<%BANNER%>

Analytic Methods for Predicting Biosettlement on Patterned Surfaces

Permanent Link: http://ufdc.ufl.edu/UFE0024908/00001

Material Information

Title: Analytic Methods for Predicting Biosettlement on Patterned Surfaces
Physical Description: 1 online resource (134 p.)
Language: english
Creator: Long, Christopher
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: attachment, biofouling, biomaterials, fouling, marine, materials, model, prediction, predictive, settlement, topography, ulva, wettability
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: Marine organism fouling of surfaces has significant impact on our environment and the economy. Increased fuel use due to drag costs approximately $600 million annually in the United States alone. The efficiency of marine vessels substantially decreases due to fouling. Toxins in some antifouling paints accumulate in the marine environment and produce negative effects on the marine ecology. There is a critical need for effective non-toxic, anti-fouling, marine coatings that reduce operational costs and the overall environmental impact of ocean vessels on the environment. Our approach is to investigate the interaction between the wettability of surfaces with the response of fouling organisms. One of the ways the wettability can be influenced is through the use of topography. Since the topographies have directionality, the direction dependence of the wettability was determined on several microscale topographies that have previously shown antifouling potential. Breaking microscale ridges into the discontinuous features in the antifouling topographies reduced the anisotropies in the contact angles, but did not eliminate anisotropy. The number of distinct features in the design and the area fraction of the topographic features were found to influence settlement of the fouling alga Ulva linza. A biosettlement model, refined from previous work, predicts the settlement of Ulva linza to three previously untested surfaces. These surfaces significantly reduced the settlement of these spores in vitro by up to 78%. The attachment of another species of fouler, the diatom Navicula perminuta, was reduced by approximately 35% on several surfaces that reduced Ulva linza settlement. The Navicula cells responded differently to the topographies than the Ulva linza spores. A mapping technique was developed to determine the two-dimensional settlement pattern of cells on the topographical surfaces. This technique revealed and quantified several preferential locations for Ulva linza settlement on engineered topographies. The characteristics of these locations can be further investigated to elucidate the driving factors for the interaction of these cells. Other applications, such as the medical devices and tissue scaffolds, could benefit from investigating the localized interactions between various cells and surface patterns.
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 Christopher Long.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Brennan, Anthony B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024908:00001

Permanent Link: http://ufdc.ufl.edu/UFE0024908/00001

Material Information

Title: Analytic Methods for Predicting Biosettlement on Patterned Surfaces
Physical Description: 1 online resource (134 p.)
Language: english
Creator: Long, Christopher
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: attachment, biofouling, biomaterials, fouling, marine, materials, model, prediction, predictive, settlement, topography, ulva, wettability
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: Marine organism fouling of surfaces has significant impact on our environment and the economy. Increased fuel use due to drag costs approximately $600 million annually in the United States alone. The efficiency of marine vessels substantially decreases due to fouling. Toxins in some antifouling paints accumulate in the marine environment and produce negative effects on the marine ecology. There is a critical need for effective non-toxic, anti-fouling, marine coatings that reduce operational costs and the overall environmental impact of ocean vessels on the environment. Our approach is to investigate the interaction between the wettability of surfaces with the response of fouling organisms. One of the ways the wettability can be influenced is through the use of topography. Since the topographies have directionality, the direction dependence of the wettability was determined on several microscale topographies that have previously shown antifouling potential. Breaking microscale ridges into the discontinuous features in the antifouling topographies reduced the anisotropies in the contact angles, but did not eliminate anisotropy. The number of distinct features in the design and the area fraction of the topographic features were found to influence settlement of the fouling alga Ulva linza. A biosettlement model, refined from previous work, predicts the settlement of Ulva linza to three previously untested surfaces. These surfaces significantly reduced the settlement of these spores in vitro by up to 78%. The attachment of another species of fouler, the diatom Navicula perminuta, was reduced by approximately 35% on several surfaces that reduced Ulva linza settlement. The Navicula cells responded differently to the topographies than the Ulva linza spores. A mapping technique was developed to determine the two-dimensional settlement pattern of cells on the topographical surfaces. This technique revealed and quantified several preferential locations for Ulva linza settlement on engineered topographies. The characteristics of these locations can be further investigated to elucidate the driving factors for the interaction of these cells. Other applications, such as the medical devices and tissue scaffolds, could benefit from investigating the localized interactions between various cells and surface patterns.
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 Christopher Long.
Thesis: Thesis (Ph.D.)--University of Florida, 2009.
Local: Adviser: Brennan, Anthony B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2010-02-28

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0024908:00001


This item has the following downloads:


Full Text

PAGE 1

ANALYTIC METHODS FOR PREDICTING BIOSETTLEMENT ON PATTERNED SURFACES By CHRISTOPHER JAMES LONG 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 2009 1

PAGE 2

2009 Christopher James Long 2

PAGE 3

To my wonderful wife, Victoria 3

PAGE 4

ACKNOWLEDGMENTS My progress would not be possi ble without the help of many people. I extend my sincere gratitude to my advisor, Dr. Anthony Brennan, and I thank him for his guidance, support, and motivation for the past four years. I tha nk my committee membersDr. Christopher Batich, Dr. Simon Phillpot, Dr. Ronald Baney, Dr. Henry Hess, and Dr. Patrick Antonellifor their guidance and for serving on my committee. I am gr ateful to Dr. Eugene Goldberg and Dr. Susan Sinnott for substituting on my committee. Many thanks go to Jennifer Wrighton for her friendship and assistance. I sincerely thank Al Ogden for his providing me his expertise and training in the UF Na nofabrication Center. My colleagues and friends were crucial to helping me throughout my time at the University of Florida. Id like to extend special thanks to my mentor, Dr. James Schumacher, for his friendship and guidance. For providing vital feedback, assistance, and guidance, I thank my group members both past and present: Dr. Miche lle Carman, Dr. Thomas Estes, Dr. Leslie Wilson, Dr. Clifford Wilson, Kenneth Chung, Chel sea Magin, Scott Cooper, Liwen Jin, James Seliga, Julian Sheats, David Jacks on, Angel Ejiasi, and Jack Chen. I thank Sean Royston for his exceptional work and diligence. I thank Paul A. C. Robinson II for his work in developing new sample production methods. My progress could not be possible without the help of my collaborators: Dr. Maureen Callow, Dr. James Callow, Dr. John Finlay, and Angel Sampson. I thank Dr. Jeffrey Fergus for his guidance early in my academic career. Donald Parker is thanked for his guidance and mentorship. I cherish the friendship of Dr. Iris Enriquez Schumacher, Dr. Amin Elachchabi, Dr. Shema Freeman, Jennifer and Mike Wohlwend, Leah and Darren Gibson, James Ristow, David Reigada, and Caley Burke. I want to extend sp ecial thanks to Ken C hung and Dr. Cliff Wilson for their friendship. 4

PAGE 5

My research and progress would not have b een possible without th e love and support of my family. My wife, Victoria, has been by my side through the ups and downs of the process and provides me support and encouragement. My parents, Allen and Karen, my brother Scott, and my sister Jenny have shaped me so much throughout my life, and continue to provide abundant love and support. I th ank my grandparents, Ada and Pe te, and Mary Anne and Robert for providing me with so much love and passing down the values and qualities that have been incredibly valuable. I thank my father-in-law Mi ke Salazar, my brother-in-law James Salazar, and the rest of my wifes fam ily for accepting me into their family and providing me with so much support. I thank the rest of my fa mily for their continued love and support. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................4 LIST OF TABLES................................................................................................................. ..........9 LIST OF FIGURES.......................................................................................................................10 ABSTRACT...................................................................................................................................12 CHAPTER 1 INTRODUCTION................................................................................................................. .14 Introduction................................................................................................................... ..........14 Specific Aims..........................................................................................................................15 Specific Aim 1: Identify Magnitude of Static and Dynamic Contact Angle Anisotropy on Engineered Topograp hies with 2 m Dimensions...............................15 Specific Aim 2: Predict the Settlement of Ulva linza Spores on an Engineered Topography..................................................................................................................15 Specific Aim 3: Reduce the Attachment of Diatoms Using Engineered Topography....16 Specific Aim 4: Identify Preferential Settlement Sites for Ulva linza Spores on an Engineered Topography...............................................................................................16 2 BACKGROUND................................................................................................................... .17 Introduction................................................................................................................... ..........17 Process of Marine Biofouling.................................................................................................18 Economic Impacts of Marine Biofouling...............................................................................19 Historical Perspective......................................................................................................... ....20 Non-Eluting Strategies to Combat Fouling............................................................................23 Foul Release Coatings.....................................................................................................23 Release Free Antifouling Coatings..................................................................................25 Chemical approaches to rele ase free antifouling coatings.......................................25 Natural antifoulants..................................................................................................26 Natural topography...................................................................................................27 Artificial topography................................................................................................28 Multi-scale surfaces..................................................................................................33 Models for Fouling............................................................................................................. ....35 Surface-Wide vs. Local Characteristics..................................................................................35 Local Position Preference.......................................................................................................36 Summary.................................................................................................................................37 3 POTENTIAL FOR TUNABLE STATIC AND DYNAMIC CONTACT ANGLE ANISOTROPY ON GRADIENT MICROSC ALE PATTERNED TOPOGRAPHIES.........39 6

PAGE 7

Introduction................................................................................................................... ..........39 Materials and Methods...........................................................................................................46 Topographical Surfaces...................................................................................................46 Silicon Wafe r Molds........................................................................................................47 Material............................................................................................................................47 Contact Angle Characterization......................................................................................47 Prediction of Wettability Character istics on Topographical Surfaces.............................48 Results and Discussion......................................................................................................... ..49 Predicted vs. Experimental Measurements......................................................................49 Anisotropy of Measurements..........................................................................................53 Comparison of Anisotropies on Gradient Surfaces to Ridges.........................................56 Regression of Slip Angle.................................................................................................57 Conclusions.............................................................................................................................59 4 DEVELOPMENT AND IMPLEMENTATI ON OF A PREDICTIVE MODEL FOR BIOSETTLEMENT OF THE GREEN ALGA ULVA LINZA ...............................................62 Introduction................................................................................................................... ..........62 Materials and Methods...........................................................................................................63 Refinement of Engineered Roughness Index..................................................................63 Surface Designs...............................................................................................................64 Nomenclature...........................................................................................................64 Evaluating S ........................................................................................................65 Evaluating n ..........................................................................................................65 Sample Preparation..........................................................................................................66 Settlement Assay with Spores of Ulva linza...................................................................67 Results.....................................................................................................................................68 Regression Analysis........................................................................................................68 Evaluating S ...............................................................................................................71 Evaluating n.................................................................................................................73 Conclusions.............................................................................................................................80 5 NAVICULA SETTLEMENT AND RELEASE ON SHARKLET AF-BASED TRANSLATIONALLY SYMMETRIC TOPOGRAPHIES..................................................81 Introduction................................................................................................................... ..........81 Materials and Methods...........................................................................................................84 Surface Designs...............................................................................................................84 Silicon wafer molds.........................................................................................................86 Material............................................................................................................................87 Initial Attachment of Navicula........................................................................................88 Strength of Navicula Attachment....................................................................................88 Microscopy......................................................................................................................88 Results and Discussion......................................................................................................... ..89 Pillar Topographies.........................................................................................................89 Sharklet AF-Based Topographies................................................................................91 Conclusions.............................................................................................................................99 7

PAGE 8

6 TECHNIQUE FOR MAPPING SET TLEMENT ON SURFACES AND THE PREFERENTIAL SETTLEMENT LOCATIONS OF ULVA LINZA ALGAL ZOOSPORES.......................................................................................................................100 Introduction................................................................................................................... ........100 Materials and Methods.........................................................................................................101 Topographical Surfaces.................................................................................................101 Sample Preparation........................................................................................................102 Ulva Linza Assay..........................................................................................................103 Image Analysis..............................................................................................................104 Asymmetric units...................................................................................................104 Vector notation.......................................................................................................106 Spore centroid locations.........................................................................................108 Conversion of centroids to vector system..............................................................109 Settlement Maps and Smooth Histograms....................................................................111 Results and Discussion......................................................................................................... 112 Conclusions...........................................................................................................................119 7 CONCLUSIONS AND FUTURE WORK...........................................................................121 Conclusions...........................................................................................................................121 Future Work..........................................................................................................................123 Feature Dimensions.......................................................................................................123 New Sharklet AF Designs.........................................................................................125 Settlement Mapping.......................................................................................................126 Summary........................................................................................................................127 LIST OF REFERENCES.............................................................................................................128 BIOGRAPHICAL SKETCH.......................................................................................................134 8

PAGE 9

LIST OF TABLES Table page 3-1 Contact angle data for smooth PDMSe..............................................................................50 3-2 Calculated surface parameters for topographies................................................................50 3-3 Predicted and measured static, advanc ing, and receding water contact angles on topographies.......................................................................................................................51 3-4 Predicted and Measured Slip Angles for Topographically Modified Surfaces.................52 3-5 Comparison of contact angle anisotr opy on gradient topographical surfaces and ridges..................................................................................................................................56 4-1 Dimensions and prediction variables for surf aces used to evaluate the effect of area fraction of feature tops ( S) on spore settlement................................................................71 4-2 Settlement on surfaces used to evaluate the effect of solid surface fraction ( S) compared to the settlement predic ted by the biosettlement model....................................71 4-3 Dimensions and prediction va riables for n-Series surfaces...............................................73 4-4 Settlement on n-Series surfaces compared to values predicted by the biosettlement model..................................................................................................................................73 4-5 Comparison of spore settle ment relative to smooth...........................................................75 5-1 Characteristics of square p illar surfaces varying in spacing..............................................85 5-2 Characteristics of surfaces varying in width and spacing..................................................87 5-3 Characteristics of surfaces varying in n.............................................................................87 5-4 Correlation coefficients (R2) for regressions of Navicula response to Engineered Roughness Index................................................................................................................97 6-1 Surface designations for Gradient, Sh arklet AF, and Recessed Sharklet AF surfaces............................................................................................................................103 6-2 Spore preference on Gradient surfaces............................................................................113 6-3 Spore preference on Sharklet AF and Recessed Sharklet AF surfaces....................115 9

PAGE 10

LIST OF FIGURES Figure page 2-1 Scanning electron micrograph of Sharklet AF surface..................................................30 2-2 Hierarchical structure with spore-spec ific Sharklet AF and 20 m wide ridges...........34 2-2 Spores settled between adjacent diamonds on the Sharklet AF surface.......................37 3-1 Schematic of static and dynamic c ontact angles of se ssile liquid drops............................40 3-2 Scanning electron micrographs of the gradient and ridge surfaces...................................46 3-3 Illustration of the convention used to denote directionality...............................................49 3-4 Angular dependence of A) static contact angle, B) advanci ng contact angle, C) receding contact angle, D) contact a ngle hysteresis, and E) slip angle.............................53 4-1 Scanning electron micrographs of A) Sharklet AF (+1.8SK2x2) and B) Recessed Sharklet AF (-2.0SK2x2) surfaces.................................................................................65 4-2 Sharklet AF surfaces varying in th e distinct number of features (n).............................66 4-3 Regression of Ulva settlement to ERII for dataset from Schumacher et al. 2008 showing poor correlation (R2 = 0.4)..................................................................................69 4-4 Regressions of spore settlement with ERIII showing historical data correlates with ERIII...................................................................................................................................70 4-5 Regression of settlement data from 3 studies with ERIII (R2=0.88)..................................72 4-6 Regressions of n-Series Sharklet AF surfaces confirming the number of distinct features in the design is a key pa rameter in predicting settlement....................................74 4-7 Performance of the biosettlement model in predicting spore settlement...........................75 4-8 Biosettlement model correlated to four spore settlement assays.......................................76 5-1 Optical images of square pillar topographical surfaces.....................................................84 5-2 Scanning electron micrographs of A) +2.8SK2x5, B) +2.7SK2x2, C) +2.4SK5x2, and D) +2.5SK10x2...........................................................................................................86 5-3 Scanning electron micrographs of surfaces with varying numbers of distinct features, n..........................................................................................................................................87 5-4 Navicula response to Pillar topographies...........................................................................90 10

PAGE 11

5-5 Navicula response to surfaces varying in width and spacing............................................93 5-6 Navicula response to n-Series topographies......................................................................94 5-7 Comparison of Navicula response among all Sharklet AF surfaces..............................95 6-1 Scanning electron micrographs of Gr adient and Sharklet AF surfaces.......................102 6-2 Symmetry elements of the Sharklet AF surface and latti ce vector notation................105 6-3 Three asymmetric units for the Sharkl et AF, indicated by the outlined shape............106 6-4 Asymmetric units for evaluated surfaces.........................................................................106 6-5 Representative Cartesian coordinates for vector transformation on an image of Ulva spores settled on the Sharklet AF surface....................................................................107 6-6 Threshold operation of settlement image.........................................................................108 6-7 Process of transformation from lattice vectors A) La to B) a to C) a to D) a ............110 6-8 Example of duplicating asymmetric un it and spore locations for presentation...............111 6-9 Spore settlement maps on Gradient surfaces...................................................................112 6-10 Spore settlement maps for Sharklet AF and Recessed Sharklet AF surfaces..........114 6-11 Smoothed histograms of Ulva spore settlement maps on Gradient surfaces...................117 6-12 Smoothed histograms of Ulva spore settlement maps on Sharklet AF (+2.5SK2x2) and Recessed Sharklet AF (-2.0SK2x2) surfaces........................................................118 7-1 Sharklet AF surfaces overexposed to shrink spacing between features.......................124 7-2 Sharklet AF patterned ebeam resist using ebeam lithography.....................................124 7-3 Etched silicon wafer molds for Sharklet AF surfaces with 200 nm feature width and spacing.......................................................................................................................125 7-4 New Sharklet AF surfaces...........................................................................................126 11

PAGE 12

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 ANALYTIC METHODS FOR PREDICING BIOSETTLEMENT ON PATTERNED SURFACES By Christopher James Long August 2009 Chair: Anthony B. Brennan Major: Materials Science and Engineering Marine organism fouling of surfaces has si gnificant impact on our environment and the economy. Increased fuel use due to drag co sts approximately $600 million annually in the United States alone. The efficiency of marine ve ssels substantially decreases due to fouling. Toxins in some antifouling paints accumulate in the marine environment and produce negative effects on the marine ecology. Th ere is a critical need for e ffective non-toxic, anti-fouling, marine coatings that reduce operational costs an d the overall environmental impact of ocean vessels on the environment. Our approach is to investigate the interacti on between the wettability of surfaces with the response of fouling organisms. One of the ways the wettability can be influenced is through the use of topography. Since the topog raphies have directionality, th e direction dependence of the wettability was determined on several microscal e topographies that have previously shown antifouling potential. Breaking microscale ridge s into the discontinuous features in the antifouling topographies reduced th e anisotropies in the contact angles, but did not eliminate anisotropy. The number of distinct features in the de sign and the area fraction of the topographic features were found to influence settlement of the fouling alga Ulva linza. A biosettlement 12

PAGE 13

model, refined from previous wo rk, predicts the settlement of Ulva linza to three previously untested surfaces. These surfaces significantl y reduced the settlement of these spores in vitro by up to 78%. The attachment of anothe r species of fouler, the diatom Navicula perminuta was reduced by approximately 35% on se veral surfaces that reduced Ulva linza settlement. The Navicula cells responded differently to the topographies than the Ulva linza spores. A mapping technique was developed to determ ine the two-dimensional settlement pattern of cells on the topographical surface s. This technique revealed and quantified several preferential locations for Ulva linza settlement on engineer ed topographies. The ch aracteristics of these locations can be further investigat ed to elucidate the dr iving factors for the interaction of these cells. Other applications, such as the medical devices and tissue scaffolds, could benefit from investigating the localized interactions between various cells and surface patterns. 13

