate to consider the strain hardened values of aluminum for comparison purposes. The strained harden values of yield strengt h and tensile strength of al uminum are 117 MPa and 124 MPa, respectively [31]. Using the strained hardened values and Equatio ns 4-2 to 4-6, data was calculated and the following results were obtained. The 0.5 mm gr inding media exerts the most stress on the material, nearly two orders of magnitude greate r than the strength of the material. Such high stress could be responsible for low amount of de formation before breakage, this result can be seen in Figure 4-7. In the figure the maximu m particle size achieved for the 0.5 mm for any rotation rate was approximately 13 m. It should also be noted that out of all the media size the 0.5 mm milling media result in the longest amount of time to reach the maximum amount of deformation as can be seen by the figure, wher e, the 100 rpm experiment took almost 10 hrs to

PAGE 78

78 reach the maximum particle size. This result will be explained in the next section, which will discuss deformation rate. Similarly, the stress of the 1.0 mm grinding media is higher than the tensile strength of the material thus resulting in particle breakage and a minimu m amount of deformation. The high normal stress of the 1.0 mm media resulted in rapid particle breakage, which can also be seen in Figure 4-7 and in more detail in Figure 4-1. Conv ersely, the 1.5 mm media exerts a stress that is closest to the yield stress of a strain hardened aluminum mate rial. Each compression event for the 1.5 mm media could be considered a loading to the yield strength of the particle resulting in finite amount of deformation. Media of this si ze can therefore apply the maximum amount of deformation to the material without failure. Th e largest media, 2.0 mm, apply a stress to the particle that is below the yield stress of the mate rial and thus resulting in less deformation then the 1.5 mm media. These results can also be seen in Figure 4-7 where the 1.5 mm results in the most rapid deformation and the largest magnitude of deformation. However, it can be seen in Table 4-4 that the stress acting on the pa rticles for a specific media size does not change significantly with rota tion rate. This stability in stress values may explain the data in Figure 4-7, which shows that the maximum deformation was not significantly affected by rotation rate. These results are in cont rast to the breakage models that state rotation rate has a large effect on breakage. The m odels described above give a fundamental interpretation of particle deformation as a function of mill parameters. However, this study has so far neglected the influence of number freque ncy of media collisions on particle deformation. The next section will discuss the effect of me dia size on the frequency of media collisions and the resulting affect on deformation and deformation rate.

PAGE 79

79 Deformation Rate This study dem onstrated that rotation rate a ffected the rate at which deformation was achieved. Eskin and Kwade developed models that expressed the rate at which collisions between grinding media take place. Kwades Equa tion 4-7, was used to determine the rate at which media collide. In Equation 4-7, stress frequency, SF is equal to angular velocity, s, times diameter of the stirrer, Ds divided by the diameter of the bead, Db, multiplied by milling time, T T D D SFb s s 2 4-7 Kwade determined the necessary stress frequency for reaching a specific material fineness. The author found a specific fineness could be obtained by using a hi gh stress frequency and low kinetic energy of grinding media, or vise versa. In addition, Kwade found this relationship held up to a point of a minimum amount of kinetic en ergy. Kwade considered a constant volumetric media loading, for comparisons of stress frequenc y. Eskin derived a similar model, but chose to measure the number of times a particle enc ounters the media. Both authors found that the frequency was dependent on the media size and ro tational rate. Table 4-5 gives the values of stress frequency for this study as calculated using Kwades mode l. The model shows that stress frequency increases with media size. The m odel also gives a linear relationship between frequency of collisions and rotati on rate. However, values in the ta ble refer to the collision of the media and fail to give the number of co mpressions the particles experiences. The time at which the particle size begins to decrease for each experiment, was estimated from the data in Figure 4-7, and was recorded in Table 4-6. Kwades model indicates that smaller media size had the largest collision frequency, as s hown in Table 4-5. When this result is coupled with the high stress acting on the particle a ssociated with the 0.5 mm one would expect this experiment to approach the maximum amount of particle deformation more rapidly than the

PAGE 80

80 larger media sizes. However, it was shown in th e previous section that the 0.5 mm reached the maximum deformation slowly. It is challenging to compare the effect of media size on the onset of particle breakage for all the media sizes due to the complex nature of breakage. However, it is certain that rotational rate affects particle brea kage. Higher rotation rates result in a more rapid onset of particle breakage. In addition, Kwades model would predict a more direct correlation between media size and breakage, which is not the case for particle deformation as observed in this study. In order to explain the discrepancies in these models with the experimental data it was necessary to formulate a new model. Therefore, in this investigation the stress frequency (Table 4-5 and Equation 4-7) was multiplied by the number of particles in the contact area (Table 4-4 and Equation 4-5) to obtain the total number of particles stressed as a f unction of milling time, Table 4-7 gives the total number of particles stressed in 60 minutes. From Table 4-6, it can be seen that the total number of stressing events does not change as a function of media size. However, the total number of stressing events does change as a function of rota tion rate. These results explain the increase in the breakage as the rotation rate increases, as shown by the expe rimental data in Table 4-6. The frequency model suggests that the number of collisions is only a function of rotation rate and not of media size. This is way rotation rate only e ffects the time it take to achieve the maximum amount of deformation and does not effect th e actual maximum deformation achieved. In summary, it has been found that milling media size determines the maximum amount of particle deformation until failure. This was explained by calculating the stress acting on the particles using Hertzs theory. Rotational rate was found to have only a minor effect on the maximum achievable particle size, Table 4-2. Th e calculation of stress acting on the particle showed that there was a minimal amount of change in stress as a function of rotational rate. Thus

PAGE 81

81 confirming the result that rotation rate did not have a significant effect on maximum particle deformation. The time it took a particle to break was modeled using stre ss frequency. Rotation rate was shown to affect the rate of deformati on. Existing models failed to explain how rotation rate affects particle breakage. Therefore, a ne w model was derived that showed rotation rate directly influenced the number of stressing even ts. The models give the fundamental relationship between media size, and rotationa l rate on particle deformation. However, the models above do not predict the energy efficiency of the milling process or give any indication on how the milling parameters affect efficiency. Strain Energy and Milling Efficiency Strain Energy One of the greatest challenges in any milling operation is the measurement and calculation of milling efficiency. It is relatively easy to measure the amount of energy input to a milling process, however, it is very difficult to determ ine how the energy is uti lized. Many authors have studied the amount of energy requir ed to break either a single particle or an ensemble, a review of this work was given in Chapter 2 [43, 47, 75]. However, the milling process consists of numerous breakage events and a particle will likely fracture several times before it reaches the desired particle size. It is then difficult to determine the energy required for a particle to break to a specific size then break again to another size. This makes cal culating the energy of breakage difficult. Generally, milling efficiency is estimated by measuring the energy input and subtracting the energy that is expended as heat. Most mills are cooled to maintain a constant temperature, and the heat is usually measured from the coolant. This method does not directly establish how much of the energy we nt into the material itself, only the energy dissipated as heat. Efficiency calculated from this method does not re present the milling efficien cy to cause particle breakage, thus making it difficult to optimize the milling process.

PAGE 82

82 However, the particle deformation process is unique in that the number of particles does not change and the energy that goes into deforma tion is stored in the material as strain energy. From measurements of the amount of deformati on the material experienced, it was possible to calculate the strain of the material. From the st rain, it was then possible to determine the strain energy. The strain energy was then used to ca lculate milling efficiency as a function of milling parameters. Usually strain is directly measured by studying the change in dimensions of the object. In this study, it is difficult to measure all the dimensions of the particles being strained. It was shown in Chapter 3 that light scat tering could be used to character ize that major dimension of the particles. However, there is no direct method for characterizing the minimum dimension, i.e., thickness of the flakes. Thickness was calculate d by considering there was no change in the volume of the material up to the point where the mate rial fails. Failure is taken to be the point at which the mean particle size begins to decrea se. Equations 4-8, 4-9, and 4-10 were used to calculate the thickness of the flake by calculating the volume of the as received particles. In Equation 4-8, Vs is the volume of the as received spherical aluminum particles and Dp,o is the diameter of the as received aluminum particle s. The volume of the as received sphere is equal to the volume of the flak e in Equation 4-9. The volume of the flake was calculated using the equation for the volume of a cylinder, Vcyl (a flake is essentially a cylinder with a small height, ie., thickness), and th e diameter of the flake, Dp,f, at some sampling time; tp,f is the thickness of the flake at the corresponding sa mpling time. The thickness of the flake was calculated by equating the volume of the sphere to the volume of the flake using Equation 4-10. 3 ,23 4 op sD V 4-8

PAGE 83

83 fp fp cylt D V, 2 ,2 4-9 2 3 ,2 23 4 fp op fpD D t 4-10 The strain, was then calculated form thickness of the flake by using the true stain, Equation 4-11. A diagram for the change of the spheri cal particle to a flake can be seen in Figure 4-9. The true strain was calculated for every sampling time up to the point where the particle size decreases. op fpD t, ,ln 4-11 The work hardening equation, Equation 4-12, was us ed to calculate the stress. The coefficient and exponent (K and n, 180 MPa and 0.2 respectively) for the work hardening equation were obtained form literature for aluminum [31]. nK 4-12 Figure 4-10 is the stress-strain curves at 1000 rpm for each media sizes. The 1.5 mm media yields the highest strain for a ll the experiments at 1000 rpm rotati onal rate. Similar behavior can be seen in the 1500 rpm and 2000 rpm stress-strain curves, as seen in Figure 4-11 and 4-12, respectively. The number of data points on each set of data decreases as rotational rate increases due to the more rapid particles breaking at the high rotational rates and thus shorter milling time. Mill Efficiency As discusse d in Chapter 2, stirred media milling processes have been estimated to operate at energy efficiencies of less than 2%. These es timates are for particle breakage processes and