PAGE 14

CHAPTER 1 INTRODUCTION Introduction All stable surfaces observed to this point foul over time in the marine environment, including glass, metals, stone, pol ymers, shells of organisms, a nd other biological surfaces. This process contains several steps, including conditioning by macrom olecules and fouling by diverse species varying in size.1 This fouling greatly increases drag on marine vessels which leads to increased fuel consumption, reduced speed, and increased the production of gasses that are of environmental concern.2, 3 The increase in fuel is estimated to cost the U.S. Navy and U.S. commercial, fishing, and recreational ve hicles approximately $600 million annually.4 Toxic strategies to reduce fouling ha ve found some success, but these toxins have shown toxicity toward other organisms and in some cases have been banned.5, 6 A variety of approaches have been investigat ed to reduce fouling on surfaces in ways that do not release toxins. One of these approaches is the use of topographic patterns on the surface. Several species of fou ling organismsincluding algae and ba rnacleshave been shown to alter settlement rates in respon se to topographic patterns.7-9 These organisms do not all respond to the same topographies in the same ways. Additiona lly, the interaction betw een the cells and the topographies is not fully understood. Prior to this work, there was not a consistent model capable of predicting biosettlement. As a result of this work, an empirical model that is based on a correlation between topography-altered wettability and biosettlement was refined. The static and dynamic contact angles of topographies are examined because to pography both affects wettability and influences biosettlement. Previous work is built upon to de velop a predictive model for the settlement of the fouling alga Ulva linza. The fouling response of another organism, Navicula perminuta, is 14

PAGE 15

tested using this model. A method to describe preferential settlement to localized regions on the topography is developed. This method can be used to investigate lo calized characteristics of the surface to better understand cell-surface interactions. The role of topographically-altered surface energy on settlement behavior is investigated in this work. Specific Aims Specific Aim 1: Identify Magnitude of Static and Dynamic Contact Angle Anisotropy on Engineered Topographies with 2 m Dimensions Identify if static and dynamic contact angle anisotropy larger than 1 exists on engineered topographies with features 2 m wide and spaced by 2 m. Several engineered topographies developed for anti-fouling purposes were evaluated for static and dynamic contact angles using th e tilting plate method. The channels topography exhibited anisotropy of 5-42 in the various contact angles. Making the features discontinuous decreased the anisotropy to 2-6. The anisotr opy in the static and dynamic contact angles on these surfaces is statistically significant, but was greatly reduced from the values found on the continuous channels of similar dimensions. Specific Aim 2: Predict the Settlement of Ulva linza Spores on an Engineered Topography The settlement of Ulva linza spores on a previously-untes ted engineered topography of poly(dimethyl siloxane) elastomer (PDMSe) will be predicted to within 5% in terms of the percent reduction versus a smooth PDMSe surface. A previously developed model, the Engi neered Roughness Index, was re-examined and modified to fit data for multiple historical Ulva linza settlement data sets. This biosettlement model was then used to predict the settlement relative to smooth on three previously un-tested engineered topographies in tw o separate experiments. The Ulva linza settlement assays were performed by Drs. Maureen Callow, James Callo w, and John Finlay at the University of 15

PAGE 16

Birmingham, UK. Each experiment was used to c onfirm the role of different surface descriptors in the biosettlement model. The biosettlem ent model showed corre lation between surface descriptors and the settlement of Ulva linza spores for four separate studies. Specific Aim 3: Reduce the Attachment of Diatoms Using Engineered Topography A PDMSe Sharklet AF engineered topograp hical surface designed with appropriate width and spacing will reduce the attachment of Navicula perminuta diatoms by at least 30% versus smooth PDMSe. The Sharklet AF engineered topographical design was mo dified by altering the width, the spacing, and the number of dist inct features in the design. Th ese surfaces were evaluated for Navicula perminuta diatom attachment and release by Dr s. Maureen Callow, James Callow, and John Finlay at the University of Birmingham, UK. Surfaces with 2 m width and 2 m spacing reduced the attachment of thes e diatoms by an average of ap proximately 35%, while surfaces with larger widths or larger sp acing between features had simila r attachment of diatoms as the smooth surface. Specific Aim 4: Identify Pref erential Settlement Sites for Ulva linza Spores on an Engineered Topography A mapping technique to identify th e location within the pattern that Ulva linza spores (or other cells) settle will identify locations in whic h the density of settled spores is higher than other locations within the pattern. Images from several Ulva linza settlement assays performed by Drs. Maureen Callow, James Callow, and John Finlay at the Universi ty of Birmingham, UK we re analyzed using a newly developed mapping technique. Preferential se ttlement sites were identified on each of the surfaces, and quantitative analysis was performed. Several preferential settlement sites were not previously identified with observational evidence. 16

PAGE 17

CHAPTER 2 BACKGROUND Introduction Biofoulingthe accumulation of biological mol ecules, cells, and organisms to a surface has implications for many aspects of everyday lives. Bacteria are a concern in nearly every surface implanted in the body, increasing the cost of treatment, the length of stay, and mortality rate.10 Some short term medical devices, such as ur inary catheters, have extremely high rates of disease related to bacteria fouling these devices.11 Oral plaque other oral diseases are caused by bacteria fouling surfaces in the oral cavity.12 Bacteria foul the membra nes of water purification systems for drinking water and wa stewater treatment. Fouling of these membranes reduce output, reduce efficiency, and increase corrosion.13 In the marine environmentthe environment of focus in this workfouling from single cells to multicellular animals creates considerable problems in the shipping industry and other marine applications. The industry of preventing marine fouling ha s recently undergone a major shift: the most effective commercial technology was recently ba nned by the international community due to toxicity. Other technologies that also rely on toxicity are in danger of being similarly banned. Non-toxic strategies to contro l marine biofouling are required not simply for improving current technology, but to take the place of toxicity-based technologies. New technologies must focus on the interaction of the fouling organisms with the surfaces, since the organisms cannot be killed at a distan ce from eluted toxins. The fouling response is affected by surface chemistry, topography, mechanical properties, and other characteristics. The fouling response to topography is of particular interest, as technologies based on topographies have the potential to be combined with othe r technologies based on chemical and compositional 17

PAGE 18

properties. A model for predicting the response of fouling organisms to topographies would provide a means for creating non-toxic su rfaces capable of preventing fouling. Process of Marine Biofouling All stable surfaces observed to this point foul over time in the marine environment, including glass, metals, stone, polymers, shells of organisms, and other biological surfaces. The colonization of the surfaces differs in the spec ies fouling the surfaces and in the rate of colonization. Colonization occurs in four step s, starting with a pris tine surface that becomes colonized with an advanced community.1 Macromolecules condition the surf ace in the first step of colonization. This conditioning step begins as soon as a surface is introduced to the water. The formation of this film is a physical process and is purely thermodynamic. In the second step, bacteria colonize the surface through a variety of driving forces. These driving forces include water motion, cell motility, van der Waals interactions, and elec trostatic intera ctions. Long range motion such as water flow bring the cells near the surface, and short te rm motion through cell motility, Brownian motion, and microturbulance bring the cells into near-contact. Rearrangement of excreted polymers and other processes bring the cell into intimate cont act with the surface. This second step begins within hours of immersion of the surface. 1 The third step of marine biofouling is th e settlement and attach ment of unicellular eukaryotic foulers such as diatoms, and begins on the order of days to weeks. The fourth stage is fouling by macrofoulers, including algae and larval forms of animal foulers. This stage begins on the order of weeks after immersion. The macrof oulers foul the surface through a colonizing form, such as a spore or larva, and develop into much larger organisms.1 The term biosettlement refers to the biological process of a cell or larval stage initially attaching to the surface in a seri es of steps. The organism approaches a surface, senses the 18

PAGE 19

surface, contacts the surface, and makes initial attachment to th e surface. Some organisms have propulsion mechanisms to approach the surface (e.g. flagella), while others rely on gravity and water currents. The organism senses the surface in some manner before or during contact. Initial attachment is often provided by some biological adhesive to anchor the cell or organism. Biosettlement does not mean the cell or organism has strong attachment to the surface and does not provide any indication of attachment strength. Economic Impacts of Marine Biofouling Fouling of ship hulls creates several economic and ecological problem s. Fouling increases the weight of the vessel and increases the frictional resistance from the water. The increased drag and weight require higher fuel consumption.2 Slimes, an early occurring stage of biofouling, increase drag by up to 25%. An increase in resistance of 80% was reported when a 1 mm thick slime was present along with extensive macrofou ling algae. Hard foulers were estimated to increase the drag by up to approximately 85%, and even small surface coverage provides large increases in drag.14, 15 In 2000, when the price of oil was approximately equal to the current range (approximately $40 per barrel), estimations of the increased fuel cost for U.S. military and non-military were evaluated.4 In the United States alone, increased fuel use due to fouling was estimated to be approximately $600 million for both military and non-m ilitary vessels. For the U.S. Naval fleet, the increase in cost was estimated to be approximately $250 million annually. For U. S. nonmilitary vessels, increased fuel use was es timated to cost approximately $300-400 million annually. Additional cost increases come from increased wear on machinery and cleaning time and expense. Fouling can exacerbate corrosion by damaging protective coatings, requiring additional maintenance and costing both time and materials. Biofilms of bacteria a nd algae create anodic 19

PAGE 20

and cathodic regions based on the lo cal concentrations of oxygen a nd other ions. The salt in the water magnifies the corrosive effect s of these cathodic and anodic regions.16 Organisms which have fouled a ship hull are tr ansported to new geographic regi ons causing the spread of nonnative species.2 Historical Perspective Antifouling coatings are surface treatments used to prevent biofouling and have a varied history dating back approximately 3000 years.2 Pitch, wax, and tar were used by ancient cultures, and metal sheathing was used as early as 700 B. C. Copper, a material currently used in antifouling coatings, was used by both the Romans a nd the Greeks in the form of nails to secure lead sheathing. Pitch and tallow were used fo r many centuries to protect hulls. Copper use for antifouling has been reported as early as the 17th century. Widespread use of copper for antifouling occurred in the late 18th century in the British Navy. It wasnt until around 1800 that the antifouling property of copper was due to the dissolution of the metal into the water. The field of antifoulant paints arose as a result of the introduc tion of iron ships. Copper as an antifoulant was not perfect and was largely di scontinued on iron ships due to the corrosive effect of copper on iron. The antifouling propert ies of the highly toxic, broad antifoulant tributyltin (TBT) were discovered in the 1950s after testing a variety of formulations of antifouling paints with limited effectiveness. This highly effective antifoulant was commercialized in the 1960s, and remained the primary compound for use in antifouling until its negative consequences led to its ban in the 2000s by the inte rnational community.2 Despite the problems with tributyltin, this effective antifoulant reduced fuel consumption and decreased emissions. Tributyltin-based antifoulants were estimated in 1999 to save the shipping industry $5.7 billion and 7.2 million tons of fuel annually. Emissions of various environmentally concerning gases such as carbon dioxide and sulfur dioxide from the shipping 20

PAGE 21

industry were greatly reduced by the use of this effective antifoulant. Th e spread of invasive species was decreased by the use of tributyltin by preventing organisms from attaching to vessels moving throughout the worlds harbors.3 Unfortunately, the high toxicity of tributyltin that makes it effective as an antifoulant also makes it troublesome ecologically. Reproduction pr oblems caused by tributyltin nearly destroyed oyster farming in the prominent Arcachon Bay area of France in the late 1970s and early 1980s.17 Imposex, a reproductive problem in which females develop male sexual organs, has been linked to tributyltin in various gastropods. Entire populations of gastropods disappeared when imposex prevented fema les from releasing eggs.18 Tributyltin enters the food chain through abso rption by bacteria and algae or by adsorbing onto particles in the water.19 The toxin accumulates in tissues in fish, birds, and mammals, presenting a threat to animals throughout the food chain. A large number of marine species have exhibited toxic effects from tribut yltin, particularly the early and larval stages. Humans may be exposed to tributyltin through water which contains the comp ound or through consumption of contaminated seafood. Detectable levels of dibuty ltin, a metabolite of tributyltin as well as a toxic compound itself, have been found in human liver samples in various populations.20 The international community agreed in 2001 to ban tributyltin because of the problems with this popular and effective antifoulant. Al ready banned by some countries, tributyltin could not be applied to ships starting in 2003. A comp lete ban on tributyltin including the removal of any of the biocide app lied prior to 2003 took e ffect worldwide in 2008.5 Tin-free biocidal antifouling pa ints have become popular due to the ban on tributyltin. Copper is the primary biocide used in these pa ints, because copper is toxic to a variety of species. However, several species of algae are tolerant to copper, and booster biocides are 21

PAGE 22

required for effective antifouli ng activity. Both diatoms and Ulva two highly important fouling algal speciesform biofilms even on relatively new tin-free biocidal paints.16 Leaching biocides including copper in antif ouling paints have their own concerns. The leaching rate for copper from antif ouling paints is higher than th at for tributyltin, leading to higher environmental concentrations21. Copper has been shown to cause mortality in mollusk larvae at concentrations which are observed in some coastal regions.6 Copper concentrations in some areas with high vessel traffic are higher than federal and state water quality criteria.22 The concentration of zinc pyrithione a common biocide used along w ith copper, is large enough in some regions to effect the developm ent of mussel and sea urchin embryos.23 Copper concentrations in the seawater near some she llfish farms have increased twofold since the 1980s. The elevated copper concentration is approachin g levels which cause faster aging, decreased rates of reproduction, and decreased growth in clams for farming.24 Antifouling paints have been suggested to be th e main source of heavy metals in seawater. These metals, including copper and zi nc, are absorbed by algae and isopods.25 Zinc and copper have synergistic toxicity, leadi ng to difficulty evaluating the full effect of antifouling paints on the ecological environment using laboratory studies.26 Even some supposedly non-toxic commercial coatings leach toxic material into the water.27 These concerns have led some European countries to limit th e copper release from vessels and increased concern from other governments.28 Little information is available on the toxi city of booster biocides or their effect on the environment. The booster biocides used w ith copper may prove to have deleterious long term effects and may not be a safe alternative to tributyltin.29 With increasing concerns over the release of toxic agents into the environment, bans on any biocide-base d antifouling paint are likely to be forthcoming. 22

PAGE 23

Non-Eluting Strategies to Combat Fouling Problems with released biocides and the prospe ct of bans on the use of released biocides have directed fouling control rese arch into new approaches for coatings. The first such approach is to design coatings which prevent strong attach ment of fouling organisms to the surface. These surfaces, called foul release co atings, allow easy cleaning of biofouling from the surface. The movement of a vessel through the water crea tes enough shear force to dislodge fouling organisms in some cases. The second approach to combat fouling without eluting biocides is to create release-free antifouling coatings. These coatings prevent attach ment of foulers via means other than released biocides. Technologies employing th is approach include specialized chemistries, incorporation of bioactive molecules in the surface, surfaces with speciali zed topographies, and surfaces designed to change properties dynamically. Foul Release Coatings Foul release coatings are typically fluoropolym ers or silicones, whose low surface energies and low moduli prevent strong adhesive bonds between foulers and the surfaces. Several problems with foul release coatings exist. Some early foulers like diatoms attach strongly to commercial foul release coatings. Because these coatings do not prevent fouling when the vessel is not moving, fouling communities on a stationa ry vessel may grow and become even more difficult to remove.30 Additionally, foulers respond to the wettability of these surfaces differently. Diatoms are more strongly a ttached to hydrophobic surfaces than hydrophilic surfaces.31 In contrast, Ulva linza algal spores attach in higher numbers on hydrophobic surfaces, but attach more strongly to hydrophilic surfaces.32 Amphiphilic chemistries have been investigated as potential foul-release surfaces. Surfaces that contain both hydrophilic a nd hydrophobic regions were created with the aim of allowing 23

PAGE 24

removal of both Ulva linza and diatoms organisms. These am phiphilic surfaces are typically block copolymers with amphiphilic molecules at tached to one of the blocks. These block copolymers can be created by either modifying sma ll molecular weight polymers prior to further polymerization or by selectively modifying blocks in a block copolymer. The amphiphilic component typically contains both fluorinated regions and ethylene oxide regions. Both diatoms and Ulva algal sporelings had higher removal rate s on the amphiphilic surfaces than on the control surfaces.28 In another study, the diatoms did not release as well from amphiphilic surfaces.33, 34 This discrepancy may be due to the lack of suitable and generally-accepted control surfaces for testing. Dynamic coatings, which are designed to cha nge properties, have been proposed to allow for increased release of foulers. One example of dynamic coatings is the application of polymers which have a lower critical solu tion temperature (LCST) near am bient temperature. The polymer is soluble in water below the LCST, but insoluble above the LCST. In these two states, the surfaces are structurally different and have different wettabilities. Two organisms which prefer surfaces with diffe rent wettabilities were both shown to have high release from one of these dynamic su rfaces based on poly(N-isoproylacrylamide).35 Halomonas marina showed release of 95% during the transition to the hydrated state. Staphylococcus epidermidis showed release of 93% during th e transition to the non-hydrated state. These surfaces were tested with unfiltere d seawater from Puget Sound, and showed release of 95% of the surface coverage of the mixed speci es foulers. These surfaces lasted only a few repetitions before becoming ineffective, and c ontrolled switching may be difficult to implement on most marine surfaces. 24

PAGE 25

Release Free Antifouling Coatings The second approach for new fouling-control co atings is to preven t the initial fouling through a means other than released biocides. This developing area includ es control of fouling using a variety of technologies currently bei ng developed and tested for potential use. Technologies employing this approach have a larg e range and include specialized chemistries, incorporation of bioactive molecules in the surf ace, surfaces with spec ialized topographies, and surfaces designed to change properties. Chemical approaches to release free antifouling coatings One approach to release free antifouling coatings is to immobilize bioc ides in the coating. Quaternary ammonium salts, which are effectiv e at killing a variety of bacteria and other pathogens, have been tethered to crosslinked poly(dimethyl siloxane).36 A stable system was able to incorporate these quaternary ammonium salts w ithout degrading the mate rial or leaching toxic compounds. These surfaces reduced biofilm formation of the marine bacterium Cellulophaga lytica and the diatom Navicula incerta by more than 80%, as determined by colorimetric and chlorophyll extraction as says, respectively. A similar approach to immobilized biocides involves incorporating catalysts that produce antifouling compounds into the surface. McMaster et al. have developed xerogels that incorporate catalysts that prom ote the production of positive halogen species from chemical species existing within seawater.37 The seleniumand tellurium -based catalysts produce hypohalous acid from halide salts an d hydrogen peroxide. These ions exists in natu ral seawater in small concentrations and therefore do not need to be part of the coating itself. Barnacle cyprids of Balanus amphitrite and tubeworm larvae of Hydroides elegans showed lower settlement on these surfaces when a small concentration of H2O2 was added to artificial seawater to mimic than when no H2O2 was added. Ulva zoospores did not, in general, respond to the 25