PAGE 84

84 are generally obtained from monitoring heat genera ted in a mill. There is no research, which measures the amount of energy supplied to the material. The following section will demonstrate that milling processes do not innately operate at low energy efficiencies, and will describe methods of improving the milling efficiency for particle deformation. This study used measurements of particle deform ation to calculate the amount of energy stored in the material. Strain energy was obtained by integrating the stress-strain curves and dividing by the volume (volume of the particle remains constant at this point on the stress -strain curve) of the particles. Table 4-8 is the strain energy as a f unction of media size and ro tation rate. It is clear that the particles milled with the 1.5 mm grinding media have the highest strain energy. It is important to note that nearly the same strain ener gies were obtained at the higher rota tional rates in a much shorter amount of time. The total strain energy divided by the energy in put to the mill is milling efficiency. The total strain energy was obtained by multiplying th e strain energy by the total number of particles in the mill. Table 4-9 gives th e total strain energy from each milling experiment. The 1.5 mm milling media resulted in the highest strain ener gy values, which was due to the large amount of deformation produced by the media. The energy input into the mill was calculated using the equations developed in Eskin [20]. In order to determine the energy input Eskin calculated the energy for a stirred mixing tank modeled by Naga ta [76]. Equation 4-13 describes the power input, Pw, to the mill calculated from tank geometry, mill speed, and slurry density. The terms N, H, D, and are the rotation rate, tank height, stirrer diameter and tank diameter, respectively. H DN MPm w 2 33 4-13 Slurry density, m, was calculated from the density of media, b, liquid density, L, and media solids loading, c as shown in Equation 4-14.

PAGE 85

85 ccLbm1 4-14 M is a coefficient calculate d using Equations 4-15 through 4-20. In Equation 4-15, Re is the Reynolds number for the tank, which is calc ulated by Equation 4-16. The coefficients A, B, and exponent h, are empirically determined constants [76]. hB A M 66.0 3 66.0 3Re2.310 Re2.110 Re 4-15 m mND 2Re 4-16 In Equation 4-17, b is total stirrer diameter, and for this study b is equal to the diameter of the stirrer multiplied by the number of stirrer, four (four stirrer arms on the agitator). 1856.0670142Db A 4-17 D bB14.15.043.1210 4-18 4 275.05.241.1 b Db h 4-29 The dynamic slurry viscosity, m, used in Equation 4-16 was calculated by Equation 4-20. c ccLm20exp0019.0105.212 4-20 Table 4-10 gives the values of the mill geometry and mill operating constants used in Equations 4-13 through 4-20. Mill power was then multipli ed by the mill volume and milling time to obtain the energy of the mill as a function of time. Equation 4-21 gives milling efficiency, which is equal to total strain energy, SET, divided by mill energy, Em. 100*m TE SE 4-21

PAGE 86

86 The mill energy was estimated for a stirred tank, which may underestimate the true mill energy. This is because a mill requires more power to move the high mass of milling media in the chamber. This is not a major concern, because most milling processes can measure milling energy directly from measurements of torque or power draw and th erefore obtain a better measurement of milling power and efficiency. Figure 4-13 shows a plot milling efficiency versus milling time to breakage for the 0.5, 1.0, 1.5, and 2.0 mm media at 1000 rpm. At short mill ing times, the milling efficiency is very high due to the rapid deformation of the material. It also can be seen that for the 1.5 mm media the efficiency approaches 100 percen t. This result is due to the me thod for estimating mill energy. It was observed that the 1.5 mm milling process is mu ch more efficient than the rest of the media sizes, due to the more rapid and greater amount of deformation than the other media sizes. Figure 4-13 gives the efficiency for 1 hour of milling, however, the milling experiments were conducted for 10 hours. It can be seen at 1 hour of milling that the efficiencies begin to approach the values reported in literature of <2%. Further milling leads to an even more inefficient process. This result would suggest that milling time is the l eading cause of inefficiency. Therefore, it is important to select parameters that will result in the most rapid deformation, such as, a high rotation speeds. Similar results are observed at rapid rotationa l rates as shown in Figures 4-14 and 4-15, for 1500 and 2000 rpm respectively. The 1.5 mm me dia results in the most efficient milling process, however it is difficult to determine from these plots whether efficiency was affected by rotational rate. In orde r to understand the effect of rotationa l rate on mill efficiency, the results for the 1.5 mm media was plotted in Figure 4-16 at varying rota tional rates. By plotting the milling efficiency at maximum strain, it is possible to obtain a relationship between rotational

PAGE 87

87 rate and efficiency. Figure 4-17 is milling effici ency as a function of rotational rate. The plot focuses on the data as it approaches particles brea kage. It can be seen from the graph that the 2000 rpm speed resulted in a greater efficiency. This is due to the more rapid rate of particle deformation. Ultimately, milling at rapid rotati on rate would result in energy saving due to increased efficiency. Rapid rotation rates woul d also increase time saving due to reduced processing time. This study has shown that the benef its can be realized with no detrimental effect to the product quality, i.e. strain. By investig ating milling efficiency it was also possible to determine the mechanisms by which energy is wasted, as discussed below. As mentioned earlier milling processes are ge nerally considered inefficient processes, with efficiencies of approximately a fraction of a percent. Usually, most of the energy was considered lost to viscous friction. This study fo und that milling could be a very efficient process at short milling times, indicating that energy lost to viscous friction was grossly overestimated. It can be assumed that the energy lost to friction is constant throughout the milling process due to the low solids loading in the mill, and minimal change in viscosity during the milling process. This indicates the low efficiency observed at lo nger milling times in this study must be due to another mechanism. Table 4-11 shows the milling efficiency as a function of percent strain, and table indicates that the efficiency decreases as percent strain increases. This could be explained by the fact that it becomes increasingly difficult to cause plastic deformation in the material, due to the increase in strain hardening with incr eased strain. The media contacts with the powder results in increased elastic contac ts that do not cause strain and re sult in a waste of energy or in efficient milling. In summary, this study has determined the most effective milling parameters for maximizing particle deformation. The study de monstrated how stress acting on the particle

PAGE 88

88 ultimately determines deformation. The strain energy stored in the material during processing was also calculated and the effect of milling parameters on efficiency was identified. However, one of the goals of this study was to develop a m odel that predicts particle deformation based on milling parameters. In order to achieve this goal, statistical design of experiments was used to develop a milling model for pred icting particle deformation. Empirical Modeling of Particle Deformation through Statistical Design of Experiment Statistical design of experiment (DOE) is a powerful tool used for, identification of significant process parameters, development of se mi-empirical models, and process optimization. Many industries regularly employ the use of DO E in the development and implementation of new process [77]. In Chapter 3, screening designs were used to identify the significant variables affecting particle deformation. These designs su ccessfully determined the effect of grinding media loading on deformation. The results of these designs were used to plan future experiments to study other milling parameters. The data presente d in this chapter was the result of using DOE to develop a model capable of predicting part icle deformation. A central composite design (CCD) was used for this purpose. The CCD was us ed to obtain the coefficients to a quadratic model and to obtain a response su rface. The surface response was used to obtain a relationship between design variables (milling parameters) and the response variable (strain). A detailed description of the CCD methodology can be f ound in Montgomery and in the design of experiment software package Stat-Ease [78]. Two CCD designs were used to model the m illing experiments. Both designs were twofactor designs. The first design s data can be seen in Table 4-12. The media size for this design was varied from 0.5 to 1.5 mm and the rota tion rate was varied from 1000 to 2000 rpm. Repetitions were used for the [0.5 mm, 1500 rp m], [1.0 mm, 1000], and [1.0 mm, 2000rpm] to insure repeatability. The total design involved 12 experiments and the parameters modeled in this

PAGE 89

89 study were rotation rate, media size, and the inte raction between media size and rotation rate. The analysis of variance (ANOVA) results ar e shown in Table 4-13. This study used a significance level of 5 % or a confidence interval of 95 %, because this is a standard method of determining statistical significance. It can be seen that the p-value for the model is 0.0165, which means the model is statistically significant .When the p-value is less than 0.05 then it means the terms in the model (media size and rotation rate) have a significant effect on the response variable (strain). Equation 4-22 is developed based on the DOE and is in terms of coded factors. The coded factors allow for a dire ct comparison of the coefficients in the equation. If the terms were uncoded or the actual terms it would be necessary to normalize the equations by the magnitude of the factors in order to evaluated th ere affects. For example, rotational rate has a numerical value three orders greater than that of the media size 1000 rpm compared to 1.0 mm. Which means the coefficient of the rotational rate in actual terms would n eed to be three orders of magnitude less for the media and rotation ra te to have a similar effect on strain. The coefficients of the coded equation represent the significance of that factor effect on the response, higher magnitudes of the coefficients i ndicate greater significance. In Equation 4-22, A is the coefficient for rotation rate, B is the media size, and is the strain at failure. 2 254.0062.0067.02.012.018.2 BA AB BA 4-22 From the equation, it can be seen that the media size (B) has the greatest effect on strain and that the interaction between media size and rotation rate (A) ha s the smallest effect on strain. To accurately model the mill perfor mance it is important to model the process with respect to the actual milling parameters.

PAGE 90

90 Equation 4-23, models the design in terms of the mill parameters. The model was used to predict the strain with respect to rotation rate an d media size, within the framework of this design (i.e. between 1000-2000 rpm and 0.5-1.5 mm media). 2 27 4 418.210*48.210*70.256.310*65.738.3 BA AB BA 4-23 The contour plots and surface response can be s een in Figures 4-18 and 4-19, respectively. Figure 4-20 is a plot of standard error associated with the experime nts. It can be seen that the error was low at the center points in the study but increased towards the limits. From the surface response, Figure 4-19, a minimum strain at failu re was obtained from the 1.0 mm media. The existence of the minimum correlates with the anal ysis performed previously, that also showed that 1.0 mm media resulted in the least amount of deformation. The surface response further confirms the result that rotational rate does not significantly influence strain. This result can be seen by the effect on strain at different rotatio nal rates and constant me dia sizes. One of the problems with this design is that extrapolat ion beyond the limits of the design may provide misleading information. For example, the surf ace response and model indicate larger media size will result in higher strains. However, this st udy has shown from empirical and theoretical data that this does not occur. In order to develop a stronger model a second central composite design was constructed that would investigate the effect of milling paramete rs on strain, beyond the limits of the first design. The second composite design investigated the eff ect of a larger media size on strain. Table 4-14 gives a list the experiments and response used in the second design. There were 11 experiments modeled in this design with repe titions of [1.0 mm, 1000 rpm], and [1.0 mm, 2000 rpm]. Table 4-15 gives the analysis of variance results for the sec ond design. The p-value for this design was 0.008, which means the m odel is statistically si gnificant. Coded variable were used to

PAGE 91

91 determine the significance of the milling parameters on strain. Equation 4-24 is the second designs model in terms of coded variables. 2 258.005.0045.018.012.085.2 BAAB BA 4-24 The coefficients of the coded equation indicate the magnitude of the effect on the strain. The rotation rate has a negative affect on strain meaning increasing rotational rate will decrease strain. However, the media size has a greater eff ect than rotational rate with a strong negative second order effect. The second order effect of media size ultimately leads to a maximum in strain at intermediate media sizes. The interac tion term AB has a small coefficient, which means that there is only a minor effect on strain from the combination of media size and rotational rate. Equation 4-25 is the model for the second central composite design in actual terms. This model can be used to predict maximum deform ation within limits of the design. 2 27 4 332.210*01.210*80.104.710*11.168.1 BA AB BA 4-25 Figures 4-21 and 4-22 are the contour and surf ace response plots for the second model. Both graphs indicate that an optimum amount of stress may be obtainable with grinding media of approximately 1.6 mm. The plots al so show that there is little effect on the deformation with respect to rotational rate, which further confirms the qualitative result observed earlier in Figure 4-7. Figure 4-23 give a plot of the interaction between media size and rotational rate. If there was a significant interaction the curves would inters ect, but as they are, there is only a slight interaction at the 1000 rpm rotation rate. The small interaction terms seen in the first and second designs shown here strengthens the hypothesis that stress determines th e magnitude of particle deforma tion. This is because stress is a function of media size and only moderately effect ed by rotational rate. It also strengthens the theory that stress frequency is constant as a function of media size. The stress frequency only

PAGE 92

92 varies as a function of rotational rate and constant as a function of media size. Both the proposed models for stress and frequency i ndicate only a small interaction between the rotational rate and media size. The standard error for the second ce ntral composite design can be seen in Figure 424. In this model, the error increase at the limits of the rotation rate and at the highest media size. In summary, design of experiments is a power ful tool for process optimization. In this study, this experimental methodol ogy enabled the development of models capable of predicting particle deformation. This was proven by Equatio n 4-25 that relates media size and rotation rate to strain. Design of experiments also allowed for the validation of the proposed milling theory discussed previously by quantifying the effect an d interaction of milling parameters. One more method of validating the above results was to look at the changes in the microstructure of the aluminum flake. If the number of dislocation is quantified as a function of milling time, it may be possible to recalculate strain energy, which would offer another source of validation of the milling models. However, this study only was able to perform a qualitative analysis with respect to dislocation density as shown below. Microstructure Analysis The study of the microstructure and change to microstructure is an extensive part of material science research. If the change in the number of dislocations could be characterized, it may be possible to determine the energy per disl ocation and ultimately the strain energy. An attempt was made in this study to characteri ze the dislocations in the material by using transmission electron microscopy (TEM). However, only information of cr ystallinity and flake thickness were obtained, and the quantification of dislocation density was unsuccessful. It was determined that the study of dislocation de nsity was beyond the scope of this research. Many of the processes that defo rm metallic flakes plan on th e material to coldwelding to refine the microstructure[48]. This study did not observe coldwelding taking place, and that the

PAGE 93

93 wet milling process used in this study prevented th e coldwelding of the metal powder. In order to confirm these results transmission microscope images and diffr action patterns of the powder before and after milling were taken. Figure 4-25 is a TEM image of an as received aluminum particle and Figure 4-26 is the di ffraction pattern for the particle. It can be seen from the diffraction pattern that the alumin um particle in this study is si ngle or nearly a single crystal. Many more pictures would be n eeded to obtain a statistical representation of crystallinity, however, the particles sampled here all gave si milar results for crysta llinity. Upon initial inspection of the images, the dark regions in the image were thought to be grains; however, Figure 4-26 shows a single crysta l diffraction pattern which mean s they cannot be grains. The discrete points indicate a single crystal pattern; a polycrystalline material would have filled or nearly filled rings. Samples from the 1.5 mm, 100 0 rpm milling experiments after 4 hr of milling were sectioned and then imaged using a transm ission electron microscope. The flakes in Figure 4-27 are approximately 100-200 nm thick after 4 hour s of milling and this compares closely with the estimated thickness from calculation of 200 nm. This indicates that th e assumption that only a small amount of failure has occurred and that the flakes maintain a constant volume was correct. Figure 4-28 gives the diffraction pattern of a milled flake, which appears to be that of a single crystal material as described above. This result confirms that coldwelding has not taken place. If coldwelding was present the diffrac tion pattern of the milled flake would be polycrystalline due to the welding of multiple particles together. It is possible to obtain information about lattice strain and dislocation density from the TEM. However, this work was beyond the scope of this dissertation. Summary This study has developed a fundamental re lationship between m illing parameters and particle deformation. Through investigations of milling mechanics, it was possible to determine

PAGE 94

94 the stress exerted on a particle as a function of rotational rate a nd grinding media size. The study determined, for the first time, the frequency at which a particle was stressed during milling. It also established a relationship between milling parameters and frequency. A novel approach was used to calculate milling efficiency based on stra in measurements and stra in energy calculations. Conformation of the frequency and stress models were obtained from statistical design of experiment. A predictive model for particle de formation as function of milling parameters was also developed using statistical design. Finally, TEM studies were performed that confirmed the thickness calculations used in the strain measurements and the crystallinity of the material. The above work is a significant step in the understanding of the deformation process during milling.

PAGE 95

95 0 1 2 3 4 5 6 7 8 0.1 1 10 100 Particle diameter ( m)Volume percent (%) "as received" 15 minutes 2 hours 4 hours 10 hours Particle breakage Particle deformation 0 1 2 3 4 5 6 7 8 0.1 1 10 100 Particle diameter ( m)Volume percent (%) "as received" 15 minutes 2 hours 4 hours 10 hours Particle breakage Particle deformation Figure 4-1. Volume percent part icle size distributions for expe riments using 1.0 mm grinding media at 2000 rpm. Bimodality in the particle size distribution is indicative of particle deformation and particle break age occurring simultaneously.

PAGE 96

96 0 1 2 3 4 5 6 7 8 9 0.1 1 10 100 Particle diameter ( m)Volume percent (%) "as received" 15 minutes 2 hours 4 hours 10 hoursParticle deformation 0 1 2 3 4 5 6 7 8 9 0.1 1 10 100 Particle diameter ( m)Volume percent (%) "as received" 15 minutes 2 hours 4 hours 10 hoursParticle deformation Figure 4-2. Volume percent part icle size distributions for the experiments with 1.5 mm grinding media at 1000 rpm, a visible shift in the par ticle size distribution to larger sizes is seen due to particle deformation.

PAGE 97

97 6 8 10 12 14 16 18 0 0.5 1 1.5 2 2.5 Media SIze (mm)Mean particle size ( m) 1000 rpm 1500 rpm 2000 rpm Mill Speed 6 8 10 12 14 16 18 0 0.5 1 1.5 2 2.5 Media SIze (mm)Mean particle size ( m) 1000 rpm 1500 rpm 2000 rpm Mill Speed Mill Speed Figure 4-3.Maximum mean particle size achieved for each milling experiment versus media size, the 1.5 mm milling media produced the largest amount of deformation.

PAGE 98

98 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 0.5 mm zirconia media 1.0 mm zirconia media 1.5 mm zirconia media 2.0 mm zirconia media Figure 4-4. Particle deformation versus milling time at a milling speed of 1000 rpm for varying grinding media sizes.

PAGE 99

99 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 0.5 mm zirconia media 1.0 mm zirconia media 1.5 mm zirconia media 2.0 mm zirconia media Figure 4-5. Particle deformation versus milling time at a milling speed of 1500 rpm for varying grinding media sizes.

PAGE 100

100 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 0.5 mm zirconia media 1.0 mm zirconia media 1.5 mm zirconia media 2.0 mm zirconia media Figure 4-6. Particle deformation versus milling time at a milling speed of 2000 rpm for varying grinding media sizes.

PAGE 101

101 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm a) 0.5 mm media b) 1.0 mm media d) 2.0 mm media c) 1.5 mm media 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm 0 2 4 6 8 10 12 14 16 0500010000150002000025000300003500040000 Milling time (sec)Mean particle diameter ( m) 1000 rpm 1500 rpm 2000 rpm a) 0.5 mm media b) 1.0 mm media d) 2.0 mm media c) 1.5 mm media Figure 4-7. Mean particle size versus milling tim e as a function of rotation rate, a) 0.5 mm grinding media, b) 1.0 mm grinding media, c) 1.5 mm grinding media, and d) 2.0 grinding media. Table 4-1. Calculated stress intensity for experime nts performed in this study. The table indicates that it is possible to achieve equivalent energies and different milling conditions. Table 4-2. Maximum mean particle size of milli ng experiments. Even though the kinetic energy of some of the milling experiments are equivalent the maximum particles sizes differ. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 3 21 69 160 1500 6 46 160 370 2000 10 82 280 660 Stress intensity *10^-5 (joule) Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 11 11.0 +/0.6 16 12 1500 13.0 +/0.3 9 15 12 2000 11 9.4 +/0.8 14 11 Maximum mean particle diameter at failure ( m)

PAGE 102

102 FbFb b Rb Aluminum particles Media Media FbFb b Rb Aluminum particles Media Media Figure 4-8. Contact between coll iding elastic spheres. Aluminum particles are caught in the contact area between the media.