PAGE 26

concentration of H2O2. Some of the reduction in settlement for the barnacle cyprids and tubeworm larvae may be due to physical changes of the surface by the H2O2 or by biological response to the presence of H2O2 instead of the supposed hypohalous acid products. Chemical treatments to alter th e surface chemical behavior have been attempted to prevent fouling. Poly(ethylene glycol) (PEG) has been used to modify surfaces, because of the ability to prevent adsorption of a vari ety of proteins. Diatoms, Ulva spores, and the marine bacterium Cobetia marina have shown reduction in settlement on su rfaces treated with PEG chemistries. Zwitterionic polymers have shown good resistance to protein fouling, and have the potential to reduce marine fouling.38, 39 Natural antifoulants Certain species of marine organisms have de veloped their own versions of antifoulants. Species of marine bacteria and diatoms have b een shown to inhibit a va riety of other foulers.40 At least 14 marine bacteria sp ecies have been found to produ ce compounds that inhibit the settlement of barnacle or tubeworm larvae. Various algal species are inhibited by specific compounds produced by at least 12 species of ma rine bacteria. Some diatom species produce antifouling compounds targeting various species. These natural products have several problems regarding their use as commercial antifoulants. The bacteria are themselves very difficult to cultivate, with approximately 95 % of strains considered uncultivable. Even when a bacterium is known to produce an antifouling compound, the identity of the activ e compound may not be known. The compounds are naturally produced in such small quanti ties that cultiv ation is not feasible and are too complex to be synthesized commercially. Instead of collecting the antifouling compounds produced by some species and making a coating using the compounds, one proposed approach is to produce a coating that incorporates a constant supply of the co mpounds: the cells themselves.41 Biofilms of certain species of marine 26

PAGE 27

bacteria have been shown to inhibit barnacle larvae, tubeworm larvae, algal spores, other bacteria, and fungi. Unfortunately, no species so far identified inhibits all of these foulers, and a combination of these species would be required. Surfaces that incorporate biological cells to prevent biofouling are called living paints.42 A major problem in formulating a living paint is choosing the correct matrix material. The matrix must be stable in seawater, must allow the encapsulated cells to thrive, and must control the diffusion of the biologically-produc ed antifouling compound. The matr ix must be selected such that all steps in which the cells are present ar e performed in conditions, which will not kill the cells. Any process which requires h eat or solvents may render the liv ing paint ineffective. In one report, encapsulated cells were viable after up to a year of storage, indicating that some formulations are possible to allow for long e nough storage life to allow for manufacture and storage. In the field, these surfaces inhibited fouling for up to 7 w eeks, but the cells died shortly thereafter. Natural topography Some marine organisms have developed speci alized topographies to prevent fouling. Crabs, mussels, the eggcases of dogfish, and brit tle stars all have surfaces that resist fouling.43 Cancer pagurus a crab species, has surfaces with multis cale topographical structures. Features include large circular features approximately 200 m in diameter and smaller, 2 m pillar-like structures. Ridges or ripples are found on the Mytilus edulis a blue mussel, and Scyliorhinus canicula dogfish eggcases. The mussel surface has ripples approximately 1-1.5 m wide, while the eggcases had ridges separated by be tween 15 and 115 m. The brittle stars Ophiura texturata have surfaces with pillar-like structur es approximately 10 m in diameter. The natural topographies of the above surfaces were examined as potential designs for antifouling coatings.44 The topographies were isolated from other potential antifouling factors 27

PAGE 28

and mechanisms by molding the surfaces and cr eating castings. These surfaces all elicited different fouling responses. The crab surface re pelled macrofoulers, while the dogfish eggcase and star repelled microfoulers. The dogfish eggcase and mussel topographies decreased barnacle cyprid settlement, but the effect of topography on barnacle cyprid se ttlement decreased over time. The surfaces of several bivalve shells and molds with a pronounced topography prevented fouling in early exposure times. Even the sanded su rface showed inhibitory effects on fouling for short submersion periods. The grooved shell of Mytilus galloprovincialis remained resistant to fouling after 12 weeks, while the mold of this surface and the sanded surface became fouled. Mollusk and gastropod shells, each with different natural topographies, were shown to have different resistance to fouling and different foul release properties.45 These natural surfaces provide some direction for antifouling coating design, but factors and mechanisms other than topography likely contribute to the antifouling characteristics of these surfaces. Artificial topography Other research has attempted to create new topographical patterns. Artificial topographies allow for the design of surfaces following knowle dge and theories of cell behavior. These surfaces are, in general, easier to define and have the potential to be produced in a scaled-up process. Typically, work begins w ith simple structures such as channels or pillars, and may adapt to involve more complex topographies. The proc ess of engineered topographical design has typically focused on one organism at a time. This approach allows for theories to be developed for the fouling response of each organism to topographies. The eventual goal is to design topographies to prevent fouling by many species Most of the work has been performed on species of Ulva linza barnacles, and diatoms. These specie s range from microfoulers (diatoms) to micro/macrofoulers ( Ulva linza ) to macrofoulers (barnacles). 28

PAGE 29

Ulva linza: The response of the fouling macroalga Ulva linza to engineered topographies has been progressed from relatively simple ri dge/channel topographies to complex topographies involving multiple types of features. The Ulva spores, whose bodies are approximately 5 m in diameter, settled in higher numbers into channels that were at least 5 m wide than on the smooth surface.7 The spores settled in depr essed regions (i.e. bottoms of channels instead of tops of ridges), specifically in locations where the spore could cont act both a wall and the channel floor. In the case of 5 m channels, the spores could come into contact on three sides with the substrate: both sides of the cha nnel walls and the channel floor. Pillars were also evaluated, and the spores settled in higher numbers as the spacing between the 5 m pillars decreased. The spores groupe d in contact with the pillars, similar to the preference to contact at least one wall in the channel surfaces. In both the channels and pillars surfaces, taller (or deeper) features caused th e increase in settlement to be more pronounced. The theory of the critical dimension was deve loped from the data on Ulva spore settlement on ridges/channels and pillars. The authors theori zed that if the spore could penetrate between the features and contact multiple surfaces, it would settle in hi gher numbers, but if the spore could not penetrate the region between the features settlement would be decreased. For the case of the Ulva spores, whose diameters are approxi mately 5 m, features spaced by less than 5 m should reduce settlement. A surface was designed to prevent the Ulva spore from penetrating the topographical features in part to test the critical dimension theory.46 The Sharklet AF is a pattern of repeating rectangles and was designed around the 2 m di mension. This surface is shown in Figure 2-1. The Sharklet AF was the first topography shown to reduce settlement of Ulva spores, reducing Ulva settlement by 86% relative to smooth poly(dimethyl siloxane ) elastomer (PDMSe). Wider 29

PAGE 30

channels increased settlement as was observed in the previous study. The effect of height of the Sharklet AF surface was evaluated, and the height heavily influenced the settlement of Ulva spores.9 Figure 2-1. Scanning electron micr ograph of Sharklet AF surface Further testing of the critical dimension th eory was performed when the Sharklet AF was tested against several other topo graphies with 2 m feature spacing.47 Channels and hexagonally packed pillars 2 m wide and spaced by 2 m and a pattern of both pillars 2 m wide and triangles 10 m long were evaluated for Ulva settlement. All of these surfaces inhibited Ulva spores relative to the smooth surface, as wa s expected from critical dimension. However, not all of the surfaces performed with equal efficiency. The Sharklet AF showed the greatest inhibition, inhibiting spores by 77%, and showing approximately 65% better performance than both the pillars and the channels. The triangle an d pillars topography, containing both triangles and pillars, performed better than the channels and the pillars topographies, but not as well as the Sharklet AF. The Engineered Roughness Index (ERI) theory was developed from the data on the 2 m spaced surfaces.47 The Ulva spore settlement correlated to the ERI, an empirical relation 30

PAGE 31

involving the Wenzel roughness rati o, the area fraction of solid surface, and the degrees of freedom of movement for a spore exploring the surface. The effect of engineering nanoforce grad ients between topographical features was examined using surfaces composed of one or two features existing in the Sharklet AF.48 Six of these surfaces, along with smooth and Sharklet AF, were tested for Ulva spore settlement. The calculated gradient between the forces required to deflect adjacent features 10% varied between 0 nN and 370 nN. The Ulva settlement in general decreased w ith increasing force gradient, but the Sharklet AF, which contains 4 distinct rectangular featur es, outperformed all of the oneand two-element surfaces, despite having gradients between features of only 125 nN. This result indicated that the Ulva spores respond to thes e gradients, and that th e gradients alone are not enough to explain the differences in settlement. This theory has the potentia l to be incorporated into a predictive model, but the theory is not itself predictive. The settlement behavior of the Ulva linza spores on topographies is highly influenced by the dimensions of the topography. The response to topography is more complex than simply preventing the spore from touching the floor of the topography, as evidenced by the differing spore response to the various surfaces with 2 m feature width and 2 m spacing. Further examination of the topographical response to the Engineered Roughness Index is needed to develop a predictive model for Ulva spore settlement on engineered topographies. Diatoms: Information on diatom response to topogr aphy has been limited to a few studies on simple ridge/channel topographies.8 Several species of raphid di atoms were used to evaluate diatom response to these sinusoidal ridge/chann el topographies. The species ranged in width from 1 m to 7 m and in length from 3 m to 14 m. Topographies with ridges 1 m separated by 1 m channels and ridges 2 m separated by 2 m channels were tested along with a smooth 31

PAGE 32

control. The largest two of th e four species were inhibited from attaching to the ridge topographies, with the 2 m wide 2 m spaced ridges outperforming the 1 m wide 1 m spaced ridges. The projections topogra phy provided limited reduction in the attachment of these two largest diatom species. The smallest species did not respond differently among the surfaces. The species intermediately sized attached in highe r numbers on the 1 m wide 1 m spaced ridges than to the smooth control, but the other topographies did not a ffect the attachment. These results led to the Attachment Point Theory : the theory that the attachment of diatoms is determined by the number of points to which the diatom cell can attach.49 This theory attempted to explain the dependence of size on the response to the topographies. An expanded study used sinusoidal ridge/channel topographies va rying in scale from 1 m to 250 m in width and spacing for several species of fouling organism s. The results of this study were consistent with and expanded to new species the theory presented by Callow et al., 2002 indicating some fouling organisms have a critical dimension at which fouling is decreased. Attachment point theory has not been developed to the point of be ing predictive. The theory should have the power to determine the rank of surfaces with limited att achment points, but this approach has not been verified. Barnacles: The fouling response of the larval stage of barnacles of several species has been evaluated on several types of topographies ranging from castings from meshes to single element designs to a larger version of th e Sharklet AF. In a field study in Sweden, Balanus improvisus barnacle cyprids settled in fewer numbers on pyramids and sawtooth channels at least 46 m tall and twice as wide at the base.50 The sawtooth channels were more effective at inhibiting the barnacle cyprids than the pyramids of similar dimensions. Topographical surfaces cast from a variety of mesh structures provi ded topographical features ranging in size. The 32

PAGE 33

resulting protrusions ranged in height from a pproximately 20 m to 100 m and from 20 m to 500 m in lateral dimensions. These t opographical surfaces inhibited fouling by Balanus improvisus by approximately 80% relative to smooth, and topographies with approximately 20 m in height showed the best performance.51, 52 Balanus amphitrite barnacle cyprids have been shown to have reduced settlement on both channels and the Sharklet AF when the f eature size and spacing is 20 m, an order of magnitude larger than for Ulva spores. The Sharklet AF surface inhibited settlement of the cyprids only slightly bett er than the channels. This small diffe rence is a contrast to the response of Ulva where the Sharklet AF surface greatly outperformed the channels. As with the Ulva spores, height was found to be a major factor in influencing settlement of Balanus amphitrite cyprids. Interestingly, both the Ulva spores and Balanus amphitrite cyprids responded with nearly identical slopes to the aspect ratio of the features.9 The searching behavior of barnacle cyprids is influenced by topogr aphies, indicating a likely reason for the reduction in settlement. Balanus improvisus cyprids spent approximately 90% of their time exploring a smooth surface in the laboratory. In contrast, cyprids spent most of the time swimming above the surface after a sh ort probing period when e xploring topographical surfaces at least 20 m in dimension. These behavioral patterns indicate the cyprids cannot find suitable settlement locations on the topographical surfaces and swim away to begin searching for another location.52 Multi-scale surfaces Fouling organisms respond to different topogra phy length scales, as indicated by the work described above. Surfaces with multiple length scales of regular topography have been developed in an attempt to reduce fouling by se veral organisms. One such surface combined a barnacle-inhibiting topography and an Ulva -inhibiting topography, shown in Figure 2-2. In this 33

PAGE 34

surface, the Sharklet AF with features 2 m wide and spaced by 2 m were imposed on channels 20 m wide and spaced by 20 m. Unfortunately, Ulva spores settled in higher numbers on these hierarchical surfaces relative to smooth.9 Figure 2-2. Hierarchical structure with spore-specific Sharkl et AF and 20 m wide ridges Another hierarchical surface cont ained less well-defined wrinkles with five generations of hierarchical wrinkling. This surface was create d using a process in which a strained PDMSe substrate was exposed to UV and ozone to creat e a rigid skin, and the imposed strain was subsequently released.53 The Ulva spores settled in the wrinkles and cracks of the wrinkled surfaces. Field tests showed that the hierarch ical wrinkled surfaces inhibited fouling by barnacles, and some of the other fouling was more easily removed than on the smooth surface. However, diatoms adhered strongly and could not be removed from the recessed regions. The experimental work on these foulers indicat es that further examination is required to understand the response of these organisms to engi neered topographies. Each of the organisms described above responds to topography, but these f oulers respond to at leas t two different length scales. 34

PAGE 35

Models for Fouling Several models have been developed for fouling, but none have been shown to be predictive. The critical dimension th eory appears to be applicable to Ulva diatoms, and barnacle cyprids, though this theory only provides a guid e to the length scale. The critical dimension theory does not account for differences in settle ment among surfaces that prevent the organism from penetrate areas between features. Ulva has been shown to respond differently to topographies of similar dimension. Limited evidence with diatoms suggests there may be differences in attachment for different geomet ries. The theory that force gradients direct Ulva settlement did not fully explain differences among Ulva settlement, but did show promise if it were to be incorporated into a more de veloped model. The Engineered Roughness Index correlated with Ulva settlement, but the ERI theory has not been tested with other data sets. The ERI theory has the potential to be predictive for Ulva but the predictive pow er has not yet been tested. Attachment point theory for diatoms has not progressed to the poin t of being predictive, and is very similar to the critical dimension theory. Additional work would provide insight into ne w models for fouling or to further develop current models. Predictive models would be usef ul in the design of new surfaces. Combining models and theories for multiple organisms would pot entially lead to surfaces that resist fouling by multiple organisms. Surface-Wide vs. Local Characteristics The work being performed on many of these ex perimental surfaces can be split into the scale of their focus. Some work is focused on the characteristics of the surface as a whole, while others are focused on the characteristics on scales which match the fouling organisms. Both sets of characteristics have been s hown to affect cell response. Ulva spores and marine bacteria C. marina respond to changes in uniform surface chemistry, but Ulva spores are also affected by 35

PAGE 36

patterns of surface chemistry on a scale similar to the cell body. Topography changes the wettability of a surface, but the local properties of the t opography have been found to be important. Organisms which search for a suitable position on the surface include Ulva spores, diatoms, and barnacle cyprids. This searching behavior indicate s that they are sensing local properties, though they may also sense longer-range properties. Local Position Preference Several foulers have been shown to prefer certain localized regions of topographical and chemical surfaces. However, studies of this pr eference have been limited to either simple geometries (i.e. channels/ridges or pi llars) or to observational evidence. Ulva linza spores have been shown in a number of st udies to settle in depressed regions of topographies, especially when the spore can touch both the side of a feature and the floor region.7, 54, 55 Both Ulva spores and diatoms settled in depr essions on hierarchically wrinkled surfaces.53 Several species were found to foul pref erentially in pits between hemispherical protrusions on the millimeter scale.56 Various bacteria species were found to attach in channels and to align within channels.57, 58 Pseudomonas aeruginosa and Pseudomonas fluorescens under flow attached to downstream edges of ridges an order of magnitude larger than the cells. The preferential settlement of Ulva spores to localized locations exte nds to chemical patterns, at least for one dimensional stripes (analogous to ridg es/channels). The spores prefer fluorinated chemical stripes instead of poly(ethyle ne glycol)-modified chemical stripes.59 Preferential locations for Ulva settlement have been observed on the Sharklet AF and two-element topographical designs. Spores were obse rved to settle centere d between at least two features of the Sharklet AF, specifi cally between adjacent diamond patterns.46 The spores could not penetrate the pattern, and instead br idged the features. This preference was again observed on surfaces with two rectangular elemen ts spaced at the 2 m critical dimension for 36

PAGE 37

Ulva .48 One surface with 2 m pillars and 10 m rectangles, the Ulva spores avoided the triangles, despite being able to entirely fit on the triangles.47 An example of spores settling in locations between features is shown in Figure 2-2. Figure 2-2. Spores settled between adj acent diamonds on the Sharklet AF surface The preferential settlement locations were not quantitatively ev aluated and relied on observational evidence in all these studies of surf aces more complex than a single element. There has been limited work performed using one-dimen sional analysis for cell preference on channel topographies.7 No studies on two-dimensional quantitativ e analysis of cell preference have been identified. Quantitative analysis would allow for the identification of prefer ential sites, some of which may not be apparent with observational analysis. Studies indicating organism preference within a surface, such as within depressed regions, channels, etc. are further evid ence for local properties infl uencing cell response. Because organisms may respond both to long range and local properties, both approaches may provide insight into cell behavior. Summary Biofouling is a problem that affects many as pects of our lives, and creates significant economic and ecological concerns in the marine environment. New technologies are needed to 37

PAGE 38

effectively combat marine fouling without re leasing toxic substances. Topography is one approach to combat fouling, and th is approach can be combined w ith other technologies such as surface chemistry and bulk materials. Surface-wide characteristics and models and localized models would aid in developing new mode ls and new surfaces to combat fouling. 38

PAGE 39

CHAPTER 3 1 POTENTIAL FOR TUNABLE STATIC AND DYNAMIC CONTACT ANGLE ANISOTROPY ON GRADIENT MICROSCALE PATTERNED TOPOGRAPHIES Introduction Topographies, including those used for antif ouling, change the shap e and movement of water on the surface. The wettability of the surf ace has been linked to biofouling by bacteria and algae,38 and linked to cell attachment.60 Surface wettability has been sh own to affect the adhesion strength of fouling Ulva spores and the shape of the adhesive pad the spores use to attach to the surface.32, 61 Changes in the wettability due to top ography may influence the cell behavior and should be carefully investigated. The shape and movement of the fluid are ch aracterized by the static contact angle and dynamic contact angles. There is some limited evid ence that anisotropy existed in the static and dynamic contact angles on continuous channel or ridge topographies of various sizes.62-66 Experimental work on anisotropic wetting on topographies of disconti nuous features is lacking in the literature. The most common measures of wettability ar e the static contact angle (also called the apparent contact angle) ( ), the advancing contact angle ( A), the receding contact angle ( R), the slip angle ( ), and the contact angle hysteresis (AR). The advancing and receding contact angles are the contact angles as a liquid wets or dewets a surface, respec tively. The contact angle hysteresis is the difference between the advancing and receding contact angles. The slip angle is the angle required to cause a li quid drop to slide off of a surf ace, causing the drop to wet the surface at the advancing front, and de wet the surface at the receding front.67 These values are shown in Figure 3-1. 1 Reproduced with permission from Langmuir, in press. Unpublished work copyright 2009 American Chemical Society 39

PAGE 40

Surface topography has been shown to alter the apparent contact angle by either increasing the true contact area of the solid-liquid interface,68 or by entrapping air and forcing the apparent solid-liquid interface to beco me a composite of solid-liquid and solid-vapor interface.69 Wenzel described the change in contact angle as an in crease in the true solid-liquid interface using Equation 3-1. Figure 3-1. Schematic of static and dynamic contact angles of sessile liquid drops cos cos r (3-1) In Equation 3-1, is the apparent contact angle, is the contact angle on a smooth surface, and r is the ratio of the total surf ace area to the projected surface area.68 The influence of trapped air was described by Cassie and Baxter by Equation 3-2. 1cos cos ff (3-2) In Equation 3-2, f is the area fraction of the projected surface area which is filled by topographical feature tops. The value f represents the area fraction of solid-liquid interface below the liquid drop when the drop is supported by both solid area and trapped air. The Cassie-Baxter 40