PAGE 103

103 Table 4-3. Force, radius of contact area, and stress acting on the milling media as a function of media size and rotation rate (cal culated using Hertz theory). Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 10003.5143256 15005.7235292 20008.13273130 Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 10001.42.94.35.7 15001.73.45.06.7 20001.93.85.77.5 Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 10008300830083008300 15009700970097009700 200011000110001100011000 Force of milling media (N) Radius of contact area of milling media *10^-5 (m) Stress acting on milling media (MPa) Table 4-4. Number of particles stressed in a single collision between grinding media and the stress exerted on an individual particles. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 100073067119 1500 10 41 92 164 2000 13 52 116 206 Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 1115 279 124 70 1500 948 237 105 59 2000 845 211 94 53 Stress acting on an individual particle (MPa) Number of 4.2 m diameter "as received" aluminum particles in contact area Table 4-5. Stress frequency as calculated from Kwades model. The frequency is the number of times milling media collide with each other. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 54 14 6 3 1500 81 20 9 5 2000 110 27 12 7 Stress frequency (Hz) *10^8 in 60 minutes of mi lling

PAGE 104

104 Table 4-6. The time required to reach the maximu m particle diameter. As milling speed increases the time it takes to obtain the maxi mum particle diameter decreases. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 36000 14400 36000 28000 1500 14400 1800 14400 5400 2000 14400 3600 3600 2700 Milling time at breakage (second) Table 4-7. Total number of particle compressi ons in 60 minutes, as calculated using stress frequency and number of particles, Equation 4-2 and 4-6. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 4 4 4 4 1500 8 8 8 8 2000 14 14 14 14 Number of particle compressions in 60 minutes *10^10 (Hz) Dp,0Dp,ftp,f Dp,0Dp,ftp,f Figure 4-9. Diagram of the change in dimensions of particle and the measurements used to calculate stress and strain.

PAGE 105

105 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm mediaa) c) d) b) 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm mediaa) c) d) b) Figure 4-10. Stress-strain behavior for experiments performed at 1000 rpm a) 0.5 mm media, b) 1.0 mm media, c) 1.5 mm medi a, and d) 2.0 mm media.

PAGE 106

106 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm mediaa) c)d) b) 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm mediaa) c)d) b) Figure 4-11. Stress-strain behavior for experiments performed at 1500 rpm a) 0.5 mm media, b) 1.0 mm media, c) 1.5 mm me dia, and d) 2.0 mm media.

PAGE 107

107 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 0.5 mm mediaa) c)d) b) 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 0.5 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 2.0 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 1.0 mm media 0 50 100 150 200 250 0123 StrainStress (MPa) 1.5 mm media 0 50 100 150 200 250 00.511.522.5 StrainStress (MPa) 0.5 mm mediaa) c)d) b) Figure 4-12. Stress-strain behavior for experiments performed at 2000 rpm a) 0.5 mm media, b) 1.0 mm media, c) 1.5 mm me dia, and d) 2.0 mm media. Table 4-8. Strain energy per particle for deformation study. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 10001.61.52.21.7 15001.91.32.11.8 20001.61.221.7 Strain energy per particle *10^-8 (joule)

PAGE 108

108 Table 4-9. Total strain energy for deformati on study. The 1.5 mm media milling experiments resulted in the largest amount of stain energy. Rotation rate (rpm)0.5 mm media1.0 mm media1.5 mm media2.0 mm media 1000 3900 3700 5300 4200 1500 4400 3100 5000 4300 2000 3800 2970 4700 4000 Total strain energy for each milling experiment (joule) Table 4-10. List of constants for equa tions used to calculate milling power 0.533 0.00286 kg/(m*s) 0.028 m 785 kg/m^3 6000 kg/m^3 0.09 m 0.06 m 0.05 m 1000, 1500, 2000 rpm Values Constants 0.533 0.00286 kg/(m*s) 0.028 m 785 kg/m^3 6000 kg/m^3 0.09 m 0.06 m 1000, 1500, 2000 rpm Values Constants c b D H N L L b

PAGE 109

109 0 20 40 60 80 100 120 0 500 1000 1500 2000 Milling time (sec)Milling efficiency, (%) 0.5 mm media, 1000 rpm 1.0 mm media, 1000 rpm 1.5 mm media, 1000 rpm 2.0 mm media, 1000 rpm Figure 4-13. A plot of milling e fficiency as a function of milling time at 1000 rpm, for all media sizes.

PAGE 110

110 0 5 10 15 20 25 30 35 40 0 500 1000 1500 2000 Milling time (sec)Milling efficiency, (%) 0.5 mm media, 1500 rpm 1.0 mm media, 1500 rpm 1.5 mm media, 1500 rpm 2.0 mm media, 1500 rpm Figure 4-14. A plot of milling e fficiency as a function of milling time at 1500 rpm, for all media sizes.

PAGE 111

111 0 5 10 15 20 25 0 500 1000 1500 2000 Milling time (sec)Milling efficiency, (%) 0.5 mm media, 2000 rpm 1.0 mm media, 2000 rpm 1.5 mm media, 2000 rpm 2.0 mm media, 2000 rpm Figure 4-15. A plot of milling e fficiency as a function of milling time at 2000 rpm, for all media sizes.

PAGE 112

112 0 10 20 30 40 50 60 70 80 90 100 0500010000150002000025000300003500040000 Milling time (sec)Milling efficiency, ) 1.5 mm media, 2000 rpm 1.5 mm media, 1500 rpm 1.5 mm media, 1000 rpm Efficiency at failure 0 10 20 30 40 50 60 70 80 90 100 0500010000150002000025000300003500040000 Milling time (sec)Milling efficiency, ) 1.5 mm media, 2000 rpm 1.5 mm media, 1500 rpm 1.5 mm media, 1000 rpm Efficiency at failure Figure 4-16. A plot of milling efficiency as a function of milling time for the 1.5 mm at varying rotational rates.

PAGE 113

113 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 1000 10000 100000 Milling time (sec)Milling efficiency, ) 1.5 mm media, 2000 rpm 1.5 mm media, 1500 rpm 1.5 mm media, 1000 rpm Figure 4-17.A plot of milling efficiency at failure right side of Figure 4-13 used for comparison of efficiencies as a function of rotational rate. Table 4-11.Milling efficiency as a function of percent amount of strain the material has experienced at 1000 rpm and varying medi a size. The 1.5 mm media reaches over 60 % of the maximum strain while milling at 37% efficiency. The last few percent of the maximum strain result in the most inefficient milling. Percent of total strainMill efficiencyPercent of total strainMill efficiencyPercent of total strainMill efficiency 2 383 71 46 43 7 50 2 57 9 76 17 69 0.8 76 3 83 4 91 0.4 92 1 91 2 0.5 mm media 1.0 mm media 1.5 mm media

PAGE 114

114 Table 4-12. Experimental design fo r central composite design 1. Factor 1 Factor 2Response RunA: Rotation rate (rpm)B: Media size (mm)Strain at failure 1 1500 1.5 2.89 2 1000 1 2.41 3 2000 1.5 2.72 4 1500 1 1.95 5 1000 1 2.26 6 1500 0.5 2.70 7 2000 0.5 2.31 8 2000 1 2.17 9 2000 1 1.88 10 1000 1.5 3.04 11 1000 0.5 2.36 12 1500 0.5 2.63 Table 4-13. Analysis of variance for the central composite design 1 described in Table 4-12. Sum ofDegrees ofMeanFp-value SourceSquaresFreedomSquareValueProb>F Model 1.22250.2447.1410.0165 A: Rotation rate0.12310.1233.5790.1074 B: Media Size0.26110.2617.6130.0329 AB 0.01810.0180.5320.4931 A^2 0.00910.0090.2760.6182 B^2 0.81210.81223.7150.0028

PAGE 115

115 1000 1250 1500 1750 200 0 0.50 0.75 1.00 1.25 1.50Strain at failur e 2.17 2.34 2.34 2.52 2.52 2.70 2.88Rotation rate (rpm)Media size (mm) 1000 1250 1500 1750 200 0 0.50 0.75 1.00 1.25 1.50Strain at failur e 2.17 2.34 2.34 2.52 2.52 2.70 2.88Rotation rate (rpm)Media size (mm) Figure 4-18.Contour plot for central composite design 1, a minimum strain of 2.17 can be seen at approximately 0.9 mm milling media.

PAGE 116

116 1000 1250 1500 1750 2000 0.50 0.75 1.00 1.25 1.50 1.9 2.2 2.5 2.8 3.1 A: Rotation rate a size Media size (mm) Rotation rate (rpm) Strain at failure 1000 1250 1500 1750 2000 0.50 0.75 1.00 1.25 1.50 1.9 2.2 2.5 2.8 3.1 A: Rotation rate a size Media size (mm) Rotation rate (rpm) Strain at failure Figure 4-19. Surface response for central compos ite design 1, a trough exists at the 1.0 mm media size.

PAGE 117

117 1 000 1250 1500 1750 2000 0 50 0.75 1.00 1.25 1.50 0.094 0.11075 0.1275 0.14425 0.161 M edia size Media size (mm) Rotation rate (rpm)Standard error (units of strain) 1 000 1250 1500 1750 2000 0 50 0.75 1.00 1.25 1.50 0.094 0.11075 0.1275 0.14425 0.161 M edia size Media size (mm) Rotation rate (rpm)Standard error (units of strain) Figure 4-20. Standard error associated with centr al composite design 1, error increases at the limits of the study.

PAGE 118

118 Table 4-14. Experimental design fo r central composite design 2. Factor 1 Factor 2Response RunA: Rotation rate (rpm)B: Media size (mm)Strain at failure 1 1000 1 2.41 2 1000 1 2.26 3 1500 1.5 2.98 4 1500 2 2.55 5 2000 2 2.39 6 2000 1.5 2.72 7 1000 2 2.50 8 1500 1 1.95 9 2000 1 1.88 10 2000 1 2.17 11 1000 1.5 3.04 Table 4-15. Analysis of variance for the central composite design 2 described in Table 4-10. Sum ofDegrees ofMeanFp-value SourceSquaresFreedomSquareValueProb>F Model 1.21650.24311.9210.008 A: Rotation rate0.10610.1065.1750.072 B: Media Size0.23010.23011.2540.020 AB 0.01110.0110.5460.493 A^2 0.00510.0050.2650.629 B^2 0.71710.71735.1290.002

PAGE 119

119 1000 1250 1500 1750 2000 1.00 1.25 1.50 1.75 2.00Strain at failure 2.15 2.33 2.50 2.50 2.68 2.68 2.85 Media size (mm) Rotation rate (rpm) 1000 1250 1500 1750 2000 1.00 1.25 1.50 1.75 2.00Strain at failure 2.15 2.33 2.50 2.50 2.68 2.68 2.85 Media size (mm) Rotation rate (rpm) Figure 4-21. Contour plot for cen tral composite design 2, show s a maximum at approximately the 1.6 mm media.

PAGE 120

120 1000 1250 1500 1750 2000 1.00 1.25 1.50 1.75 2.00 1.9 2.2 2.5 2.8 3.1 A: Rotation ra t B: Media size Media size (mm) Rotation rate (rpm) Strain 1000 1250 1500 1750 2000 1.00 1.25 1.50 1.75 2.00 1.9 2.2 2.5 2.8 3.1 A: Rotation ra t B: Media size Media size (mm) Rotation rate (rpm) Strain Figure 4-22. Surface response for central composite de sign 2, a peak can be seen at intermediate media size.

PAGE 121

121 B: Media size 1 0 0 01 2 5 01 5 0 01 7 5 02 0 0 0Interaction 1.70 2.05 2.40 2.75 3.10 Rotation rate (rpm)Strain at failure 2.0 mm milling media 1.0 mm milling mediaB: Media size 1 0 0 01 2 5 01 5 0 01 7 5 02 0 0 0Interaction 1.70 2.05 2.40 2.75 3.10 B: Media size 1 0 0 01 2 5 01 5 0 01 7 5 02 0 0 0Interaction 1.70 2.05 2.40 2.75 3.10 Rotation rate (rpm)Strain at failure 2.0 mm milling media 1.0 mm milling media Figure 4-23. A plot of the interaction between the media size and rotation rate, it can be seen that there is only a slight interaction at low.

PAGE 122

122 1000 1250 1500 1750 2000 1.00 1.25 1.50 1.75 2.00 0.073 0.08675 0.1005 0.11425 0.128 A: Rotation rate B: Media size Media size (mm) Rotation rate (rpm) Standard error (units of strain) 1000 1250 1500 1750 2000 1.00 1.25 1.50 1.75 2.00 0.073 0.08675 0.1005 0.11425 0.128 A: Rotation rate B: Media size Media size (mm) Rotation rate (rpm) Standard error (units of strain) Figure 4-24. Standard error associated with central composite design 2.

PAGE 123

123 Figure 4-25. Transmission electron microscope im age of as received aluminum particle.

PAGE 124

124 Figure 4-26. Diffraction pa ttern obtained from particle in Figure 4-20.

PAGE 125

125 Figure 4-27. Transmission electron microscope imag e of sectioned flake af ter 4 hours of milling.

PAGE 126

126 Figure 4-28. Single crystalline diffraction pattern of sectioned flake after 4 hour sof milling.

PAGE 127

127 CHAPTER 5 CASE STUDY: MATERIAL SELECTI ON AND DEVELOPMENT FOR INFRARED OBSC URANTS Introduction U.S. Army has used obscurant smokes and clouds as a defense against spectral targeting for many years. Recent advancements in tec hnology have led to communication and detection systems functioning in the infrared and microw ave region of the electro magnetic spectrum and new obscurant materials are needed to meet these wavelengths. According to globalsecurity.org, virtually a ll nations have access to sensor s and guided munitions utilizing the infrared portion of the electromagnetic sp ectrum. The widespread availability of these devices has led the United States Military to search for the best countermeasures to mitigate their effectiveness. Janon Embury measured the mass extinction co efficients of many materials and later theoretically calculated the extin ction coefficients for materials depending on their shape and size [7, 8, 79-82]. The calculations indicated th at electrically conduc tive high aspect ratio particles with the minor dimension in the nanomet er size range and the major dimension of about 1/3rd the wavelength of light would exhibit many fo ld improvement over the materials currently available. The method of choice for this purpose is to disperse fine metal flakes or rods into the atmosphere between friendly forces and the enemy. With higher obscuration efficiency, less material will be required, thus reducing the weight or volu me of obscurant munitions carried by soldiers or combat vehicles. Attenuation by a particulate cloud follows Beer-Lambert law, Equation 5-1, where Io and I are incident and attenuated light intensities, is the extinction coefficient, C is the concentration of the particles in the light path of length L

PAGE 128

128 lnoI CL I 5-1 In Equation 5-1, is dependant on size and optical propert ies of the particles and its units may be expressed on either mass or volume basis w ith the use of corresponding concentration and path length units. The goal of this study was to obtain a four fold improvement over the existing IR obscurant material. In order to achieve this goal the army uses Equation 5-2 to gauge the performance of the obscurant. Figure of Merit = FoM = YP 5-2 In Equation 5-2, is the extinction coefficient on the mass basis, is the material density, Y is the yield factor that is a measure of the ease of a material to be dissemina ted and is the ratio of the mass of particulates successfu lly disseminated and the total mass of the material in a grenade. The factor P is the packing factor that measures the mass fraction of material in the fill volume. From these terms it can be seen that the figure of merit is a function of the type of material, material properties, loading of the grenade and dissemination of the powder in the grenade. The Particle Engineering Research Center applied a systems approach to achieve this goal by improving every factor in the fi gure of merit equation (FOM). However, this study will only focus on the selection and development of obscurant materials, for example, thinner flakes will be used that will increase the area of coverage at a cost of less material. The other aspects of the FOM equation can be found in a thesis by Tedeschi [64]. Development of New Ob scurant Materials In order to achieve the improvement goal, th e FOM has to increase four fold. The right hand terms in the Equation 5-2 are not mutually exclusive. First, the is dependant on the material. Higher values are achieved by using high aspect ratio and electrically conductive

PAGE 129

129 particles. Based on previous work, metal flakes a nd rods were expected to be ideal candidates [7, 8, 79-82]. The best shape and size of material s are flakes and rods with a major dimension greater then 3 m and a minimum dimension of less than 100nm and lower the better. These high aspect ratio materials (major/minor dimension) are difficult to manufacture. The cost of the material and processing into desired shape and dimensions must be minimized to reduce the overall cost of the grenade. Metal nanorods with these dimensional requirements are difficult and expensive to produce. Entanglement of fibers also results in poor dispersi on and dissemination. On the other hand, flakes nearing these dimensi ons have been manufactur ed in large quantities for decades by milling. Metal ductility plays a signifi cant roll in the aspect ratio of the flake that can be produced. It is possible to obtain a higher aspect ratio w ith softer materials, which means that softer materials can be made thinner, t hus increasing the aspect ratio. Higher density materials have been shown to have greater dissemination efficienci es, resulting in higher performance. It was possible then to construct a chart of relevant material properties (ductility, electrical conductivity, and density) to aid in the selec tion of candidate materials. The relative magnitudes of density, ductility, and conductivity for several materials can be seen in Figure 5-1. Theoretically, a material with high values in each of these areas would be and ideal infrared obscurant. The fi gure shows that one such materi al is copper, which has a high ductility, conductivity and density. Another possible candidate would be silver; however, silver is a precious metal and is expensive. Alumin um is also a good candidate with respect to conductivity and ductility; however, its low densit y will hurt its overall performance on a volume basis. It is also possible to us e alloys of some of these materi als to increase performance. As shown above the data presented in Figure 5-1 is a useful tool for selecting possible materials, however, density, ductility, and conductiv ity are not the only material properties that

PAGE 130

130 are important to material selection. As mentioned earlier, cost of the ra w and processed material is also important. Aluminum has been found to be pyrophoric, which means that it is possible to ignite the aluminum powder. Copper powder does not form a passivation later, which means it will eventually oxidize to copper oxide whic h will reduce its conductivity. There are also environmental concerns for all the materials. Addi tional work in coating the material also being performed in this project, however, this work is beyond the scope of this specific study. Several materials were selected for proces sing from Figure 5-1. A comparison of the experimentally measured performance of different materials is shown in Figure 5-2. This figure shows that copper gives the highest perfor mance followed by silver-coated copper and aluminum. The reason for the high performance of copper is due to its high ductility, high density and high conductivity. The performance of aluminum is relatively low because of its low density (roughly a third that of copper). Even though tin has a high ductility and high density, it has a low performance due to it low electrical conductivity. A wet stirred media milling process was used to produce high aspect ratio flakes from the raw material. This is the same process that was the focus of this dissertation, however, it was found that milling just beyond the point of partic le breakage gave the best obscurant material, this result will be discussed later in this chapter. The milling equipment and procedures for this study are the same that were used in the dissert ation work with the exception that most of the experiments in the case study involved longer m illing times. Several of the characterization techniques discussed in Chapter 3 were devel oped specifically for the measurements of the figure of merit of the materials. However, the next section will give a brief review of these techniques and how they were used to calculate FOM.