PAGE 41

equation has been used to predict the static, a dvancing, and receding contact angles for a liquid in the Cassie-Baxter regime if these contact angles fo r the smooth surface are known.69 The wetting regimeWenzel, Cassie-Baxter, or a wicking regimecan be predicted using a lower and upper critical contact angle for an equivalent smooth surf ace. Non-smooth surfaces are expected to exhibit Cassie-Baxter wetti ng when the equilibrium contact angle on an equivalent smooth surface, exceeds an upper critical value, given by Equation 3-3. fr fUC1 cos1 (3-3) In Equation 3-3, f is the area fraction of the projected surface area which is filled by topographical feature tops and r is the Wenzel roughness value. A wicking regime, characterized by fluid spreading between the featur es, is predicted to occur when is below the lower critical value given by Equation 3-4 with the same va riable nomenclature as in Equation 3-3: fr fLC1 cos1 (3-4) Wenzel wetting is expected to occur when lies between the LC and UC.70 Surfaces with translationally symmetric topographies such as regularly spaced channels,46, 50, 65, 71, 72 arrayed pillars,46, 65, 67, 71-78 arrayed pits,72, 75, 79 and patterns of combinations of various geometric shapes46-48, 50have shown increases in contact angles due to the roughness of the surface. The increase in contact angle due to the translationally symmetric topographies has been largely attributed to Wenzel and CassieBaxter wetting as described by Equations 3-1 and 3-2. Recently, the issue of whether the contact a ngle of the drop is defined by the properties of the entirety of the surface under the drop body or only the surf ace immediately under the three-phase contact line has been discussed.80-83 41

PAGE 42

For the slip angle, a dynamic contact angle m easurement, Miwa et al. derived an equation relating the slip angle to the apparent contac t angle for rough surfaces that exhibited mixed regimes of Wenzel and Cassie-Baxter behavior: 3/1 3 2 2coscos32 3 1cos 1cos'sin2 sin m rg rk (3-5) In Equation 3-5, r is the r oughness ratio of the true solid -liquid interface, k is the interaction parameter for the liquid and smooth surface, is the contact angle of the smooth surface, is the apparent contact angle of the rough surface, g is the gravita tional constant, m is the mass of the drop, and is the density of the drop. In Equation 3-5, the term ( cos + 1)/(rcos +1) is equal to the value of f in Equation 3-2. The inte raction parameter k used in Equation 3-5 can be calculated using a smoot h surface of the same material and the following relation: sin6 sin cos3cos3293/1 3/1 2 3 2g m k (3-6) Equations 3-5 and 3-6 can be used to predict the slip angle for drops in the Wenzel and Cassie-Baxter regimes by altering the parameters as described in Miwa et al. for mixed regimes.84 Theoretical calculations by Wola nsky and Marmur predict that anisotropy of the static contact angle on translationally symmetric t opographical surfaces will not occur when the features are sufficiently small compar ed to the size of the water droplet.85 One numerical examination predicts that when the features are large enough su ch that a liquid drop resides on only a small number of ridges, contact angle an isotropy of the static contact angle occurs.62 Several studies have reported angular depende nce of the static contact angles on hydrophobic ridges at least 20 m in width. On these surf aces, the contact angle has been reported to be 42

PAGE 43

between 15 and 23 higher when the drop is vi ewed perpendicular to the ridges than when viewed parallel to the ridges.62, 63, 71 Hydrophobic ridges 5 m in widt h, separated by between 5 and 20 m have been reported observationally to alter the shape of 2 l drops, but the angular dependence was not quantified.46 One report was found of contact angles on surfaces containing ridges of center to center dist ance of 396 to 513 nm, two orders of magnitude smaller than the reports cited above.64 In the report, 3 l drops exhibited anisotropy in th e static contact angles by between 2.1 and 13.7, increasing both with de pth and with increasing wavelength. Even these small wavelength ridges were not small enough comp ared to the 3 l drops to satisfy the size limit described by Wolansky and Marmur required to eliminate static contact angle anisotropy. The literature addressing the anisotropy in the dynamic contact anglesthe wetting and dewetting behavioron translationally symmetric topographical surfaces is even more limited than that of static contact a ngle studies. One study reported anis otropy in the contact angles of water on ridges 1 m wide and tall.65 The advancing and receding a ngles were higher by 18 and 7, respectively, when the drop wa s moving perpendicular to the ri dge features compared to the drop moving parallel to the ri dge features. Another study using ridges with wavelengths of 396 and 513 nm reported anisotropy in the contac t angle hysteresis on these small features.64 Aluminum ridges 30 m in width exhibited approximately 30 higher advanc ing contact angle in the direction perpendicular to the ridges than parallel to the ridges.66 Evaluation of the anisotropy in slip angles on translationally symmetric topographies is also extremely limited in the literature. One reference showed a 20 anisotropy in the slip angle of water on ridges 50 to 150 m wide when the drop is movi ng perpendicular to the ridges.71 Several natural surfaces have been observed to exhibit dynamic contact angle anisotropy. The rice leaf Oryza sativa, which has microscopic papilla arranged with or der in one dimension 43

PAGE 44

and randomly oriented in the othe r, exhibits anisotropy of up to 15 in the slip angle for an unreported drop size.86 Water striders exploit sliding angle anisotropy of groove-like structures on the hair of their legs.87 The anisotropy allows the striders legs to glide on top of the water or to generate large propulsion forces depending on the orientation of the structures with respect to motion against the water. Microstructures and flexible nanostructures on the wings of the Morpho aega butterfly control the sl iding of water droplets.88 The water droplet s slip radially outward from the butterflys body in a Cassie-Baxter state, but are pinned by a Wenzel-like state if the drop motion is toward the body. Translationally symmetric chemical pattern s compose another form of surfaces with engineered heterogeneity. Several studies have reported contact angle anisotropy on microscale chemically patterned surfaces. Chemically hete rogeneous stripes approximately 1-20 m in width have been shown to exhibit anisotropic static, advancing, and receding contact angles and slip angles. The values for all of these characteristics were higher in the direction perpendicular to the stripes than in the direction parallel to the stripes.89 Advancing and receding contact angles on chemically heterogeneous stripes 2.5 m th ick were 2-10 higher when drops traveled perpendicular to the stripes than when drops traveled parallel to the stripes.90 Chemical gradients on scales much larger than the drop size have ex hibited interesting behavior, such as a liquid drop moving up an incline when the surface en ergy change is large and the contact angle hysteresis is small.91 The authors in the above cited literatu re examined only surfaces based on channels/ridges/stripes, which ar e continuous along the length of the features, or complex natural surfaces. Anisotropy in the static and dynamic contact angles on ridge or channel surfaces have been attributed to discontinuity in the three-phase contact line and the resulting energy barriers.64, 44

PAGE 45

71 When the contact line is prevented from moving the contact line is said to be pinned. The continuous nature of the features in one direction imposes no ba rrier to motion in the direction along the features, while creating en ergy barriers in the direction pe rpendicular to the features. In contrast, discontinuous small-s cale topographical features, such as microscale rectangular features, may cause the surface to exhibit little, or even no, anis otropy in the static and dynamic contact angles due to energy barrie rs to motion in all directions. We have developed a series of micro-t opography gradients inte nded as anti-fouling surfaces for use in the marine environment.48 Of special note is the gradient surface denoted Sharklet AF, which has shown consiste nt antifouling behavior over numerous in vitro studies.9, 46-48 The features of these surfaces are at th e smaller end of the range of dimensions examined for contact angle anisotropy in prev ious reports. These gradient surfaces exhibit anisotropies in feature lengths and arrangements. The micrometer-s cale features are rectangular, parallel, and discontinuous along the length of the surface. The features are approximately 2 m wide, 3 m tall, and spaced by approximately 2 m (Figure 3-2). The Sharklet AF surface contains features 4, 8 12, and 16 m long; GR 0 contains features 4 m long; GR1 contains features 4 and 8 m long; GR2 contains features 4 and 12 m long; GR3 contains features 4 and 16 m long; GR4 contains features 8 and 12 m long and GR5 contains features 12 m long. The well defined, regular geometries allow fo r prediction of static and dynamic contact angles and wetting regimes. The Ridges topography was expected to exhibit significant anisotropy on both the static and dynamic contact angles. In contrast, the gradient surfaces (GR0 through GR5 and Sharklet AF) were expected to exhibit greatly reduced static and dynamic contact angle anisotropies, with the possibility that no cont act angle anisotropy would be 45

PAGE 46

observed on these surfaces. This report includes the directional dependence of the static and dynamic contact angles for this series of microtopographies. Materials and Methods Silicon wafer mold fabrication, poly(dimethyl siloxane) elastomer (PDMSe) material preparation, casting method, and slide preparati on have been described in detail previously.48 A brief overview is provided here. Figure 3-2. Scanning electron micrographs of the gradient and ridge surfaces. Topographical Surfaces The topographical surfaces used in this study have been described previously.48 The Sharklet AF four-element gradient and two-element Gradient surfaces 0 through 5 (Figure 3-2) were designed to have heights of 3 m tall a nd widths and spacing of 2 m. The feature lengths vary from 4 m to 16 m. The term gradient surfaces refers to the Sharklet AF and Gradient surfaces. A ridge topography, consis ting of ridges approximately 3 m tall, 2 m wide, and spaced by 2 m was used for comparison. 46

PAGE 47

Silicon Wafer Molds Silicon wafers 4 inches in diameter and of the (100) orientation were patterned using photolithographic techniques and etched using deep reactive ion etching as described previously.48 The etched wafers were used as negative molds for the topographically modified surfaces. Using these methods, photoresist coat ed silicon wafers were exposed through a photomask containing the pattern designs, and the photoresist was developed. Deep reactive ion etching was used to etch the ne gative pattern to a depth of 3 m into the silicon wafers. Each wafer was cleaned of photoresist with an O2 plasma etch, followed by vapor deposition of hexamethyldisilazane to prevent bonding of the s ilicon wafers to the SILASTIC T-2 PDMSe in the replication process. This process produced silicon wafer molds with topography inverted so that the cast product will cont ain the desired topography. Material The platinum catalyzed SILASTIC T-2 PDMS e (Dow Corning ) system was prepared by mixing ten parts resin and one part curing agent by weight for 5 minutes, followed by degassing under vacuum for 30 minutes. The dega ssed mixture was cured against silicon molds for 24 hours at approximately 22 C. Following casting, the surf aces were rinsed with 95% ethanol in water, blown with compressed nitrogen gas, and stored for 24 hours prior to testing. Contact Angle Characterization Contact angle characterization was performed using a Ram-Hart contact angle goniometer with an automated drop dispenser, image capture system, and a tilting stage. Contact angles were measured on captured images using the angle function in ImageJ.92 Nanopure water with resistivity greater than 17 M -cm was used for all contact angle measurements. Static water contact angle images were captured after placing a 5 l drop on the surface. Slip angle images were captured by tilting the stage at approximately 1/s until the drop began to slip. The angle at 47

PAGE 48

which the stage was tilted was recorded as the slip angle. Images captured at the slip angle were analyzed to measure the advancing water contact angles and receding water contact angles at the leading and following edges, resp ectively, of the slipping drop. To evaluate the contact angles and slip a ngle as a function of su rface orientation, the orientation of the topographical surface was varied with resp ect to viewing angle and drop motion systematically. The orientation was varied such that the drop slid in a direction parallel to, perpendicular to, or at a 45 angle to the elong ated direction of the feat ures. The directions are illustrated in Figure 3-3. Three drops were used on each surface at each rotation, for a total of nine drops per sample. Nine drops were used on the smooth surface, as no directionality was present in the surface. Advancing and receding water contact angles and slip angles were measured using tilted 20 l drops for the smooth and ridge s surfaces only, because the 5 l drops did not slip off of the smooth surface and did not slip off of one rotatio n of the ridges surface. The static water contact angles on the smoot h and ridges surfaces were measured using 5 l drops for comparison to the gradient surfaces. The contact angle hysteresis was calculated by subtracting the receding water contact angle fr om the advancing water contact angle for each captured image at the slip angle. Prediction of Wettability Characte ristics on Topographical Surfaces Scanning electron micrographs of the cross-se ction of each surface were analyzed using ImageJ for dimensions of feature height, wi dth, and spacing. The roughness ratios and area fractions of feature tops for each surface were calc ulated from the feature dimensions determined from SEM analysis. The wetting regime for each surface was predicted by comparing the static contact angle for the smooth PDMS e surface to the critical valu es described by Equations 3-3 and 3-4 for each topographical surface. The static, advancing, and receding water contact angles 48

PAGE 49

were predicted for each topographically modified surface using the appropriate equation for the wetting regime. Figure 3-3. Illustration of the convention used to denote directionality. Arrows represent direction of drop motion: A) perpendicular to features, B) 45 offset from features, and C) parallel to features. All surfaces were predicted to exhibit CassieBaxter wetting, and contact angles were predicted using Equation 3-2. The predicted co ntact angle hysteresis for each surface was calculated by subtracting the predicted reced ing water contact angle from the predicted advancing water contact angle. The interaction parameter for PDMSe-water was calculated using Equation 3-6 and the slip angle and static contact angle values measured for 20 l drops on smooth PDMSe. The predicted slip angle for each surface was calcula ted with Equation 3-5 using the measured static water contact angles for each surface and th e interaction parameter calculated from the smooth PDMSe surface. Results and Discussion Predicted vs. Experimental Measurements The static water contact angles and the s lip angles for 5 l drops and the dynamic wettability measurements for 20 l drops measur ed on smooth PDMSe are shown in Table 3-1. 49

PAGE 50

All measurements are reported as the average of the measurements and the 95% confidence interval. The static and dynamic contact angles on smooth PDMSe are in excellent agreement with measurements previously reported on the sa me material formulation using the Wilhelmy Plate method.54 The slip angle was not measured in the previous report, so no comparison can be made for this measurement. Analysis of the scanning electr on micrographs reveal that the heights of the surfaces varied from 2.3 m to 2.6 m, widths varied from 1.8 m to 2.4 m, and the spacing varied from 1.6 m to 2.2 m among the sample surfaces. The roughne ss ratios, area fractions of the tops of the features, and predicted wetting regimes for the surfaces tested are shown in Table 3-2, as determined from the measured dimensions. Table 3-1. Contact angle data for smooth PDMSe Drop size Wettability characteristic Measurement Static contact angle 113 1 5 l Slip angle >90 Advancing contact angle 118 1 Receding contact angle 72 7 Contact angle hysteresis 46 7 20 l Slip angle 81 19 Table 3-2. Calculated surface parameters for topographies Surface r f LC UC Wetting regime Ridges 2.3 0.60 71 104 Cassie-Baxter GR0 2.2 0.35 77 110 Cassie-Baxter GR1 2.3 0.43 75 108 Cassie-Baxter GR2 2.3 0.42 76 109 Cassie-Baxter GR3 2.2 0.38 76 110 Cassie-Baxter GR4 2.2 0.44 75 108 Cassie-Baxter GR5 2.2 0.45 74 109 Cassie-Baxter Sharklet AF 2.3 0.48 74 107 Cassie-Baxter Roughness ratio (r), area frac tion of feature tops ( f ), lower critical contact angle ( LC), and upper critical contact angle ( UC). All of the topographical surfaces used in this study were predicted to exhibit Cassie-Baxter wetting, due to each surface having a UC below the experimental value for smooth PDMSe. The 50

PAGE 51

predicted advancing, receding, and static water contact angles and the corresponding experimentally measured contact angles are show n in Table 3-3, and ar e represented with 95% confidence intervals. Because the predictive models do not account fo r any directionality, measurements from all three surf ace orientations are included in the measured values in Table 33. The measured static contact angle on the Sharklet AF surface was consistent with the previously reported value of 135.46 Table 3-3. Predicted and measured static, a dvancing, and receding wa ter contact angles on topographies with 95% confidence intervals A R AR Surface Pred. Meas. Pred. Meas. Pred. Meas. Pred. Meas. Ridges 130 128 () 133 138 () 102 106 () 31 32 () GR0 142 138 () 144 163 () 123 128 () 22 35 () GR1 138 134 () 140 156 () 116 119 () 25 37 () GR2 138 136 () 141 154 () 117 118 () 24 36 () GR3 141 134 () 143 152 () 121 121 () 23 31 () GR4 137 135 () 140 158 () 115 116 () 25 42 () GR5 137 137 () 140 157 () 114 116 () 25 41 () Sharklet AF (+3SK2x2) 135 135 () 138 160 () 112 112 () 26 48 () For the surfaces tested, the Cassie-Baxter equation predicted the static contact angles to within 3 on average and the re ceding contact angles to within 2 on average. The advancing contact angles and contact angle hys teresis values were not predicted as accurately as the static and receding contact angles. The advancing cont act angle was an average of 15 higher than predicted. The contact angle hyste resis was also an average of 13 higher than predicted, largely due to the under-prediction of the advancing contact angle by a similar numb er of degrees. All of 51

PAGE 52

the measured contact angles over all of the surfaces were statistically significantly different than the predicted values ( = 0.05). The large standard deviatio ns on the measurements for the ridges surface are due to the large anisotro py in contact angles on ridges. The interaction parameter was calculated from the contact angle data for the 20 l drops on smooth PDMSe using Equation 3-6. From 4 distin ct 20 l drops, the inte raction parameter for water on PDMSe was calculated to be k = 17.8.9 mJ/m2. This value was used to predict the slip angles for 5 l drops on each of the surf aces using Equation 3-5. The predicted contact angles for each surface were used as the apparent contact angles, ', for use in Equation 5 in predicting the slip angles The roughness ratio for the tops of f eatures used in the calculation of slip angles was assumed to be 1, as this is the case for Cassie-Baxter wetting. The predicted values for the slip angles are shown in Table 34, along with the 95% conf idence intervals for the measured values. Water drops 5 l in volume were predicted not to slip on the ridges surface, even at 90 tilt. The experimental slip angles for the ridges surface are omitted from Table 3-4 due to the high direction-dependence of the slip angles on this surface, with some drops not slipping. The slip angles were well predicted from Equation 3-5. The predic ted slip angles were average of less than 1 different than the e xperimental values, and this difference was not statistically significant ( = 0.05). Table 3-4. Predicted and Measured Slip A ngles for Topographically Modified Surfaces Slip angle, Surface Predicted Experimental GR0 32 27 () GR1 46 43 () GR2 44 38 () GR3 36 39 () GR4 47 49 () GR5 50 51 () Sharklet AF (+3SK2x2) 57 66 () 52

PAGE 53

Anisotropy of Measurements The dependence of the measured wettability characteristics on the underlying topography are illustrated in the angle-dependence graphs fo r static, advancing, and receding contact angles, slip angles, and contact angle hys teresis, shown in Figure 3-4, A through E, respectively. Only the parallel and perpendicular di rections are shown for clarity. Figure 3-4. Angular dependence of A) static contact angle, B) advancing contact angle, C) receding contact angle, D) contact angle hysteresis, and E) slip angle. Error bars represent standard deviation. Asterisks note statistically significant anisotropy (Tukeys test, = 0.05). 53