PAGE 131

131 Material Characterization and Obscurant Performance Measurement Method Particle size characterization was performed on the samples using the Coulter LS Model 13320 laser diffraction particle sizer. The specific surface area measurements are obtained from a Quantachrome NOVA 1200 gas sorpti on instrument. Images of the particles were taken using a JEOL 6330 field emission sca nning electron microscope. The FOM values were calculated from transm ission data obtained from the FTIR using a dual pass transmission (DPT) cell and from measurem ent of the amount of material per area. In this measurement the powder was dispersed onto DPT slides and the transmission of IR light through the slide was measured in a reflection mode. By this method the degree of dispersion and material performance could be evaluated. The amount of powder used was within the measurement error of lab balance so for a highe r degree of accuracy Indu ctively Coupled Plasma Spectroscopy (ICP) was used and is described in deta il below. This allowed the accurate measurement of masses well below those seen for these experiments. The material performance was determined based on the material FOM, or the improvement of the new materials compared to a standard material (brass fl akes) supplied by the US Army. The standard procedure used for FOM measurem ents is as follows. The dry powders were dispersed in a Galai vacuum dispersion cham ber onto a glass microscope slide and a DPT slide. The powder was allowed to settle for forty five minutes after which the DPT slide was placed in the FTIR for measurement. The gla ss microscope slide was placed in a Nalgene bottle with a 1M HCL acid solution to dissolve th e metal in order to accurately measure the concentration using the ICP. Mass was calcula ted by measuring the concentration of Al in solution, and normalized by the area of the glass slide. The Perkin-Elmer Plasma 3200 Inductively Coupled Plasma Spectroscopy (ICP) system is equipped with two monochromators covering th e spectral range of 165-785 nm with a grated

PAGE 132

132 ruling of 3600 lines/mm. The ICP operates on the principle of atomic emission by atoms ionized in the argon plasma. Light of specific wavelengt hs is emitted as electrons return to the ground state of the ionized elements, quantitatively identifying the species present. The system is capable of analyzing materials in both organic an d aqueous matrices with a detection limit range of less than one part per million. This detecti on limit is well below the concentrations used in these experiments and therefore allowed accu rate measurement of the material mass disseminated. Production of Infrared Obscurant In this project several metallic materials were measured, their FOM was compared with the current obscurant material brass which has a FOM of 1. This study evolved over a several years, the first material chosen to work with was the current brass material. Further milling was used to improve the current material. The starting materi al was brass flake obtained from US Bronze. The flake was produced by dry ball milling with the addition of stearic acid to prevent oxidation. The stearic acid coating also aids in the di spersion of the brass by making the brass surface hydrophobic. A screening design of experiment (DOE) was followed to determine the effects of media loading and rotational rate on the massed based FOM of brass. The milling media used in all experiments was high-density yttria stabilized zirconia with a mean particle size of 500 microns. The rotational rate was varied from 700, 2000, and 3300 rpm. The experimental parameters for this study can be found in Table 5-1. The results of the initial screening design can be seen in Figure 5-3. It was determined that rotational rate had very little effect on FOM va lue and showed that as media loading increased FOM increased. It should be noted that the effect of rotational rate on FOM is similar to that observed in the theoretical study of deformation of aluminum. An increase of 25% in the FOM

PAGE 133

133 was obtained from the screening design. The DOE determined that larger media loading increased the FOM. Using the information obtai ned form the DOE, Figure 5-4 gives FOM for the as received material, and that milled with a media volume of 480 ml of milling media. It can be seen in Figure 5-4 that a much larg er FOM is obtained when the media volume was increased However, it can also be seen that at high wavelengths the FOM value for high media loading decreases more rapidly. The decrease at high wavelengths is due to the reduction in particle size form milling. When the particle sizes become le ss than 1/3 the size of that wavelength no longer scatters light at the wa velength [7, 8]. This design enabled future experiments to set the media loading at a high concentration. The second design used in this study was a full factorial (22) design; this involves choosing two factors (rotational rate and me dia size) and two levels per factor for a total of 4 experiments. Table 5-2 is a list of the milling parameters used in the 22 statistical design. The media loading in this experiment was fixed at 300 ml, which is the maximum capacity of the mill. This design measured responses of FOM and particle size at milling times of 6,12, 18, and 24 hrs. The surface response of figure merit versus the design factors can be seen in Figure 5-5. At increasing media sizes the figure merit increased It was determined that the 1.5 mm media resulted in the highest figure of merit. The increase in the FOM is due to the increased deformation of particles. It is important to note that particle sizes roughly 1/3 the wavelength of IR light do not effectively scatter light at that wavelength [8]. Particle breakage in the mill can reduce the effectiveness of the material at longer waveleng ths. Figure 5-6 gives the effect of milling time on the particle size distribution of brass. As received brass has a mean particle size of 10 m with a nearly normal distribution and as milling time increases the particle size decreases. The particle

PAGE 134

134 size distribution becomes bi-modal, which is indi cative of a breakage mech anism. After 24 hr the mean particle size was approximately 2.3 m. This means that the flakes should no longer effectively scatter light above 7 m. Figure 5-7 is a plot of FO M versus wavelength at varying milling times. As milling time increased the FOM increased, this is due to increase in particle deformation. At a milling time of 18 hrs the FOM begins to decrease at the higher wavelengths, due to particle breakage, the aff ect of particle breakage on the d ecrease in more dramatic at 24 hrs of milling time. At 24 hrs the particle size is 2.3 m and the wavelength where the FOM begins to decrease is 7 m, this agrees with the calculations made by Embury. The combination of these designs and the theoretical studies es tablished a basis for empirically relating milling parameters to FOM. The designs determined th at high media loadings and that the 1.5 mm grinding media resulted in the highest performing flake. Rotation rate was found to have only a minimal affect on FOM. These re sults correlate with the theoretical study of particle deformation. Media density was not investigated in the st udy of deformation of aluminum particles, however, density is a component of all the models. Theoretically me dia density should affect the stress acting on the particle and in turn affect particle deformation. Experiments were performed to mill copper with stainless steel, zirconia, a nd tungsten carbide grindi ng media (densities of 9.8, 6.0, and 16 g/cc, respectively), Figure 58 gives these results. Tungsten carbide media results in the largest amount of deformation be fore breakage, followed by stainless steel, and zirconia. That indicates the highest density medi a, tungsten carbide densit y (16 g/cc), performed the best and, zirconia, the lowest density media (6 g/cc) was the worse.

PAGE 135

135 Summary The above studies proved that is possible to improve the obscuration properties of existing materials and to produce new and improved obscu rants by stirred media milling. It was found that high aspect ratio flakes could be produced by milling soft metals, that the softer the metal the thinner the flake that could be produced. A relationship based on fundamental material properties was derived for selecti ng materials that would produce the best infrared obscurant. The relationship suggested that c opper, aluminum, and silver would yield the best flakes from milling. The study found that larger media produced a larger thinner flake; the results indicate that the smaller media fractured the materials cau sing the materials to have a lower aspect ratio that was detrimental to their performance. Denser media yield higher aspect ratio in shorter milling time. The shorter milling times would ultima tely result in economic saving if the process was scaled up. Milling speed did not appear to have any effect on the quality of the product; however, greater rotational rate resulted in much faster deformation. Greater rotational rates would result in energy and time saving if the pro cess was scaled up. During this study particle size was closely monitor by sample from the milling during the experiments. All the experiments showed and initial increase in the particle size at short milling times due to deformation followed by a decrease due to particle breakage. This st udy found that a final opt imum particle size of approximately 5 m yielded the highest performing obscurant.

PAGE 136

136 Iron Titanium Tungsten Zinc Tin Brass Al-Bronze Aluminum Copper Silver 0 10 20 30 40 50 60 Density (g/cc) Conductivity *10^-7 (ohm-cm) Ductility (%) Figure 5-1.A chart of relevant material properties of materials for use as an infrared obscurant. High values in each property are preferred for candidate materials.

PAGE 137

137 Figure 5-2. The performance of various milled me tals, the copper has the highest performance to due to the properties being the hi ghest in all the desired area. Table 5-1. Experimental parameters used in the statistical design of experiment used to determine milling parameters. ExperimentMedia volume (ml)Rotation rate (rpm) 1 0 2000 2 40 700 3 40 3300 4a 60 2000 4b 60 2000 4c 60 2000 5 120 700 6 120 3300 7 160 2000

PAGE 138

138 R o t a t i o n r a t e ( r p m ) M e d i a v o l u m e ( m l )Figureof Merit R o t a t i o n r a t e ( r p m ) M e d i a v o l u m e ( m l )Figureof Merit Figure 5-3. The results of the initial design of experiment, indicating th at higher media loading increase the performance of the material.

PAGE 139

139 0.7 0.9 1.1 1.3 1.5 1.7 1.9 358101315wavelength (um)Figure of Merit Brass As Received Run 8 ( 480ml/700r p m ) As received brass Media volume of 480 ml 0.7 0.9 1.1 1.3 1.5 1.7 1.9 358101315wavelength (um)Figure of Merit Brass As Received Run 8 ( 480ml/700r p m ) As received brass Media volume of 480 ml Figure 5-4. Plot of FOM for as received and gr ound brass at specific m illing condition of 480ml media volume and 700 rpm rotational speed Table 5-2. Experimental parameters used in the full 2 factorial (22) statistical design of experiment. ExperimentMedia size (mm)Rotation rate (rpm) 10.5500 20.51500 3a1.01000 3b1.01000 41.5500 51.51000

PAGE 140

140 1.67 2.04 2.40 2.78 3.13 500 750 1000 1250 1500 0.50 0.75 1.00 1.25 1.50 M e d i a s i z e ( m m )R o t a ti o n r a t e ( r p m )Figure of Merit 1.67 2.04 2.40 2.78 3.13 500 750 1000 1250 1500 0.50 0.75 1.00 1.25 1.50 1.67 2.04 2.40 2.78 3.13 500 750 1000 1250 1500 0.50 0.75 1.00 1.25 1.50 M e d i a s i z e ( m m )R o t a ti o n r a t e ( r p m )Figure of Merit Figure 5-5. The results of the affect of varying media size and ro tation rate on the performance of a milled brass flake.

PAGE 141

141 0 1 2 3 4 5 6 7 8 0.1 1 10 100 Particle diameter ( m)Volume percent (%) As Received (10.3um) 6hr (7.5um) 12h (5.4um)r 18hr (3.4um) 24hr (2.3um) Figure 5-6. Differential volume particle size distribution at increasing milling times. 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 3579111315 Wavelength( m)Mass extinction (m^2/g) 6hr 12hr 18hr 24hr Figure 5-7. Figure of merit versus IR wavelength for increasing milling times.

PAGE 142

142 0 5 10 15 20 25 30 0 5101520 Milling time (hr)Mean particle diameter ( m) Tungsten Carbide Stainless Steel Zirconia Figure 5-8. Mean particle size as a function of milling time for different density milling media, zirconia (6 g/cc), stainless steel (9.8 g/cc), and tungsten carbide (16 g/cc).

PAGE 143

143 CHAPTER 6 SUMMARY, CONCLUSIONS, AND FUTURE WORK Summary and Conclusions The goal of the study was to develop a f undamental and predictable understanding of particle deformation during stirred media milling. This study confirmed that particle deformation during stirred media milling is a complex phenome non. It is difficult to understand and predict the interaction of milling parameters with each other and the material being processed. Novel characterization techniques and methods were used to understand the deformation process. Statistical analysis and design of experiments were used to inte rpret data, develop predictive models, and develop screening experiments for determining the most significant milling parameters and their effect on deformation. Relationships between milling parameters, particle deformation, and milling efficiency were establ ished based on the fundamental theories of contact mechanics and mechanical behavior of materials. Finally, knowledge gained from this study was then used to develop a material fo r the US Army to obscure infrared light. In order to achieve the goals of this study a systematic approach was used to study the process including: Development of characterizati on techniques and sample prep aration methods capable of measuring high aspect ratio particles Determine the milling parameters that have the most significant affect on particle deformation Establish a relationship, based on fundamental theories between, milling parameters and particle deformation Calculate milling efficiency with respect to milling parameters Develop milling models capable of predicting particle deformation To achieve these objectives it was necessary to develop particle si zing techniques capable of measuring high aspect particles. The predomin ate characterization techni que used in this study

PAGE 144

144 to measure particle size was light scattering, which was designed for measuring spherical particles. The results proved that light scatte ring was capable of measuring the major dimension of high aspect ratio particles. The case study required the ability to measure the capability of metallic flakes to obscure IR light. The measurement needed to be representative of how the metallic flakes would perform in an aerosolized cloud in a combat environment. In order to obtain this measurement three challenges had to be addressed: Obtain a dry dispersible powder Disperse powder Characterize the dispersed powder A dry powder was obtained through supercriti cal drying. This technique involved the exchange of the solvent (alcohol ) to liquid carbon dioxide. Followi ng the solvent exchanges, the sample was then dried using the supercritical drying technique described in Chapter 3, this method yielded an easily dispersible powder. The sample was then disseminated onto a dual pass transmission slide and placed into and FTIR for characterization. These procedures allowed for a measurement of grenade performance that clos ely relates to that obs erved in the field. Several statistical designs of experiments were performed in this study. Preliminary designs were used to determine which milling parameters were significant and required a more rigorous study and which variables could be held constant. The preliminary designs determined that a maximum media loading was important for causing the maximum amount of particle deformation. The early designs also showed that rotation rate and media size could affect media loading and required further inves tigation to determine their exact behavior. Milling studies at varying media size and rotation rate showed that the 1.5 mm milling particles resulted in the highest amount of deformation. Ro tation rate was found to affect th e rate of particle deformation

PAGE 145

145 and had little to no influence over the maximum atta inable deformation. Hertzs theory was used to explain the relationship between rotation ra te, media size and particle deformation. Hertzs theory was used to determine the st ress acting on the media, the theory was then expanded to determine the stress acting on the particle by calculati ng the number of particles in the contact area between colliding milling media. The data indicated that as media sizes decreased the stress acting on the particle incr eased. The 1.5 mm media resulted in the highest amount of deformation, due to the stress being close to the yield stress of the aluminum material. The media sizes smaller than 1.5 mm caused stresses greater than the yield stress and thus resulted in an early onset of particle breakage. Models that describe media frequency of contacts failed to describe the rotation rate effects on deformation obser ved in this study. New models were derived that demonstrated that media size had no effect on the number of times that particles were stressed during the experiment. The new models also determined that rotation rate effect the number of times a particle was stre ssed during the milling experiments. The empirical data agreed with the results of the rotational rate models, however, further experimentation would be required to obtain a strong valid ation of the rotation rate models. The strain energy was calculated from assump tions of the conservation of volume of the particles during deformation. For the first time it was possible to calculate milling efficiency based on the strain energy. The study found that milling efficiency was the greatest for the 1.5 mm media and that efficiency decreased with m illing time. A rotation rate of 2000 rpm resulted in the highest efficiency at the point of particle break age. The study also determined the reasons for inefficiency during the milling processes. Previous studies had proposed that viscous friction and heat generated in the milling process were the leading cause of inefficiency. This study found that elastic collision which resulted in no deformation led to the inefficiency.

PAGE 146

146 Statistical design of experiment was used to develop an empirical model capable of predicting particle deformation during milling ba sed on rotation rate and media size. The model also validated the fundamental theories used to explain the behavior of milling parameters on deformation. The statistical design showed that the interactions between media size and rotation rate had a minimal effect on deformation. Th is result was also obtai ned by the fundamental theory based on contact mechanics which showed the stress acting on the particle only varied slightly with rotation rate. The statistical desi gn predicted that a medi a size of 1.5 mm would yield the most strain in the particle. For the first time relationships between particle deformation and milling parameters based on fundamental theories were es tablished using Hertzs theory and collision frequency. The relationship between milling parameters and the st ress acting on the particle could be used as a scaling parameter to improve and design milling pro cesses. Results of this investigation showed that it was possible to predict milling efficiency which could be directly applied to many milling operations currently in use. The case study proved that the understanding developed in this study could be directly applied to developing pr oducts used in real world applications. Future Work This study developed a knowledge base for milli ng processes where particle deformation is desired. However, most milling processes involve particle breakage, a suggestion for future work would be to determine correlation between the m odels developed in this study and the particle breakage models. An example of this is that many researches have recognized the non linear breakage rate kinetics in a milling process [83, 84]. A similar relation would exist between stress acting on a particle and particle size as a function of milling time. As particle size decreases in a breakage process, the number of particles in th e contact area would increa se, thus decreasing the stress acting on the particle. It is also known that as particle size decreas e the number of defects

PAGE 147

147 in the particle also decreases, requiring higher stress to fracture the particle. These results could be the reason that a so called grinding limit is observed in many milling process. One method of testing the applicability of th e models developed in this stu dy to breakage processes would be to reduce the number concentrati on of particles during the milling process, thus maintaining a constant level of stress acting on the particles. Another area of interest for future work would be to investigate the e ffect of media density on particle deformation. Density has a direct e ffect on the stress acting on the particle, it would be interesting to determine if the effect on deformation is as predicted by the models. One area, which requires more study, is that of the freque ncy of media and particle contacts. The results obtained in this study were not sufficient and conc lusive enough to make up a significant theory. More empirical data and short sampling intervals would be re quired to obtain a more robust model capable of describing the effect of media size on frequency of particle contact. One of the most novel aspects of this study wa s the ability to obtain an actual value for milling efficiency. It is possible to test th is method of measuring milling efficiency by manipulating other milling variables such as viscosity and determine the effect on energy utilization. It would also be interesting to further investigate the mechanisms of milling efficiency such as the result of elastic collisions and their effect on inefficiency. Further work is needed to construct milling models that are capable of predicting deformation of materials based on milling parameters and material properties. Finally, the goal of this study was to devel op a fundamental understanding of particle deformation during milling and develop a predictive model. In most respects this goal was accomplished, however, it is important to develop a model that includes a time parameter. It may

PAGE 148

148 also be possible to construct a model that incl udes a material property parameter and properties of milling media, such as, media density, and hardness.