PAGE 54

Figure 3-4. Continued. 54

PAGE 55

The ridges surface had the largest anisotropy of all surfaces for the static contact angle, advancing contact angle, contact angle hysteresis, and slip angle. The anisotropy in the receding contact angle was similar on all of the surfaces A comparison of the anisotropy on the ridges surface to the gradient surfaces, including the Shar klet AF surface, is shown in Table 3-5. The discontinuous gradient surfaces decreased the an isotropy of the static and dynamic contact angle measurements by between 76% and 90%, excluding the receding contact angle. The anisotropy in all of the static a nd dynamic contact angles were found to be statistically significant (p<0.01) by Tukeys test following two-factor analysis of variance. The static contact angle and all of the dynamic contact angles, excluding the receding contact angle, were on average highest in the direction perpendicular to the features. The receding contact angle was on average lowest in the direction perpendicular to the features. Of th e gradient surfaces, th e angular dependence was most pronounced on the GR3 surface, which had th e highest degree of di ssimilarity between adjacent features. The direction of the anisotropy on all of the static and dynamic contact angles is consistent with larger energy barriers to drop motion in th e perpendicular direction than in the parallel direction. When the advancing three phase contact line is pi nned by the energy barrier, the contact angle at the advancing front must increase to overcome the energy barrier. In contrast, when the receding three phase contact line is pi nned by the energy barrier, the contact angle at the receding front must decrease to overcome the energy barrier. Th e hysteresis increases as both the advancing contact angle increases and the receding angle decreases due to pinning. The anisotropy in these three dynamic wettability characteristics indicated that, in general, the topographical features produced a higher energy for drop motion when the drop was moving in 55

PAGE 56

the direction perpendicular to the features than when the dr op was moving parallel to the features. Table 3-5. Comparison of contact angle anisot ropy on gradient topographi cal surfaces and ridges Anisotropy: perpendicula r vs. parallel directions Ridges Gradients Decrease in anisotropy 16 2 85% A 33 3 90% R -5 -6 None AR 38 9 76% >42 8 >80% Positive values indicate larger measurements in the perpendicular direction, while negative values indicate larger measurements in the parallel direction. Comparison of Anisotropies on Gradient Surfaces to Ridges Comparison of the anisotropies on the gradient surfaces to the anisotropies on the ridges surface reveals several interesting points with respect to the en ergy barriers. The decreases in wettability anisotropy on the gradient surfaces compared to the ridges surface indicate the discontinuous features impose an energy barrier to motion in bot h directions. When breaking up the ridges to create the gradient surfaces, an en ergy barrier to motion in the parallel direction is created that is nearly, but not quite, as large as the energy barrier in the perpendicular direction. The smaller anisotropy in the static contact an gles on the gradient surfaces relative to the ridges surface indicates that the discontinuous gradient features re strict the drop spreading in the direction parallel to the featur es, whereas the ridges allow the drop to spread along the parallel direction. As shown in Figure 3-4 the advancing c ontact angle in the para llel direction on the ridges is much lower than on the gradient surf aces, while it is similar in the perpendicular direction on the ridges and gradie nt surfaces. This indicates that the ridges and gradient surfaces have similar energy barriers for the advancing c ontact line in the perpendi cular direction, while the discontinuous features of the gradient surfaces create an en ergy barrier in the parallel direction. 56

PAGE 57

In contrast, the receding contac t angles show similar anisotro py on the gradient surfaces as on the ridges surface. This indicates that the energy barrier to the receding contact line motion is not significantly increased by breaking up the ridges into disc ontinuous gradient features. To the contrary, the receding contact line barrier may be slightly decreased by the discontinuous features, as the receding contact angles are lower for the ridges in both directions than on the gradient surfaces. The slip angles and slip angle anisotropies point to another intere sting consequence of breaking the ridges into disconti nuous features. The slip angles on the gradient surfaces are generally lower than the slip angles on the ridges surface. This is lik ely due to a lower area fraction of elevated solid surface supporting the drop ( f in Equation 3-2). The lower area of solidliquid contact increases the stat ic contact angle, reducing the area of contact between the drop and the surface for a given drop volume. The smalle r area of contact reduces the area-dependent force resisting drop motion in a manner analogous to a frictional force. The decrease in the receding contact line barrier, as described above, may also play a role. The large slip angle anisotropy in the ridges indicates a large difference between direc tions on the barriers to drop motion. The greatly reduced anisotropy in the slip angles on the gradient surfaces indicates the barriers to motion are nearly, but not quite, the same in both directions. Regression of Slip Angle Multiple regression analysis showed that the sl ip angle is best correlated to the receding contact angle on the gradient topographical surf aces. The ridges surface was omitted, as not all of the 5 l drops slipped. The regression (p<0.001, R2 = 0.75) using each drop as an independent data point gave Equation 3-7. dropR drop ,cos36.132.1sin (3-7) 57

PAGE 58

The regression (p<0.001, R2 = 0.93) using each gradient topographical surface as an independent data point gave Equation 3-8. surfaceR surface ,cos77.153.1 sin (3-8) The slip angles were highly correlated to the receding contact angles on the gradient surfaces tested. Equations 3-7 and 3-8 are not gene ral equations to predict slip angles, but are correlations for the drop size and surfaces test ed. The slip angles would be undefined in Equations 3-7 and 3-8 if the receding angle we re below about 107. An undefined slip angle indicates the drop would not slip off the surface ev en at 90 tilt due to insufficient gravitational force to overcome the forces on the drop. The slip angle showed no si gnificant dependence on the advancing contact angle, static contact angle, or the hysteresis. The advancing contact angles did not show any significant correla tion to the receding contact angle. These findings have several significant impli cations. First, pinning of the receding liquid front was the limiting factor for determining when the liquid drop slipped on the surfaces tested. This result indicates neither pinning of the advancing front nor contact angle hysteresis determined slip angle. Second, the advancing co ntact angles were not determined by the slip angles or receding contact a ngles. Drops that were pinned by the receding front did not necessarily show artificially hi gher advancing contact angles. The regression between the slip angle and the r eceding contact angle is interesting in light of the fact that both the receding contact angles and slip angles were well predicted by the models. Prediction by theoretical models and corr elation between to valu es are not necessarily connected. The model may fail to accurately predict the contact angle even when this contact angle plays an important role in determining the slip angle. The slip angle is typically described as being a function of both the advancing a nd receding contact angl e (i.e. contact angle 58

PAGE 59

hysteresis).93 The results indicate the advancing contact angle and hys teresis did not contribute sufficiently to the slip angle on these surfaces The connection between receding contact angle and slip angle is similar to conclusions by M ougin et al. that nanoscal e friction properties are related to the receding contact angle of water on chemically uniform and chemically heterogeneous substrates.94 Both the slip angle and friction properties are related to forces between the surface and on the drop or AFM tip. However, th e topographical surfaces and chemical surfaces would not be expected to aff ect the wetting exactly the same way, and the interaction area of the water droplet and AFM tip are of different magnitudes in scale. The slip angle did not appear to be a function of the static contact angle. The static contact angles on the topographical surfac es tested did not have as hi gh of a range as the other wettability characteristics. The contributions to the slip angle by the st atic contact angle are expected to be very small in th is study due to the small range in static contact angles. Studies using surfaces with a larger rang e for static contact angles would likely see an effect on the slip angle by the static contact angle due to changes in drop shape, as predicted by Equation 3-5. A smaller static contact angle w ould produce a larger area in cont act between the liquid and the surface, creating a larger energy barrier for motion. Conclusions The data presented indicate that breaki ng up continuous ridges into discontinuous rectangular features greatly reduces, but does not eliminate, static a nd dynamic contact angle anisotropy at feature sizes of approximately 2 m in width and spacing. The ridges surface exhibited considerable anisotropy in the static contact angle, a dvancing contact angle, contact angle hysteresis, and slip angle. The receding co ntact angle was statistically anisotropic on the ridges surface, but the anisotropy was much sma ller than for the other characteristics. The gradient surfaces, including the Sharklet AF, exhibited greatly reduc ed, but statistically 59

PAGE 60

significant, anisotropy in the st atic contact angle, advancing contact angle, contact angle hysteresis, and slip angle. The receding angle anisotropy was statistically significant on the gradient surfaces and similar in value to the anisotropy on the ridges surface. Compared to the continuous ridges, the discont inuous features of the gradient surfaces increased the barrier to advanci ng contact line motion in the direc tion parallel to the features and reduced the ability of a static wa ter droplet to spread along the features in th e parallel direction. The gradient features had a reduced solid area fraction supporting the liquid drop relative to the ridges as a result of breaking the ridges into features. This decreased solid-liquid interfacial area caused the gradient surfaces to have lower slip an gles than the ridge surface, despite an increase in the energy barrier to advancing contact line motion. The gradient surfaces slip angle anisotropy was greatly reduced compared to the ridges surface, indicating a smaller difference in the energy barrier to overall drop motion on the grad ient surfaces compared to ridges. The slip angle was found to be highly correlated to the receding contact angle on the gradient surfaces, indicating that the slip angle was dictated by the receding contact angle for the gradient surfaces using 5 l drops. Breaking up the ridges decreased both the slip angle and the anisotropy in the wettability measurements. If a surface is desired to have a low slip angle and high anisotropy, the feature dimensionsincluding height, wi dth, spacing, and lengthscoul d be optimized to produce a desirable compromise. This could be useful in designing specializ ed microfluidic devices, in which a low resistance to fluid motion and high anisotropy may be desired. On average, the liquid drops on gradient surface s exhibited static cont act angles 2 higher when the liquid front was moving perpendicular to the features. This anisotropy is of similar magnitude as many of the reported 95% confidence intervals on the static contact angles when 60

PAGE 61

the angular information is not taken into account, as in Table 3-3. Similarly, the anisotropy in the other wettability characteristic s is larger than many of the corresponding confidence interval ranges when the angular information is not taken into account. If not taken into account, anisotropy in the contact angles will increase th e apparent variability in contact angles of randomly oriented observations. This study and previous studies cited were performed on hydrophobic surfaces, but hydrophilic surfaces may show different anisotropic responses. This could be performed using similar surfaces using different materials, by pr oducing chemical grafts on the surface, or a combination of both. A study using the same mate rial and the same set of topography designs in varying size scales could be perfor med to further evaluate the effect of feature size on static and dynamic contact angle anisotropy with both cont inuous and discontinuous features. The surfaces tested were designed for marine antifouling ap plications and not designed to produce certain static or dynamic contact angles. Further examination of anisotropic topographical surfaces could lead to surfaces designed for directional self-c leaning surfaces, microfluidic devices, or other surfaces designed for directing fluid motion. 61

PAGE 62

CHAPTER 4 DEVELOPMENT AND IMPLEMENTATION OF A PREDICTIVE MODEL FOR BIOSETTLEMENT OF THE GREEN ALGA ULVA LINZA Introduction Understanding the processes a ssociated with settlement, leading to the permanent attachment of the colonizing stag es of fouling organisms, is fundamental to the development of novel strategies to control fou ling without the use of biocides Those organisms with motile larvae or spores such as barnacles95 and the green alga Ulva96, actively select favorable sites on which to settle. Attributes of the surface such as surface energy or topography also affect the ability of the settled larvae, spores or cells to adhere to the surface.45, 49, 97 This paper focuses on the settlement of motile spores (zoospores) of Ulva. Colonization and growth of Ulva on underwater surfaces is in itiated by zoospores, which are able to perceive and respond to various su rface-associated cues including wettability/surface chemistry,38, 98, 99 chemical patterns,59 charge100 and topography. Periodic topographical surfaces whose features were spaced by 2 m reduced the settlement of zoospores while surfaces with spacing 5 m or larger increased settlement.7, 46-48, 54 The working hypothesis in these studies was that the 2 m spacing would inhibit settlement due to the inab ility of the spore to penetrate the volume between the features. One surface, th e Sharklet AF, which consists of four different lengths of rectangular features, is of particular interest. In terms of settlement of spores, the Sharklet AF consistently outperformed all ot her surfaces tested in four separate studies, including those surfaces containing combinations of one or two of the four rectangular features in the Sharklet AF pattern.46-48 Therefore, the arrangement of th e four rectangular features in the Sharklet AF appears to be cruc ial to the performan ce of the surface. The Engineered Roughness Index (ERII) was derived from the results of one of the Ulva settlement studies and involves several geom etric aspects of the surfaces (Equation 4-1).47 62

PAGE 63

s Idfr ERI 1 (4-1) In Equation 4-1, r is the roughness ratio as described by Wenzel,68 df is the degrees of freedom (1 or 2) for movement across the surface, and S is the area fraction of feature tops as described by Bico et al.65 The value S is equivalent to the value f1 used by Cassie and Baxter for solid-liquid interface69 when describing composite (or Cass ie-Baxter) wetting. For consistency with more well-known variables, (1-S) replaces the value fD, the area fraction of the depressed regions of the surface used in Schumacher et al. 2007. It should be noted that the variables S and r are used in the ERI as geometric descriptor s of the surface and do not imply a particular wetting regime. Settlement of spores of Ulva correlated to ERII with an R2 of 0.69 and p < 0.001 when using each replicate image as a da ta point (30 images per surface type). The settlement response on the engineered topographical surfaces used in Schumacher et al. (2007) correlated with the ERI as described in Equation 4-2. ERI 5.63796 spores/mm Density Spore2 (4-2) In the present paper we have expanded the original ERI correlat ions by incorporating additional spore settlement data and we have determined the predictive power of the ERI relationship using previously untested surface desi gns. A predictive model for settlement will aid in the design of new surfaces to prevent biofouling. Materials and Methods Refinement of Engineered Roughness Index Spore settlement data on engineered topograp hies were obtained from two previously published reports.47, 48 The feature dimensions of the surf aces in Schumacher et al. 2008 were previously evaluated and reporte d to two significant features.101 For regression analysis, the 63

PAGE 64

transformation was determined by selecting the transformation which caused the variance to be constant along the range of settlement densi ties. A second Engineered Roughness Index, the ERIII, was developed and is de scribed in Equation 4-3. s IInr ERI 1 (4-3) In Equation 4-3, r is the Wenzel roughness ra tio as in Equation 4-2, n is the number of distinct features in the surface design, and S is the area fraction of th e feature tops. Comparing ERIII to ERII, the degrees of freedom (df) was replaced by the number of distinct features in the design (n). Surface Designs Nomenclature Surfaces are described by a nomenclature (E quation 4-4) that includes the surface design, the feature width, feature spacing, and the number of dist inct features in the design. In Equation 4-4, a + preceding the height indicates the features protrude from the surface, while a indicates the features are recessed (i.e. pits). The height or depth of these features is indicated by the [height] term, the feature widt h is denoted by [width], and the spacing between the features is denoted by [spacing]. The surface design for all surf aces used here is the Sharklet AF, denoted SK. The number of distinct features in the de sign is denoted by [distinct features]. When the [distinct features] is omitted, the number of distinct features is assumed to be n = 4, the same as the original Sharklet AF design. The smooth surface is denoted SM without the other parameters. All dimensions are given as m. Fo r example, a 2 element Sharklet AF surface is denoted +2.7SK2x2_n2, representing 2.7 m feature height, 2 m feature spacing, 2 m feature width, and 2 distinct f eatures in the design. Surface Designation = [height][surface design][width]x[spacing]_[distinct features] (4-4) 64

PAGE 65

Evaluating S The Sharklet AF surface was inverted such that the rectangular features form pits instead of protrusions. This new design, the Recessed Sharklet AF, increased S without altering the degrees of freedom for movement, the number of features, or the roughness ratio. Both surfaces contain features that are appr oximately 2 m wide, spaced by approximately 2 m, and are between 4 m and 16 m long. The features ar e arranged in the following order to form a diamond shape: 4, 8, 12, 16, 12, and 8 m. A feat ure 4 m long starts the sequence again next to the final 8 m feature. The Sharklet AF a nd Recessed Sharklet AF surfaces are shown in Figure 4-1. Figure 4-1. Scanning electron micrographs of A) Sharklet AF (+1.8SK2x2) and B) Recessed Sharklet AF (-2.0SK2x2) surfaces. Evaluating n New surfaces were designed by altering th e number of distinct surface features (n) in the Sharklet AF design to evaluate the effect of th e number of features on spore settlement. The number of features (n) in the Sharklet AF design was varied from 1 to 5, with a smooth surface acting as a surface with n = 0. These surfaces ar e named the n-Series surfaces for simplicity. The surfaces have similar diamond arrangements of features as the original Sharklet AF surface, and comprise selected features of th e Sharklet AF design. The surface with n = 5 65

PAGE 66

contains an additional feature ( 20 m long) not found in the orig inal Sharklet AF surface (n = 4). The surfaces are shown in Figure 4-2. The +2.7SK2x2_n1 and +2.7SK2x2_n2 surfaces were previously evaluated48 under the notations GR0 and GR1, respectively. Figure 4-2. Sharklet AF surfaces va rying in the distinct number of features (n). A)SM (n = 0), B)+2.7SK2x2_n1, C)+2.7SK2x2_n2, D)+2.6SK2x2_n3, E)+2.9SK2x2 (n = 4), F)+2.6SK2x2_n5. Sample Preparation The two-dimensional patterns were produced in a darkfield photomask. Silicon wafers 4 in diameter and of (100) orie ntation were patterned with phot olithography using Shipleys S1813 photoresist. Deep reactive ion etching was used to etch the silicon wafers to the desired height, followed by an O2 plasma clean to remove th e photoresist from the wafers. Hexamethyldisilazane (HMDS) was vapor-deposite d on the wafers to pr event adhesion of the casting materials. The mold for the Rece ssed Sharklet was crea ted by solution casting Kraton styrene-ethylene/butylene-styrene (S EBS) block copolymer onto the silicon wafer Sharklet AF mold from toluene. In this so lution casting process, a solution of 10% Kraton series G1657 in toluene was placed into an enclosure sealed to a HMDS-treated silicon wafer 66

PAGE 67

patterned with the mold for the Sharklet AF. The evaporation rate of the solvent was controlled by adjusting the area of the enclosur e exposed to the ambient air. This process produced a mold for the Recessed Sharklet AF in approximately 3 days. Dow Corning SILASTIC T-2 poly(dimethyl s iloxane) elastomer (PDMSe) was prepared by mixing the resin and curing agen t in a 10:1 ratio for 5 minutes. The mixture was degassed for 30 minutes and cured against the sample mold s for 24 hours at approximately 22C. PDMSe sample surfaces were backed to glass microscope slides using allyltriethoxysilane as a coupling agent. Resulting samples consisted of a 1x 1 square area of topographical surface in the center of the slide with smooth PDMS e covering the remainder of th e slide area. Smooth PDMSe sample surfaces consisted of glass microscope slides covered entirely with smooth PDMSe. Samples surfaces were rinsed with 95% ethano l (5% distilled water) and blown with dry N2. Surfaces were stored in air. Feature dime nsions were analyzed through scanning electron microscopy. The ERIII for each surface was determined, and spore settlement relative to the settlement on a smooth surface was predicted using the ERIII values and regression equation. Settlement Assay with Spores of Ulva linza Ulva linza zoospore settlement assays were performed at the University of Birmingham, UK by Maureen Callow, James Callow, and John Finl ay. Three replicates of each surface type were assayed for spore settlement using standard ized spore settlement protocols as described previously 47, 102. Test samples were immersed in deionized water for 24 hours, followed by 2 hours in Tropic Marin artificial seawater (ASW) prior to the assay. Each sample surface was placed into an assay dish, and 10 ml of zoospore suspension (1.5 x 106 spores/ml) were added to each assay dish. The samples were incubated in the dark at approximately 20C for 45 minutes. The samples were rinsed to remove any unattached (i.e. motile) zoospores and fixed with 2% 67