PAGE 149

149 LIST OF REFERENCES 1. American Society for Metals Interna tional, Powder Metal Technologies and Applications, Materials Park, Ohio, 1998. 2. F. Muller, R. F. Polke, Powder Technology 105 (1999) 2-13. 3. M. Senna, H. Kuno, Journal of the Amer ican Ceramic Society 54 (1971) 259-&. 4. P. S. Gilman, J. S. Benjamin, Annual Revi ew of Materials Scie nce 13 (1983) 279-300. 5. S. H. Hong, B. K. Kim, Mate rials Letters 51 (2001) 139-143. 6. P. Wissling, dmg Business Media Ltd (1999). 7. J. Embury, Emissive versus attenuating smokes, in: R. a. T. Directorate (Ed.) Edgewood Chemical and Biological Center, 2002. 8. J. Embury, Maximizing infrared extinction coeffi cients for metal discs, rods, and spheres, in: R. a. T. Directorate (Ed.) Edgew ood Chemical and Biol ogical Center, 2002. 9. D. M. Bigg, D. E. Stutz, Polymer Composites 4 (1983) 40-46. 10. C. B. Keemer, in: Communication (Ed.) Siblerline Manufacturing Co., Inc., 2005. 11. H. R. Zoz, R. Reichardt, H.U. Benz, A. Nadkami, G. Wagner, Advances in Powder Metallurgy and Particul ate Materials (1999). 12. J. Y. Huang, Y. K. Wu, H. Q. Ye, Mate rials Science and Engineering a-Structural Materials Properties Microstructure and Processing 199 (1995) 165-172. 13. S. Hwang, C. Nishimura, P. G. McCormic k, Scripta Materialia 44 (2001) 1507-1511. 14. R. Watanabe, H. Hashimoto, G. G. Lee, Materials Transactions Japan Institute of Metals 36 (1995) 102-109. 15. G. G. Liversidge, K. C. Cundy, International Journal of Pharmaceutics 125 (1995) 91-97. 16. C. Frances, C. Laguerie, B. Mazzarotta, T. Veccia, Chemical Engineering Journal 63 (1996) 141-147. 17. S. Mende, F. Stenger, W. Peukert, J. Schwedes, Powder Technology 132 (2003) 64-73. 18. S. Mende, F. Stenger, W. Peukert, J. Schwed es, Journal of Materials Science 39 (2004) 5223-5226. 19. L. G. Austin, P. Bagga, Powder Technology 28 (1981) 83-90.

PAGE 150

150 20. D. Eskin, O. Zhupanska, R. Hamey, B. Moudgil, B. Scarlett, Aiche Journal 51 (2005) 1346-1358. 21. A. Jankovic, Minerals Engineering 16 (2003) 337-345. 22. A. S. Kwade, J., KONA 52 (1997) 91-101. 23. E. Bilgili, Scarlett, B., Estimation of the selection and breakage parameters from batch grinding: a novel full numerical scheme, in: Proceedings of the AIChE Annual Meeting, San Francisco, 2003. 24. J. A. Herbst, Y. C. Lo, International Jour nal of Mineral Proces sing 44-5 (1996) 209-221. 25. A. Kwade, Powder Technology 105 (1999) 14-20. 26. H. Rumpf, Particle Technology, chapman and hall, 1990. 27. W. S. McCabe, J Harriott, P., Unit Operations of Chemical Engineering, McGraw-Hill inc, New York, 1993. 28. C. L. Prasher, Crushing and Grinding Process Handbook, wiley, 1987. 29. R. Pond, Getting metals into usable fo rm: Cold working, in: Metallurgy for nonmetallurgist, ASM International, Metals Park. 30. R. A. R. Reed-Hill, Physical Metallur gy Principles, pws publishing company, 1994. 31. W. Callister, Materials Science and Engin eering an Introduction, John Wiley and Son, New York, 2000. 32. A. J. R. Inigoraj, R. Narayanasamy, K. S. Pandey, Journal Of Materials Processing Technology 84 (1998) 143-148. 33. S. Lou, D. O. Northwood, Journal Of Mate rials Engineering And Performance 3 (1994) 344-349. 34. H. J. Kleemola, M. A. Nieminen, Metall urgical Transactions 5 (1974) 1863-1866. 35. R. P. Reed, Cryogenics 12 (1972) 259-&. 36. T. G. Nieh, P. S. Gilman, J. Wadswort h, Scripta Metallurgica 19 (1985) 1375-1378. 37. C. Tome, G. R. Canova, U. F. Kocks, N. Chri stodoulou, J. J. Jonas, Acta Metallurgica 32 (1984) 1637-1653. 38. M. A. Meyers, A. Mishra, D. J. Benson, Progress in Materials Science 51 (2006) 427556. 39. J. A. Hines, K. S. Vecchio, Acta Materialia 45 (1997) 635-649.

PAGE 151

151 40. C. R. Bemrose, J. Bridgwater, Powder Technology 49 (1987) 97-126. 41. Y. S. Cheong, A. D. Salman, M. J. Hounslow, Powder Technology 138 (2003) 189-200. 42. H. Cho, M. A. Waters, R. Hogg, Internati onal Journal of Minera l Processing 44-5 (1996) 607-615. 43. K. Schnert, International Journal of Mineral Processing 44-5 (1996) 1-16. 44. K. Schnert, in: Summer school in winter, Particle Engineering Research Center, Gainesville, Florida, 2003. 45. K. Kendall, Nature 272 (1978) 710-711. 46. K. U. Schnert, H. Kleman, W., in: Proceed ings of the 2nd International Conference of Fracture, Brighton, England, 1969, pp. 474-482. 47. K. Schnert, About the particle breakage and the possibities and limits for improving size-reduction processes, in : The International Symposium on Powder Technology, Powder Technology, Kyoto, Japan, 1981. 48. J. S. Benjamin, Metallurgical Transactions 1 (1970) 2943-&. 49. J. B. Fogagnolo, F. Velasco, M. H. Robe rt, J. M. Torralba, Materials Science And Engineering A-Structural Materials Propertie s Microstructure And Processing 342 (2003) 131-143. 50. S. X. Mao, N. A. McMinn, N. Q. Wu, Mate rials Science And Engineering A-Structural Materials Properties Microstructure And Processing 363 (2003) 275-289. 51. S. K. Moothedath, K. V. S. Sastry, Powder Technology 75 (1993) 89-96. 52. A. W. Weeber, A. J. H. Wester, W. J. Haag, H. Bakker, Physica B & C 145 (1987) 349352. 53. J. Guerrero-Paz, D. Jaramillo-Vigueras, Nanostructured Materials 11 (1999) 1123-1132. 54. Booz, Metal Flake Production, in: US (Ed.) vol 4065060, US, 1976. 55. M. M. Moshksar, S. M. Zebarjad, Iranian Journal of Science and Technology 23 (1999) 239-250. 56. H. R. H. Zoz, R. Reichardt, H.U. Benz, A. Nadkami, G. Wagner, Advances in Powder Metallurgy and Particul ate Materials (1999). 57. L. P. Sung, M. E. Nadal, M. E. McKnight, E. Marx, B. Laurenti, Journal of Coatings Technology 74 (2002) 55-63. 58. S. H. Hong, B. K. Kim, Mate rials Letters 57 (2002) 275-279.

PAGE 152

152 59. M. Riffel, J. Schilz, Journal of Materials Science 33 (1998) 3427-3431. 60. J. A. Rodriguez, J. M. Gallardo, E. J. He rrera, Journal of Materials Science 32 (1997) 3535-3539. 61. A. Jankovic, Minerals Engineering 14 (2001) 1177-1186. 62. T. Tanaka, KONA 13 (1995). 63. L. Blecher, A. Kwade, J. Schwedes Powder Technology 86 (1996) 59-68. 64. S. Tedeschi, Improving aerosol dispersion through a fundamental understanding of interparticle forces, in: Materials Science a nd Engineering, vol Ph. D., University of Florida, Gainesville, 2007. 65. S. Brunauer, P. H. Emmett, E. Teller, J ournal Of The American Chemical Society 60 (1938) 309-319. 66. T. Allen, Particle Size Measurem ents, Chapman & Hall, London, 1997. 67. D. W. Fuerstenau, A. Z. M. Abouzeid, Inte rnational Journa l of Mineral Processing 31 (1991) 151-162. 68. M. Z. He, Y. M. Wang, E. Forssberg, Powder Technology 147 (2004) 94-112. 69. J. Zheng, C. C. Harris, P. Somasunda ran, Powder Technology 86 (1996) 171-178. 70. M. Z. He, Y. M. Wang, E. Forssberg, Powder Technology 161 (2006) 10-21. 71. R. Hamey, Production of organic pigment na noparticles by stirred media milling, in: Materials Science and Engineering, vol Master s of Science, Florida, Gainesville, 2003. 72. A. Kwade, L. Blecher, J. Schwedes Powder Technology 86 (1996) 69-76. 73. A. H. Tullo, Chemical & Engineering News 82 (2004). 74. K. L. Johnson, Contact mechanics, Cambridge University Press, 2003. 75. R. A. Broadbent, Callcott, T.G., Journa l of Institute of Fuel 29 (1956) 528-539. 76. S. Nagata, Mixing: Principles and Applications, Wiley, New York, 1975. 77. C. Alamprese, L. Datei, Q. Semeraro, J ournal Of Food Engineer ing 83 (2007) 629-636. 78. D. C. Montgomery, Design and Analysis of Experiment, John Wiley and Son Inc, New York, 2004. 79. J. Embury, Technical Report US Army E dgewood Chemical and Biological Center (1993).

PAGE 153

153 80. J. Embury, Technical Report US. Ch emical Systems Laboratory (1994). 81. J. Embury, Technical Report US Army E dgewood Chemical and Biological Center (1994). 82. J. Embury, Technical Report US Army E dgewood Chemical and Biological Center (1998). 83. L. G. Austin, Powder Technology 5 (1971) 84. E. Bilgili, Jepes, J., Scarlett, B., A non-linear framework for population balance modeling of batch grinding: beyond first-order kinetics., in: Proceedings of the 2nd International Conference on Population Ba lance Modeling, Valencia, Spain, 2004b.

PAGE 154

154 BIOGRAPHICAL SKETCH Rhye Garrett Ham ey was born in Medi na, Ohio. He attended Cloverleaf High School. After high school he obtained a degree in chemical engineering at The Ohio State University. While at Ohio State, Rhye atte nded a class on particle technology instructed by visiting Professors Brian Scarlett and Dr. Brij Moudgil. Thr ough continued contact with Professor Scarlett, Rhye became interested in part icle technology and chose to pursue a graduate degree from the Department Materials Science and Engineering at the University of Florida. He performed his st udy while working at the Particle Engineering Research Center. In 2005, Rhye obtained a Masters degree in Materials Science and Engineering. After receiving his Masters R hye decided to conti nue his Ph.D. study under the supervision of Dr. Hassan El-Shall.