PAGE 68

glutaraldehyde in ASW for 10 minutes. The glutaraldehyde solution wa s rinsed off using a sequence of ASW, 50% ASW in distilled wate r, and distilled water prior to drying. Settled spores were visualized and quantifie d by the autofluorescence of chlorophyll using a Zeiss epifluorescence micr oscope and Zeiss Kontron 3000 image analysis system.7 Thirty counts for fields of view of 0.033 mm2 were obtained for each repl icate surface giving a total of 90 counts per surface type. Settlement is reported as the mean and 95% confidence interval using 90 counts per surface. Two separate settlement assays were perf ormed. The smooth, Sharklet AF, and Recessed Sharklet AF surfaces were evaluate d in one assay. The n-Series surfaces (+2.8SK2x2_n(1-5) surfaces) were evaluated in the second assay along with the smooth surface. Results Regression Analysis Examination of the dataset from Schumacher et al., 2008, showed that settlement on the surfaces tested poorly correlated with ERII, with an R2 value of 0.4, as shown in Figure 4-3. The poor fit of the sett lement data to ERII suggested a need to reexamine the relationship. This was achieved using a transformation for the settlement data and altering the predicting variable (ERII to ERIII). The natural logarithmic transformati on (ln) was chosen to transform the settlement data for several reasons. First, this transformation yields a linear correlation between settlement and a predictor, such as the E ngineered Roughness Index, while constraining the settlement to positive values. Second, the transfor mation allows for nearly equal variance in the ln-transformed space regardless of the value of settlement. In contrast, the variance of settlement increases as settlement increases in the untransformed space. 68

PAGE 69

Figure 4-3. Regression of Ulva settlement to ERII for dataset from Schumacher et al. 2008 showing poor correlation (R2 = 0.4). Each of the two previous data sets, (i.e. those in Schumacher et al., 2007 and Schumacher et al., 2008) correla ted with the ERIII after using the natural logarithmic transformation. Combining both data sets into a single data set yields a regression with a strong correlation to ERIII. All three regressions are shown in Figure 4-4. The regression line for the combined data set is described by Equation 4-5. This equation can be rewritten as Equation 4-6, the biosettlement model in which S is the settlement on a surface with a given ERIII and SSM is the settlement on a smooth surface (whose ERIII = 0). 28.6 1047.7 mm spores Settlement ln2 2 IIERI (4-5) S SMnr S S1 1047.7 ln2 (4-6) 69

PAGE 70

Figure 4-4. Regressions of spore settlement with ERIII showing historical data correlates with ERIII. Data from (A) Schumacher et al., 200747 (R2=0.94), (B) Schumacher et al., 200848 (R2=0.81), and (C) the combined data set (R2=0.86). 70

PAGE 71

Evaluating S The measured dimensions and calculated para meters for the Smooth, Sharklet AF, and Recessed Sharklet surfaces test ed are shown in Table 4-1. Table 4-1. Dimensions and prediction variables for surfaces used to evaluate the effect of area fraction of feature tops (S) on spore settlement. When no unit is given, the quantity is unitless. Surface Height (m) Width (m) Spacing (m) r n S ERIII Smooth (SM) NA NA NA 1 0 0 0 Sharklet AF (+1.8SK2x2) 1.8 1.8 2.2 1.9 4 0.38 12 Recessed Sharklet AF (-2.0SK2x2) 2.0 1.9 2.1 2.0 4 0.60 20 Spore settlement relative to smooth was pred icted on these surfaces using the rewritten regression equation, Equation 4-6. The comp arison between predicted and experimental settlement values relative to th e smooth surface are shown in Tabl e 4-2, along with the absolute settlement values with 95% confidence intervals. Table 4-2. Settlement on surfaces used to ev aluate the effect of solid surface fraction (S) compared to the settlement predicted by th e biosettlement model. Settlement is reported as average of 90 counts with 95% confidence intervals Reduction from Smooth (%) Surface Ulva Settlement (spores/mm2) Predicted Experimental Smooth (SM) 386 12 NA NA Sharklet AF (+1.8SK2x2) 216 36 59 44 Recessed Sharklet AF (-2.0SK2x2) 85 29 78 78 The inversion of the surface from the Shar klet AF to the Recessed Sharklet AF surfaces caused the spores to settle in reduced numbers. The pattern design, the number of features in the pattern, and the dimensions of the features were the same or very similar for both surfaces. The only variable that was altered between these two surfaces was S. The spores 71

PAGE 72

were observed to settle within the recessed regions of both surf aces, indicating that they respond to S as predicted. The biosettlement model using ERIII predicted the settlement on the Recessed Sharklet AF (-2.0SK2x2) surface relative to smooth near ly exactly. The prediction of the Sharklet AF (+1.8SK2x2) was also predicted closely, bu t not as well as the Recessed Sharklet AF surface. In order to illustrate the effectiveness of the ERIII to predict the settlement, the three data setsSchumacher et al., 2007,47 Schumacher et al., 2008,48 and the current data setwere combined and are shown in Figure 4-5. The data clearly show that the Ulva spores settle in lower numbers when the value of S is increased. Figure 4-5. Regression of settlement data from 3 studies with ERIII (R2=0.88). Historical data sets are plotted as white tria ngles and dark grey squares, while new data are plotted as light grey diamonds. Dashed line indicates regression. 72

PAGE 73

Evaluating n The measured dimensions and calculated parame ters for the Smooth and n-Series Sharklet AF surfaces (+2.8SK2x2_n(1-5) surfaces) tested are shown in Table 4-3. Table 4-3. Dimensions and prediction variables for n-Series surfaces. When no unit is given, the quantity is unitless. Surface Height (m) Width (m) Spacing (m) r n S ERIII SM NA NA NA 1 0 0 0 +2.7SK2x2_n1 2.7 2.3 1.7 2.4 1 0.38 3.9 +2.7SK2x2_n2 2.7 2.3 1.7 2.4 2 0.43 8.4 +2.6SK2x2_n3 2.6 2.3 1.7 2.3 3 0.46 13 +2.9SK2x2 2.9 2.2 1.8 2.5 4 0.48 19 +2.6SK2x2_n5 2.6 2.2 1.8 2.3 5 0.49 23 Spore settlement relative to smooth was pred icted on these surfaces using the rewritten regression equation, Equation 4-6. The comp arison between predicted and experimental settlement values relative to th e smooth surface are shown in Tabl e 4-4, along with the absolute settlement values with 95% confidence intervals. Spore settlement (as percent reduction versus smooth) was well predicted by the ERIII for all of the n-Series su rfaces tested, except for the +2.7SK2x2_n1 surface. Table 4-4. Settlement on n-Series surfaces compared to values predicted by the biosettlement model. Settlement is reported as average of 90 counts with 95% confidence intervals. Asterisks represent groups that are statistically different. Reduction from Smooth (%) Surface Ulva Settlement (spores/mm2) Predicted Experimental SM* 800 70 NA NA +2.7SK2x2_n1* 782 72 25 2 +2.7SK2x2_n2** 437 49 47 45 +2.6SK2x2_n3*** 253 25 62 68 +2.9SK2x2**** 213 32 76 73 +2.6SK2x2_n5**** 186 15 82 77 Figure 4-6 shows the regression of the n-Series surfaces versus n and versus ERIII. The regressions are nearly identical, si nce the other variables in the ERIII (i.e. Wenzel roughness r 73

PAGE 74

and area fraction of feature tops S) are nearly equal for these n-Se ries surfaces. The data clearly show that the variable n is a critical variable in the biosettlement model (Equation 4-6). Figure 4-6. Regressions of n-Series Sharklet AF surfaces conf irming the number of distinct features in the design is a key parame ter in predicting settlement. Settlement correlates to the number of distinct features (R2 = 0.94) and to ERIII (R2 = 0.93). The settlement data versus ERIII is arbitrarily shifted to di stinguish data points between regressions Spore settlement generally agreed with the da ta obtained in Schumacher et al. for several of the n-Series surfaces. The surfaces that we re assayed in both the n-Series assay and Schumacher et al. 2008 include the +2 .7SK2x2_n1 (previously GR0), +2.7SK2x2_n2 (previously GR1), and +2.9SK2x2 (previously SK) surfaces. The settlement relative to smooth on these surfaces is shown in Table 4-5. The feat ure heights were slightly different between the two studies. Variations in featur e height causes variations in ERIII values and are expected to alter settlement de nsity slightly. The settlement on the +2.7SK2x2_n2 and +2.9SK2x2 in the current assay agree with the settlement on the GR1 and SK surfaces, respectively in Schumacher et al. 2008. The settlement on the +2.7SK2x2_n 1 surface in the current study, however, is 74

PAGE 75

considerably different than the settlement on the GR0 surface. The predicted value for settlement on the +2.7SK2x2_n1 surface relative to the smooth surface is close to the experimental value for the GR0 surface (25% versus 26%, respectively). Table 4-5. Comparison of spore settlement rela tive to smooth for surfaces assayed in both the present study and in Schumacher et al. 2008. Surface % Reduction in settlement vs. smooth Present Study Schumacher et al. 200848 Present Study Schumacher et al. 200848 +2.7SK2x2_n1 GR0 2% 26% +2.7SK2x2_n2 GR1 45% 47% +2.9SK2x2 SK 73% 67% The experimentally determined percent reductio ns versus smooth for all surfaces evaluated are plotted versus the percent reduction predicted by the biosettl ement model (Figure 4-7). Five of the seven surfaces fall within 6% of the predicted values. Figure 4-7. Performance of the biosettlement model in predicti ng spore settlement. Surfaces include the Sharklet AF, R ecessed Sharklet AF, and n-Series surfaces. Solid line indicates that the percent reduction is equal to the predicted value. Dashed lines indicate 6% difference from the predicted value. 75

PAGE 76

The levels of spore settlement in the sec ond assay (n-Series surfaces) were in general higher than those in the other assa ys. All of the assays have vari ation in the settlement numbers, including the settlement on the smooth surface in each study. Variation in settlement numbers between assays is due to biological va riation between different batches of Ulva collected from the shore. In order to account for these diff erences, the absolute settlement values were normalized with respect to the smooth surface in the study. The biosettl ement model with the two historical data sets (i.e. Schumacher et al. 2007 and Schumacher et al. 2008) and the new data sets (i.e. Sharklet AF, Recessed Sharklet AF and n-Series surfaces) are shown in Figure 4-8. The biosettlement model fits extremely well (R2 = 0.88) with the four separate studies: two historical studies a nd two new studies. Figure 4-8. Biosettlement model correlated to fo ur spore settlement assays. The settlement densities of the surfaces with topographic features are norma lized to the settlement on the smooth surface for each assay. Each a ssay is distinguished by marker shape and color. The spores were observed to settle such th at the spore center was within the recessed features on the Recessed Sharklet AF. The recesses were smaller than the spore body, preventing the spore from fully penetrating the re cess. Similarly, the spores on the Sharklet 76

PAGE 77

AF tended to settle centered between the features. Previous work has observed the tendency of the spores to settle centere d between the features, but all of the surfaces previously tested comprised protruding features. Th e spores prefer to settle in the areas represented by the term (1-S), even if that region is discontinuous. The reduction in settlement due to S is expected, as the spores pr efer to settle in recessed regions. Minimizing the area fraction of recesse d regions reduces the density of preferred regions on the surface. The reduction in settlement due to n is more difficult to explain. The inclusion of the number of distinct features, n, in the ERIII allows for the description of an entropy-like term. Larger values for n create a larger number of different local surface conditions, creating a surface with higher disorder The spores are motile and probe the surface during the settlement phase. After a single spore se ttles, additional spores tend to settle adjacent to existing spores. The spores may probe for simila rity within a region, not simply for a suitable location for single-spore settlement The probing process may be used by the spores to locate a settlement site where additional spores can also settle. In this way, th e disorder of the surface may contribute to the response of the spores. Another possible explanati on for the contribution of n to the settlement is the number of distinct features a spore can contact simultane ously. Since the spores prefer to settle at intersections of multiple features, the number of features in the design correlates to the number of distinct features the spores would contact simultaneously in these intersection locations, up to a maximum of 4 features. The spore can only contact the types of features that are in the design, and can only contact up to four features at a tim e due in the intersections. The surface with 5 distinct elements (+2.7SK2x2_n5) would be exp ected under this hypothesis to have a similar spore settlement as the 4-element Sharklet AF (+2.9SK2x2), since a spore can only contact 77

PAGE 78

four distinct features simultaneously on each of these surfaces. The settlement on these two surfaces was not different statistically, but the +2.7SK2x2_n5 surface had slightly better performance than the +2.9SK2x2 surface. The spor es may be sensing different stresses on their cell membranes caused by the different features, similar to the hypothesis by Schumacher et al. 2008, or there may be another property of the surface the cell may be sensing. The biosettlement model predicts that the spore settlement will decrease asymptotically toward zero settlement as the number of features in the design increases. However, as additional features are added as longer features, areas of the surface begin to approach the local properties of ridges. Previously, the ridges surface was the least effective of seve ral topographical surfaces (including the Sharklet AF surface) at reducing spore settlement.47 The features do not have to be consecutively longer, but could be of different geometry. As the number of distinct features approaches infinity with features of simila r size, the surface could approach a surface with randomly placed features. Since the spore would not be able to probe th e entire surface, the correlation of spore settlement to the number of distinct features is expected to have a limit. Further evaluation of the effect of the number of distinct features in the designe.g. surfaces with larger numbers of distinct featurescoul d be used to further develop the predictive biosettlement model. All of the surfaces whose perfor mance is incorporated into the biosettlement model have depressed regions whose width is small enough to prevent spores from penetrating the recesses. Previous work has shown that spore settlement is increased on ch annels whose depressions are wide enough for the spores to touch the floor.7 The biosettlement model has not been expanded to predict settlement on surfaces which increase settle ment, such as channels at least 5 m wide. 78

PAGE 79

Further work could be performed to broaden the predictive capability of the biosettlement model beyond surfaces based on 2 m spacing between features. A surface based on the Sharklet AF surface, but with curved features instead of rectangular features woul d provide a similar ERIII (and similarly predicted settlement), but alter the mechanical properties of the features from those of the Sharklet AF. The bending moment of the features would change considerably, as would the curvature of the feature walls. The curvature of the spore membrane may interface wi th one side of the bent features, but would have smaller contact area on the other side. The bending moment of th e features could be separately evaluated using materi als with higher elastic moduli, but the curvature of the walls would need to be evaluated using a newly designed topography. The biosettlement model includes two parameters that are important in the wettability: the Wenzel roughness r, and the fraction of feature tops S. These wettability parameters clearly influence the settlement of spores. There are othe r parameters that are po tentially influencing the settlement and can be incorporated into the biosettlement model. These parameters include the wettability of the substrate material, the mechani cal properties of the mate rial, and the size of the organism. Altering the wettability of the substrat e material (i.e. a smooth surface) over a range of wettabilities will further influen ce the wettability of the topographies from very hydrophilic to very hydrophobic. The mechanical properties of th e material affect the bending moment of the features and would be expected to alter cell response to the to pographies. The size (i.e. diameter and hydrodynamic volume) and shape (e.g. rod or s phere) of the settling or ganism relative to the topography may be important in in corporating more organisms into the biosettlement model. The biosettlement model has the potential to be expanded to more organisms, more topographies, and other materials. 79

PAGE 80

Conclusions The data presented show that the Ulva spores respond independently to both the area fraction of feature tops, S, and the number of fe atures in the design, n. As S increases or n increases, Ulva spores settle in lower numbers. The data presented demonstrate the first report of the prediction of Ulva settlement on topographical surfaces. The biosettlement model based on the ERIII correctly predicted the Ulva settlement performance of the previously-untested Recessed Sharklet AF (-2.0SK2x2) surface relative to a smooth surface. In a separate assay, the biosettlement model correctly predicted the Ulva settlement on two ne wly designed surfaces (+2.6SK2x2_n3 and +2.6SK2x2_n5), and predicted th e settlement of se veral previously evaluated surfaces. Five of the seve n topographical surfaces tested had Ulva settlement within 6% of the predicted values, in terms of per cent reduction versus smoot h. The regression model has excellent correlation, especially for a biologi cal system. This predictive model may be useful for other organisms, and will be useful in designing new antifouling surfaces. 80

PAGE 81

CHAPTER 5 NAVICULA SETTLEMENT AND RELEASE ON SHARKLET AF-BASED TRANSLATIONALLY SYMMETRIC TOPOGRAPHIES Introduction Raphid diatoms are group of single-celled algae and are the mo st abundant algal component of most marine biofilms. Thes e algae are resistant to copper biocides 16 and adhere strongly to commercial foul release coatings.30, 32 Foul release coatings function only when the vessel is moving, so fouling communities on a st ationary vessel may grow and become even more difficult to remove. Diatom-dominated slim es have been reported to increase drag of marine vessels and increase propulsi on power required by approximately 20%.14 Heterogeneous biofilms, many of which contain diatoms, increas e corrosion rates of metal due to gradients in dissolved local ion concentrations. The presence of raphesslits along the leng th of the cell body from which adhesive mucilage is excreteddistinguish ra phid diatoms from non-raphid diatoms.103 Raphid diatoms first approach a surface via gravity or water motion, and the cells contact the surface via extensions of mucilage excreted from the raphe s. After contact, diatoms position themselves to bring one of the raphes into contact with the surface, allowing initia l attachment. Mucilage excreted from the raphe provides firm attachment. This firm attachment can be released by the diatom to allow the cell to leave the surface and colonize new surfaces. Raphid diatoms glide along the surface, secreting a trail of adhesive from the raphe in contact with the surface. In addition to movement in translation, these cell s can rotate to position th eir elongated bodies in a position more suitable for their colonization.104 Topographies with translationally symmetric pa tterns have been effective in reducing the settlement of several marine organisms. Each or ganism that responds to these topographies has a critical dimension to which they resp ond, as hypothesized by Callow et al. 2002. Ulva linza 81

PAGE 82

motile zoospores were inhibited from settling by surface features spaced by 2 m,46-48 while the settlement of these spores is promot ed by features spaced by at least 5 m.7 Balanus amphitrite barnacle cyprids were shown to be inhibited by several topogra phies with features 20 m and spaced by 20 m, an order of magnitude larger than the dimensions of the surfaces that inhibited Ulva spores.9 The attachment of two species of diatoms was decreased by sinusoidal ridge/channel topographies when the cell body was too large to sit w ithin the troughs.8 These two species, Amphora sp. and Navicula jeffreyi, were inhibited by ridges 2 m wide and spaced by 2 m. The sinusoidal ridge/channels decreased attachment by 47% and 36%, respectively. An additional two species were examined, but these cells were sm aller than the scale of the ridges and could sit within the troughs on both of the ridge topographies. These two sp ecies settled in similar or higher numbers on the topographies compared to the smooth surface. Although the critical dimension has been s hown to be important, not all topographies designed with the critical dimension exhibit the sa me effectiveness against fouling. The Sharklet AF surface consistently inhibited Ulva spores by 67-86% versus an otherwise identical smooth control. This surface consistently outperform ed other surfaces tested, including channels, hexagonally packed circular pill ars, a topography containing bot h triangles and pillars, and a variety of oneand two-element surface designs containing rectangular features. The diatoms Amphora sp. and N. jeffreyi tested by Scardino et al. 2006 se ttled in much lower numbers of ridges 2 m in width and spaced by 2 m than to projections approximately of the same width and spacing. Theories to explain the observed differences among surfaces in fouling response by diatoms and Ulva spores have been developed independe ntly. The Attachment Point Theory was 82

PAGE 83

proposed by Scardino et al. 2006 to explain diffe rences in diatom attachment among surfaces, while the Engineered Roughness Index wa s developed by Schumacher et al.47 and expanded as described in Chapter 4 to explain differences in Ulva spore settlement among surfaces. Attachment Point Theory is the theory that th e attachment of diatoms is determined by the number of points to which the diatom cell can attach. This theory was de veloped to explain the dependence of size on the diatom response to th e sinusoidal/ridge topogra phies. The attachment of the four species of diatoms tested increased when the number of points of cell-surface contact, though an equation relating the attachment points to cell re sponse was not developed. The second Engineered Roughness Index (ERIII) relationship correctly predicted the performance of three previously-untested topogra phical surfaces using th e relationship shown in Equation 5-1. II SMERI S S21047.7 ln (5-1) In Equation 5-1, SSM is the Ulva settlement on a smooth poly(dimethyl siloxane) elastomer (PDMSe) surface, and S is the Ulva settlement on a topographically modified PDMSe surface with a given ERIII. The value -7.47x10-2 is empirically found from the regression and can be thought of as the sensitivity of the organism to the ERIII for PDMSe surfaces. It is unclear whether this sensitivity will be affected by the subs trate material, and if the sensitivity is speciesspecific. The work by Scardino et al. establishes that the colonization of a surface by some species of diatoms is affected by topographical surfaces. However, variations in the topographies tested have been limited to scaling of ridge topograp hies and a single topography of projections. Including new variations of topography including the width of features, spacing between features, and the number of features would provi de increased insight into the attachment and 83

PAGE 84

attachment strength of these common marine foulers. The Sharklet AF surface, which has outperformed other topographies for Ulva linza settlement, may be usef ul for decreasing fouling by diatoms. Materials and Methods Surface Designs Two iterations of surface designs were incorporat ed into this study. The first iteration was designed by James Schumacher and contained surfaces with square pillars 6 m in width and 3 m tall designed with different spacing between th e features. The spacing wa s varied, starting at 2 m, which was shown by S cardino et al. to reduce Navicula attachment. The spacing between features was set at 2, 4, or 6 m. Unlike surfaces tested by Scardino et al 2006, whose features were sinusoidal and equal in width and spacing, the surfaces designed here have features with vertical sidewalls, flat tops, a nd have feature width set at 6 m regardless of spacing. These surfaces are shown in Figure 5-1, and are labe led according to Equation 5-4. The calculated characteristics of these surfaces including the values for ERII and ERIII are shown in Table 5-1. Figure 5-1. Optical images of s quare pillar topographical surf aces: A) +3P6x2, B) +3P6x4, and C) +3P6x6. Surface Designation = [height][surface design][width]x[spacing] (5-4) In Equation 5-4, protruding features have a + preceding the height term, the surface design is P to designate pillars, and the di mensions are in m. A smooth surface is simply designated SM with no height, width, or spacing terms. 84

PAGE 85

Table 5-1. Characteristics of square pillar surfaces varying in spacing Surface designation Width (m) Spacing (m) Height (m) r S ERII ERIII Smooth NA NA NA 1 0 2 0 +3P6x2 6 2 3 2.1 0.56 9.7 4.9 +3P6x4 6 4 3 1.7 0.36 5.4 2.7 +3P6x6 6 6 3 1.5 0.25 4.0 2.0 Following results from the Navicula response to the series of pillars, a second set of surface designs was created. This set of surfaces consisted of va riations of the Sharklet AF surface, which is particularly effective against Ulva spores. These surfaces were designed to determine the effect of widt h and spacing independently on Navicula attachment and attachment strength, and to determine if attachment or attachment strength correlates to the ERIII. The height of all of these surfaces was designed to be approximately 3 m. Two series of surfaces were created based on the Sharklet AF design. The first series contained surfaces whose features vary in widt h and spacing. The second series consisted of surfaces with 2 m wide features spaced by 2 m. The number of distinct features ( n in ERIII) in this series varied from 1 to 5, with a smooth surface acting as a surface with n = 0. The 4-element Sharklet AF with 2 m feature width and 2 m spacing was included in both series. The surfaces with 1 and 2 distinct features have been previously evaluated for Ulva settlement under the designations of GR0 and GR1, respectively.48 The surfaces are shown in Figures 5-2 and 5-3 with the designation as in Equation 5-5. Surface Designation = [height][surface design][width]x[spacing]_[distinct features] (5-5) The naming designation of Equation 5-5 is similar to Equation 5-4, with the addition of the number of distinct features to distinguish among Sharklet AF -based designs. These Sharklet AF surfaces have the surface design SK. When the distinct features term is omitted, the number of distinct features is n = 4, which is the number of f eatures in the original Sharklet 85

PAGE 86

AF design. The calculated surf ace parameters, including ERII and ERIII are shown in Tables 52 and 5-3 and correspond to the surfaces in Figure 5-2 and 5-3, respectively. Silicon wafer molds Sample molds were created with photolithography and deep reactive ion etching. A quartz photomask with the pattern etched out of a layer of chromium was used for photolithography. Silicon wafers 4 inches in diameter and of (100) orientation were vapor deposited with hexamethyldisilazane (HMDS) to promote adhe sion of photoresist. Shipleys Microposit S1813 photoresist was spun onto the wafers at 4000 rp m. The photoresist was exposed to 405 nm ultraviolet light through the photoresist and deve loped using AZ MIF 312. Silicon wafers with patterned photoresist were et ched using deep reactive ion etching by employing the Bosch process. In this process, cycles of etching exposed silicon with SF6 were alternated with passivation cycles using C4F8. The length of each cycle and numb er of cycles is determined by the depth of the desired etched pattern. Afte r etching, the wafers were cleaned with O2 plasma for 5 minutes. Figure 5-2. Scanning electron mi crographs of A) +2.8SK2x5, B) +2.7SK2x2, C) +2.4SK5x2, and D) +2.5SK10x2. 86

PAGE 87

Table 5-2. Characteristics of su rfaces varying in width and spacing Surface designation Width (m) Spacing (m) Height (m) r S ERII ERIII Smooth NA NA NA 1 0 2 0 +2.8SK2x5 2 5 2.8 1.7 0.21 4.2 8.4 +2.7SK2x2 2 2 2.7 2.4 0.46 8.8 18 +2.4SK5x2 5 2 2.4 1.8 0.66 10 21 +2.5SK10x2 10 2 2.5 1.5 0.79 14 29 Figure 5-3. Scanning electron micrographs of surfaces with varying numbers of distinct features, n A) SM (n = 0), B) +2.9SK2x2_n1, C) +2.9SK2x2_n2, D) +3.0SK2x2_n3, E) +2.7SK2x2 (n = 4), and F) +2.9SK2x2_n5. Table 5-3. Characteristic s of surfaces varying in n All designed width and spacing was 2 m. Surface designation Distinct features Height (m) r S ERII ERIII Smooth 0 NA 1 0 2 0 +2.9SK2x2_n1 1 2.9 2.5 0.37 7.9 3.9 +2.9SK2x2_n2 2 2.9 2.5 0.43 8.8 8.8 +3.0SK2x2_n3 3 3.0 2.5 0.46 9.4 14 +2.7SK2x2_n4 4 2.7 2.4 0.46 8.8 18 +2.9SK2x2_n5 5 2.9 2.5 0.47 9.3 23 Material The platinum catalyzed SILASTIC T-2 PDMS e (Dow Corning ) system was prepared by mixing ten parts resin and one part curi ng agent by weight for 5 minutes, followed by degassing under vacuum for 30 minutes. The dega ssed mixture was cured against HMDS-treated 87

PAGE 88

silicon molds for 24 hours at approximately 22C to form patterned films. These films were backed to allyltrimethoxysilane-treated glass microscope slides by curi ng a layer of PDMSe between the microscope slide and the sample film in a second step. Three replicate samples for attachment and three replicate samples fo r strength of attachment were cast. Initial Attachment of Navicula Attachment and release assays of Navicula perminuta were performed at the University of Birmingham, UK by Maureen Callow, James Callow, and John Finlay. Navicula perminuta cells were cultured in F/2 medium for 3 days to bri ng the cells into log growth phase. Cells were washed 3 times in fresh medium before harves ting and diluting to give a suspension with a chlorophyll a content of approximately 0.25 g ml-1. Cells were settled in individual dishes containing 10 ml of suspension at appr oximately 20C on the laboratory benches. After 2 hours the slides were exposed to a submerged wash in seawater to remove cells which had not properly attached (the immersion process avoided pass ing the samples through the air-water interface). Samples were fixed in 2.5% glutaraldehyde, air dr ied and the density of cells attached to the surface was counted on each slide using an imag e analysis system attached to a fluorescent microscope. Counts were made for 30 fields of view (each 0.064 mm2) on each slide. Strength of Navicula Attachment Slides settled with Navicula were exposed to a shear stress in a water channel for 5 minutes. The pillar samples were exposed to 53 Pa, while the Sharklet AF-based samples were exposed to 24 Pa. Samples were fixed and the nu mber of cells remaining attached was counted using the image analysis system described above. Microscopy Images of live diatoms were taken after one hour of settlement on th e pillar topographies prior to any washing procedure us ing low power magnification. Thes e images were taken as part 88

PAGE 89

of the settlement assays. From images take n under the microscope the position of diatoms relative to the pillars (i.e. above or between) was made for a singl e field of view on each surface. Results and Discussion Pillar Topographies The pillar topographies containe d similar numbers of attached Navicula cells as the smooth PDMSe surface (Figure 5-4A ) after rinsing to remove cells which had contacted the surface but did not attach. All surfaces performed similarly in initial attachment, indicating the diatoms were equally likely to attach on all surfaces. Diatom strength of attachment to PDMSe is much greater than to glass.31 Consequently removal of cells was rela tively low after shear flow of 54 Pa. One topographical surface (+3P6x2) showed improved release performance versus the smooth PDMSe surface, while the other two surfaces performed similarly to the smooth surface. The +3P6x2 surface decreased numbers of Navicula cells by 25% after exposure to shear (Figure 54B) and more than doubled the percent release ve rsus the smooth control (Figure 5-4C). The +3P6x4 and +3P6x6 surfaces had similar numbers of cells after exposure to shear as the smooth surface, with similar removal rates. Navicula cells, which are approximately 12 m lo ng and 4 m wide, were too large to fit between the features on the +3P6x2 surface, causing nearly all of the Navicula cells to remain on top of the +3P6x2 topography. The few diatoms on this surface which were not situated above the pillars were generally wedged at angle betw een them, sticking out from the surface and not properly attached. In contrast, the cells were small enough to fit between the features on the +3P6x4 and +3P6x6 surfaces, allowing most of the cells on these tw o surfaces to reside in the recessed regions between the features (Figure 54D). Cells residing between the features, as on the +3P6x4 and +3P6x6 surfaces, we re protected from the shear stress produced by the water flow. Cells on the +3P6x2 surface had only part of the cell in contact with the surface, generally 89

PAGE 90

providing only two points of atta chment. The 6 m features are not long enough to allow the full length of the 12 m long diatom to fit on top of the square pillar. The cells on the +3P6x2 surface were not protected from the flow, and less of the cell could cont act the surface compared to cells on the smooth surface. These factors likely contributed to lower attachment strength of the diatoms on the +3P6x2 surface than on the other surfaces. Figure 5-4. Navicula response to Pillar topographies. A) at tachment, B) number of cells after exposure to shear, C) percentage of cells released by shear, and D) percentage of Navicula cells above topographies. Bars show 95% confidence inte rvals. Asterisks denote groups that are statistically different (B, C: Tukeys test, = 0.05, D: Marascuilo procedure, = 0.05). The attachment occurs from contact between the surface and mucilage from the raphe. If the cell falls between the features and contacts the bottom of the recessed region, the raphe can contact the surface and provide firm attachment Similarly, raphid diatoms on a smooth surface can place their raphes in contact with the su rface. Cells centered betw een two features, but 90

PAGE 91

unable to fall between the features, have a limited area of contact between the cell and the surface. This limited contact makes raphe-surface contact difficult, especially considering diatoms reposition themselves us ing mucilage from the raphe. The surface with 2 m spacing between features outperfor med the 4 and 6 m spacing surfaces in terms of release, and performed as we ll in terms of attachment. This data reinforces conclusions by Scardino et al. 20068 that the 2 m dimensions are effective at reducing Navicula fouling. There are several impor tant distinctions between the previous study and the data included here. First, the conti nuous ripples in Scardi no et al. 2006 had feature width of 2 m, while the pillar topographies had features 6 m wide and long and were not continuous. Second, the 2 m ripples reduced attachment (attachment strength was not evaluated), while the 6 m pillars spaced by 2 m reduced attachment strength, but had no effect on attachment. Sharklet AF-Based Topographies The surfaces based on the Sharklet AF were aimed at determining the effect of feature spacing, feature width, and the ERI on Navicula attachment and attachment strength. In evaluating the effect of feature width and sp acing, the +2.8SK2x5, +2.4SK5x2, +2.5SK10x2, and the smooth surface had similar numbers of attached Navicula cells prior to shear. The +2.7SK2x2 surface, with feature width of 2 m and feature spacing of 2 m, had reduced attachment of Navicula The +2.7SK2x2 surface reduced att achment by 35% versus the smooth control. The attachment data is shown in Figure 5-5A. The attachment data supports previous data and conclusions by Scardino et al.8 The surfaces with 2 m feature width and 2 m sp acing reduced initial atta chment of diatoms by approximately 35% versus a smooth surface. This reduction is similar to the reduction on the sinusoidal ridge/channel topographies with 2 m wide ridges and 2 m wide channels found by Scardino et al. 2006 (labeled the 4 m wavelength ri pple surface). Surfaces that allowed cells to 91

PAGE 92

sit on top of the features, as on +2.5SK10x2 a nd +2.4SK5x2, did not show improvements over smooth in terms of initial attachment. Likewise, surfaces that allowed ce lls to fit between the features, as in +2.8SK2x5, performed similarl y to smooth for initial attachment. This is consistent with the pillar topographies, which al lowed the diatoms either to fit on top of the features or to fit between the features and did not reduce initial attachment of Navicula The numbers of diatom cells after exposure to flow generally follow the attachment data. The number of cells on the +2.7SK2x2 surface was approximately 40% lower than on the smooth surface after shear. The +2.8SK2x5 surface showed some reduction, though the reduction was not statistically significan t. Figure 5-5B shows the numbers of Navicula cells on each surface after shear. The percent released of attach ed diatoms by flow is shown in Figure 5-5C. The decreased numbers of cells versus smooth on the +2.7SK2x2 afte r flow is primarily due to a decrease in the number of cells initially atta ched. Unlike on the pillar topogra phies, the surfaces with 2 m spacing and wide features (+2.4SK5x2 and +2 .5SK10x2) did not show increased release compared to smooth. This may be because the wi de features in the Sharklet AF-based designs are also longer than the pillar topographies features. The lengths of the features vary from 10 m to 40 m for +2.4SK5x2 and from 20 m to 80 m for +2.5SK10x2; while the pillars were consistently 6 m x 6 m. The longer featur es of the +2.4SK5x2 and +2.5SK10x2 allow the diatom cells to entirely attach to the tops of these features. In contrast, the di atoms could only partially attach to the Pillar surfaces due to the cell being long er than the pillar. The number of cells initially attached to the Sharklet AF-based surfaces with 2 m spacing, 2 m width, and number of features va rying from 1 to 5 are shown in Figure 5-6A. All of the Sharklet AF-based surfaces with 2 m width and 2 m spacing reduced initial Navicula 92

PAGE 93

attachment by between 32% and 41% versus sm ooth. These Sharklet AF surfaces performed similarly, and all were statistically different than smooth. The number of features in the design, at least within the range of 1 to 5, did not affect Navicula attachment. Instead, a feature width of 2 m and spacing between the features of 2 m ar e the critical dimensions. At approximately 3 m in height, all of these SK2x2 surfaces performed similarly against attachment. Figure 5-5. Navicula response to surfaces varying in width and spacing. A) attachment, B) number of cells remaining after shear, a nd C) percent cells removed by shear from Sharklet AF-based surfaces with varying width and spacing. Bars represent 95% confidence intervals. Asterisks denote groups that are statistically different (Tukeys test, = 0.05). After exposure to 24 Pa shear, most of the SK 2x2 surfaces performed similarly, with the exception of the +2.9SK2x2_n5 surface (Figure 56B). The +2.9SK2x2_n5 surface had slightly fewer remaining cells than the smooth control, an d slightly more remaining cells than the other SK2x2 surfaces. The percentage of cells releas ed from these surfaces, including smooth, was 93

PAGE 94

similar with the exception of the +2.9SK2x2_n5 surface (Figure 5-6C). This surface had no release of attached diatoms, while the other surf aces released approximate ly 20% to 30% of the attached cells. Figure 5-6. Navicula response to n-Series topographies. A) Attachment, B) number of cells after exposure to shear, and C) percent cells rel eased by shear from Sharklet AF-based surfaces with varying number of unique features ( n ). Bars represent 95% confidence intervals. Asterisks denote groups that are statistically diff erent (Tukeys test, = 0.05). Observations showed that Navicula cells were too large to fit between the 2 m spacing that separated the features and consequently a ttachment was confined to the surface provided by the feature tops. Due to the small width of the features, many of the diatoms straddled several features. As on the +3P6x2, diatoms that were to uching two features but unable to fall between them would be loosely attached, as the raphe ha s limited contact with the surface. As the number of features in the pattern increased, the length of the longest available feature increased as well, 94

PAGE 95

from 4 m (+2.9SK2x2_n1) to 20 m (+2.9SK2x2_n5). The availabi lity of longer features increased the proportion of continuous surface available for adhesion, potentially increasing attachment strength of some of the diatom s. In the pattern with the most features (+2.9SK2x2_n5), the central ridges were approximately the same length as a diatom cell (20 m) All of the Sharklet AF-based surfaces were run together in a single experiment, and comparisons between the two sets are therefor e legitimate. Surfaces whose width and spacing was 2 m had improved performance in preventing attachment by Navicula consistently reducing attachment by approximately 35%. Surf aces whose width or spacing was at least 5 m performed similarly to the smooth surface in terms of Navicula attachment. This nearly equal settlement on these surfaces indicates Navicula cells simply attempt to attach if they come into contact with the surface and have little or no additional sensing ab ilities for initial attachment. Removal rates were similar across all surfaces, with the exception of the +2.9SK2x2_n5 surface, whose removal rate was zero. The +2.8SK2x5 rem oval rate was slightly higher than the other surfaces, though it was statistically different fr om only the +2.5SK10x2 surface. The attachment numbers on all of the Sharklet AF surfaces are shown in Figure 5-7. Figure 5-7. Comparison of Navicula response among all Sharklet AF surfaces. A) Attachment and B) percent release. 95

PAGE 96

The biosettlement model, when applied to Navicula predicted several th ings that were not observed experimentally. First, the +2.5SK10x2 surface was predicted to outperform all other surfaces. However, this surface was found to have similar numbers of attached cells and a similar percent release as the smooth surface. The +2.5SK10x2 surface also underperformed several other surfaces. The +2.4SK5x2 was predicted to outperform all surfaces except the +2.5SK10x2 and +2.9SK2x2_n5 surfaces. The +2.4SK5x2 surface pe rformed similarly to the smooth surface and had higher numbers of attached cells than the surfaces with 2 m feature width and 2 m spacing (+2.8SK2x2_nX surfaces). The Navicula response was predicted to correlate to the number of distinct features, n but this was not found experimental ly. All of the su rfaces with 2 m feature width and 2 m spacing had similar numbers of atta ched cells, and the release rates were not statistically different with the exception of the +2.9SK2x2_n5 surface. The Navicula attachment and release were also co mpared to the components of the ERIII and to the ERII to determine any other correlations (Table 5-4). The Navicula responses did not correlate with the ERII, ERIII, or two of the variables involved in the biosettlement model. The only variable that correlated with Navicula attachment before and after shear was the Wenzel roughness ( r ). The high correlation coefficients (R2) between attached cells (before and after flow) and Wenzel roughne ss are somewhat misleading. The values of r are not widely distributed, but instead clumped in two main groups: surfaces with 2 m feature width and 2 m spacing (+2.8SK2x2_nX surfaces), and surfaces with larger feature width or larger spacing. The high correlation coefficient is a result of differences in Navicula response to these two groups of surfaces, not necessarily a resu lt of the Wenzel roughness values. The Wenzel roughness may play a role in determining Navicula attachment, but the data suggest that the width and spacing of 2 m is the determining factor. 96

PAGE 97

The fact that the bioset tlement model predicts Ulva linza spore settlement but does not correlate with Navicula perminuta attachment is not surprising. The Ulva spores are flagellated and swim toward the surface, while the diat oms do not have a motility mechanism when approaching the surface. The Ulva spores can use their motility to search the surface prior to settling, while the diatoms cannot choose where to attach among locations. The difference in motility may be one of the reasons the two species do not respond the same way to the Engineered Roughness Index, and other biological differences ar e likely to contribute to differences in response. Table 5-4. Correlation coefficients (R2) for regressions of Navicula response to Engineered Roughness Index ERII ERIII r 1/(1S) n Navicula attachment 0.02 <0.01 0.84 0.11 <0.01 Navicula after shear <0.01 0.09 0.74 0.24 0.01 Percent cells released by shear 0.07 0.20 0.02 0.03 0.10 In terms of attachment point theo ry, proposed by Scardino et al. 2006,8 the results are supportive of the theory, though some difficulties persist. All of the surfac es that allow the diatoms to sit on top of or betw een the features had similar numbers of attached cells as the smooth surface. All of these surfaces would be described as having multiple attachment points by the attachment point theory. All of the Sharklet AF-based surfaces whose width and spacing are 2 m, have similar densities of attached di atoms. The cells cannot sit between or on top of the features in these su rfaces, limiting the number of points in contact between the cell and the surface. The decrease in the density of attached diatoms on these surfaces supports attachment point theory. Not all of the data supports a ttachment point theory. The cells are longer than many of the features in these surfaces, in particular the +2.9SK2x2_n1 surface. On this surface, each cell is 97

PAGE 98

long enough to contact at least 4 of the 4 m long features. If the cell is lying in the direction perpendicular to the length of the features, the cell is long enough to contact approximately 8 features. Both of these scenarios were observed microscopically. Diatom cells are able to contact fewer features as the number of distinct featur es in the surface design increases. According to attachment point theory, the number of attached diatoms should decrease as n increases, which was not observed from the data. The mechanism by which diatoms sense the su rface to determine whet her to attach or where to attach is not entirely understood. The raphes are known to control motility and provide adhesion through excreted mucilage, and the inte raction between the mucilage likely has a significant role in determining adhesion. C ontact between a raphe and the surface likely determines whether the diatom will attach. Of th e surfaces tested, all surfaces which had a large enough area for the entire diatom cell to contact the surface had similar numbers of attached cells. These surfaces allowed one of the raphes to fully contact the surface. Surfaces spaced by 2 m and with feature width of 2 m likely prevented or reduced raphe-surface contact by suspending the diatom cells between features. Su spended cells whose raph es had little or no contact with the surface would not be able to attach, which would explain the ability of this subset of surfaces to reduce attachment. Attachment point theory currently uses the number of points of cell-surface contact, but it may be more appropriate to approach the theory from a raphe-surface point of view, since adhesion occurs at the raphes. Area or length of raphe-surface contact or number of contact po ints between raphe and surface ma y be appropriate modifications to the attachment point theory. Further testing of diatom re sponse to topographies could be performed using additional surfaces with 2 m feature width and 2 m spacing. Several of thes e surfaces with different 98

PAGE 99

values for Wenzel roughness and S have been previously evaluated for Ulva settlement by Schumacher et al. 2007.47 The roles of these variables in th e fouling response of diatoms could be tested using these surfaces. Conclusions The data suggest that in or der to reduce fouling by cells Navicula perminuta diatoms, the dimensions of engineered topographies should be designed to prevent the cells from sitting on top of or between the features. Surfaces with areas between features la rge enough for the cell to sit on top of or between the features had Navicula attachment similar to the smooth surface. Sharklet AF-based surfaces with feature width of 2 m and spaced apart by 2 m consistently reduced Navicula attachment by between 30% and 40%. Th e percentage of cells released by shear flow was in general similar among the surfaces with several exceptions. The critical dimension of 2 m appears to be the most impor tant factor in determining Navicula attachment. The biosettlement model did not co rrelate with cell attachment or release. Differences in motility between Ulva linza spores and Navicula cells may explain differences in response to topographies. The data, in line with previous findings for patterns effective against Ulva spores and barnacle larvae, show that there is a st rong relationship between the dimensions of the colonizing organism and the dimensional sc ale of topographical features and spacing. 99

PAGE 100

CHAPTER 6 TECHNIQUE FOR MAPPING SETTLEMENT ON SURFACES AND THE PREFERENTIAL SETTLEMENT LOCATIONS OF ULVA LINZA ALGAL ZOOSPORES Introduction Several foulers have been shown to prefer certain localized regions of topographical and chemical surfaces. However, studies of this pr eference have been limited to either simple geometries (i.e. channels/ridges or pillars) or to observational evidence. Ulva linza spores have been shown in a number of studies to settle in depressed regions of topographies, especially when the spore can touch both the side of a feature and the floor region.7, 54, 55 Both Ulva spores and diatoms settled in depressions on hierarchically wrinkled surfaces.53 A variety of species were found 56 to foul preferentially in pits between hemispherical protrusions on the millimeter scale. Several bacteria species have shown preferential attachme nt on several channel topographies under static and flow conditions.57, 58 The preferential settlement of Ulva spores to localized locations extends to chemical patterns at least for one dimensional stripe patterns (analogous to ridges/channels). The spores prefer fluorinated chemical stripes instead of poly(ethylene glycol)-modified chemical stripes.59 Preferential locations for Ulva settlement have been observed on the more complex Sharklet AF and two-element topographical designs, but the preference has not been quantified. Spores were observed to settle centere d between at least two features of the Sharklet AF, specifically between adjacent diamond patterns.46 The spores could not penetrate the pattern, and instead bridged the features. This preference was observed on surfaces with two rectangular elements spaced at the 2 m critical dimension for Ulva 48. On one surface with 2 m pillars and 10 m rectangles, the Ulva spores avoided the triangles, despite being able to entirely fit on the triangles.47 100

PAGE 101

In all of these studies on surfaces more comple x than a single element, the preferential settlement locations were not quantitatively evaluated and relied on observational evidence. Some one-dimensional analysis has been pe rformed for channels, but two-dimensional quantitative analysis has not been performed Quantitative analysis would allow for the identification of preferential si tes, some of which may not be apparent with observational analysis. The field of statistically mode ling clusters attempts to identify locations of elevated incidence of some event.105 This approach is commonly used to identify clusters of disease cases in order to locate causes of these diseases.106 Geographical regions are not typically translationally symmetric in shape, geographical features, and human activity. Translational symmetry is rarely used in cluste r modeling due to the non-symmetric nature of most of the areas of interest. Similar techniques, modified for translational symmetry have the potential to map cell location on translationally symmetric t opographies. This mapping would allow for the identification of pattern features wh ich increase or reduce cell response. Materials and Methods Topographical Surfaces Engineered topographical surfaces th at were previously assayed for Ulva linza spore settlement were used for the evaluati on of preferential settlement locations.48 The Gradient surfaces (GR0, GR1, GR2, GR3, GR4, and GR5), the Sharklet AF, and Recessed Sharklet AF surfaces have features 2 m wide, spaced by 2 m, and between 2 and 3 m tall. These surfaces are shown in Figure 6-1. The dimensions of the surfaces features were analyzed using scanning electron microscopy, and the surfaces were rename d using the designation in Equation 6-1. Surface Designation = [height][surface design][width]x[spacing] (6-1) 101

PAGE 102

Figure 6-1. Scanning electron micrographs of Grad ient and Sharklet AF surfaces. A) GR0, B) GR1, C) GR2, D) GR3, E) GR4, F) GR 5, G) Sharklet AF, and H) Recessed Sharklet AF. In Equation 6-1, a + preceding the height indicates the features protrude from the surface, while a indicates the features are recessed (i.e. pits). The height or depth of these features is indicated by the [height] term, the f eature width is denoted by [width], and the spacing between the features is denoted by [spacing]. Th e surface designs for the Gradient surfaces are GR0 through GR5, and SK for the Sharklet AF design. All dimensions are given as m. The surface designations are shown in Table 6-1. Sample Preparation Images of settled Ulva linza spores on topographical surf aces were obtained during previous Ulva settlement assays. Sample preparation and Ulva settlement assays are described here briefly. The two-dimensional surface patterns were produced in a darkfield photomask. Silicon wafers 4 in diameter and of (100) orient ation were patterned w ith photolithography using Shipleys S1813 photoresist. Deep re active ion etching was used to etch the silicon wafers to the desired height, followed by an O2 plasma clean to remove the photoresist from the wafers. The 102

PAGE 103

mold for the Recessed Sharklet was created by solution casting Kraton styreneethylene/butylene-styrene (SEBS) block copolymer onto the sili con wafer Sharklet AF mold from toluene. Table 6-1. Surface designations for Gradient, Sharklet AF, and Recessed Sharklet AF surfaces Surface Designation GR0 +2.4GR0_2x2 GR1 +2.5GR1_2x2 GR2 +2.5GR2_2x2 GR3 +2.4GR3_2x2 GR4 +2.5GR4_2x2 GR5 +2.3GR5_2x2 Sharklet AF +2.5SK2x2 Recessed Sharklet AF -2.0SK2x2 Dow Corning SILASTIC T2 poly(dimethyl si loxane) elastomer (PDMSe) was prepared by mixing the resin and curing agent in a 10:1 ra tio for 5 minutes. The mixture was degassed for 30 minutes and cured against the sample mold s for 24 hours at approximately 22C. PDMSe sample surfaces were backed to glass microscope slides using allyltriethoxysilane as a coupling agent. Resulting samples consisted of a square patterned area in the cen ter of the slide with smooth PDMSe covering the remainder of the s lide. The patterned ar ea was 25 mm x 25 mm (Sharklet AF and Recessed Sharklet AF) or 15 mm x 15 mm (Gradient surfaces). Samples surfaces were rinsed with 95% ethanol (5 % distilled water) and blown with dry N2. Ulva Linza Assay Ulva linza zoospore assays were performed at the University of Birmingham, UK by Maureen Callow, James Callow, and John Finlay. Each surface was prev iously evaluated for Ulva linza spore settlement using standardized sp ore settlement protocols as described previously.102 Ulva linza plants were collected from Wembury Beach, UK (50 N, 4 W), and the zoospores were released as previously described. Samples were immersed in deionized 103

PAGE 104

water for 24 hours, followed by 2 hours in Inst ant Ocean artificial se a water prior to the Ulva assay. Each sample surface was placed into an assay dish, and 10 ml of zoospore solution (1.5 x 106 spores/ml) was added to each assay dish. Th e samples were incubated in the dark at approximately 20C for 45 minutes. The sample s were rinsed to remove any unattached zoospores and fixed with 2% glutaraldehyde in seawater for 10 minut es. The glutaraldehyde solution was rinsed off using a sequence of seaw ater, 50% seawater in distilled water, and distilled water prior to drying. Images were obtained using a Zeiss epifluorescence microscope as part of the Ulva linza zoospore assays. The images combined light from transmission through the sample and fluoresced light from the chlorophyll in the spores. The transmitted light allowed the topographical surfaces features to be distinguishable in the image, and the fluoresced light from the spores allowed the spores to be distinguishable in the image. Image Analysis The positions of the spores within the topograp hical pattern were determined in a multistep process. First, the symmetric elements of the patterns were determined in order to create an asymmetric unit that represents the pattern. Second, a vector notation was derived for each pattern, enabling multiple images to be combined. Third, the spore centroid locations were determined in the Cartesian coordinate system of the image and converted to the vector notation. Fourth, symmetry operations were performed on the spore centroid locations to bring each location into the asymmetric unit. Asymmetric units The symmetry of each surface design was determined using analysis of symmetry elements. The symmetry elements on the Sharklet AF surface are shown in Figure 6-2, along with the lattice vectors. The vector v1 is the direction perpendi cular to the features, and v2 is the 104

PAGE 105

direction parallel to the features. The plan e group for the Sharklet AF is c2mm. The asymmetric unit for each surface was determined from the symmetry elements. Multiple asymmetric units are possible for a given surface, as shown in Figure 6-3. All of the asymmetric units for a surface are functionally equivalent an d vary only in aesthetic s. The asymmetric unit shown in Figure 6-3C was chosen to represent the Sharklet AF surface. Mathematically, the unit is more complex to describe than the other two units, but Figure 6-3C presents the data in a more visually meaningful way. This asymmetric unit does not split individual features within the unit, and keeps the gap between adjacent diamonds continuous. Figure 6-2. Symmetry elements of the Sharklet AF surface and lattice vector notation. Ovals represent two-fold rotation; solid lines represent mirror lines; and dashed lines represent glide lines. The asymmetric units for the other surfaces we re determined in a similar manner as the Sharklet AF. These asymmetric units are s hown in Figure 6-4. The Recessed Sharklet AF has the same asymmetric unit as the Sharklet AF, since both surfaces have the same two dimensional pattern. 105

PAGE 106

Figure 6-3. Three asymmetric units for the Sharkl et AF, indicated by th e outlined shape. Areas of features within the asymmetric unit are colored with blue. Figure 6-4. Asymmetric units for evaluated surf aces: A) GR0, B) GR1, C) GR2, D) GR3, E) GR4, F) GR5, G) Sharklet AF and Recesse d Sharklet AF. Vectors are in terms of the lattice vectors. Vector notation The vector notation is based on the lattice vector s for the surface patterns (Figure 6-2). In order to convert the Cartesian coordinate system for each image to the vector coordinate system, the Cartesian coordinates of se veral selected points were identified. Image analysis was performed using ImageJ software.92 The vectors between latti ce points in both directions (perpendicular to the features and parallel to th e features) were calculated in the Cartesian space and compared to the lattice vector lengths. The la ttice points were chosen to be multiple lattice vector lengths apart to reduce error in determin ing the lattice vector lengths. Additionally, the process was performed in two separate locatio ns on the image to further reduce error though averaging. The ends of the largest features in the pattern were used because of the relative ease 106

PAGE 107

of determining the location of these points. The location of one lattice point in the Cartesian coordinates was found after determining one mo re point: the bottom of one of the largest features. The average of the bottom and top of this feature gives the center point of the feature and the location of the lattice point. These five points are shown in Figur e 6-5 as an example. The number of diamonds (i.e. vector lengths) between points varies among images and among surfaces tested. Figure 6-5. Representative Cart esian coordinates for vector transformation on an image of Ulva spores settled on the Sharklet AF surf ace. Inset shows higher-magnification view of locations A and B. The vectors v1 and v2 were calculated from Equations 62 and 6-3, respectively. The factor of 2 in the denominators indicates the average between two calculations. The calculations of v1 and v2 for all surfaces are calculated similarly. n DE n AC n DE n AC vyyyy xxxx22 221 (6-2) m CE m AD m CE m AD vyyyy xxxx22 222 (6-3) 107

PAGE 108

The location of the lattice point between A and B was calculated by averaging the x and y coordinates of the A and B points. Up to four lattice points on each image were determined using similar calculations. The lattice points were sel ected to minimize the distance between the spore locations and the nearest lattice point. Choosing multiple lattice points re duces the error in the final location. Spore centroid locations The locations of the spores on the image we re determined using ImageJ software. The images were converted to black and white, followed by applying a th reshold operation. The threshold operation removed the surface features from the image, leaving only the fluorescent spores in the image. An image that has undergone the threshold operation is shown in Figure 6-6. Particle analysis was then performed to de termine the centroids of the bright spores. Occasionally, a part of a topographical feature was bright enough to remain after the threshold operation and these spots were disregarded. The ImageJ particle analysis determined spore centroids in the Cartesian coordi nate system. In this process, the centroid of the region is determined by using the Cartesian coordinates of the pixels in each partic le and calculating the centroid from the pixel locations. Figure 6-6. Threshold operation of settlement im age: A) Settlement im age and B) Thresholded image. 108

PAGE 109

Only individual spores were included in the analysis. The analysis is focused on spores interacting with the topographi cal surface. Settled spores alter the local topography and produce chemical cues that attract other spores. In groups of spores, the first spore cannot be distinguished from spores that subsequently sett led. Only single spores could be guaranteed to have interacted with an area of the surface c ontaining no other spores, and thus were the only spores used in the analysis. Conversion of centroids to vector system The closest identified lattice point (up to four were identified in each image) was found for each of the spore centroids. Any lattice point would provide a registration point, but error in the final map increases as the length of the position vector increases. The vector in the Cartesian coordinate system between the lattice point and the centroid was determin ed and converted to a vector in the lattice vector space. For a spore of arbitrary label a w ith position a(x, y) = (ax, ay), the vector La between the nearest lattice point L(x, y) and a in vector space follows Equations 6-4 and 6-5. yyxxLaLayxLa ,), ( (6-4) 2 2 2 2 1 1 21, ),( v vyxLa v vyxLa vvLa (6-5) Translational symmetry was used to transform La( v1, v2) to the vector a( v1, v2) that describes the position of the spore relative to a lattice point within one unit cell: within (0, 0) and ( v1, v2). First, the mirror symmetry in both the v1 and v2 directions at each la ttice point was used in the calculation to make all of the vector components positive. Second, the fractional portion of the vector components (above multiples of v1 and v2) were taken to determine the single unit cell vector for the spore centroids. One unit cell cont ains eight asymmetric units for the GR1, GR2, 109

PAGE 110

GR3, GR4, Sharklet AF, and Recessed Shar klet AF designs. This area contains four asymmetric units for the GR0 and GR5 designs. The calculation is shown in Equation 6-6. 1,mod,1,mod),(2 121 v va a vva (6-6) The mirror symmetry within a unit cell was used to convert a ( v1, v2) into a vector a( v1, v2) which is the vector to the centroid within ( v1, v2) = (0, 0) to (0.5, 0.5). This area contains two asymmetric units for the GR1, GR2, GR3, GR4, Sharklet AF and Recessed Sharklet AF designs. The area contains one asymmetric un it for the GR0 and GR5 designs. The two-fold rotation at ( v1, v2) = (0.25, 0.25) for the GR1, GR2, GR 3, GR4, Sharklet AF, and Recessed Sharklet AF designs was used convert a( v1, v2) to a( v1, v2), which falls entirely within one asymmetric unit. These calculations are shown in Equations 6-7 and 6-8. The equation for the boundary of the asymmetric unit depends on the pattern and the asymmetric unit chosen. The process from image vector La to asymmetric vector a is shown in Figure 6-7. 2 15.0,5.0),(21 v vaa vva (6-7) 2 15.0,5.0),( :unit asymmetric outside is if ),(),( :unit asymmetric within is if21 21 21 v vaa vva a vvavva a (6-8) Figure 6-7. Process of transformation from lattice vectors A) La to B) a to C) a to D) a 110

PAGE 111

Settlement Maps and Smooth Histograms Using this vector system, spore locations from multiple images of th e same surface pattern were compiled onto a single asymmetric unit. The asymmetric units and th e spore positions were reflected and rotated to provide a presentation that is visually meaningful This duplication is represented for one exampl e location in Figure 6-8. The Ulva spore settlement maps were converted into smoothed histograms as a method of estimating the relative settlement probability distribution within a surface. The smoothed histogram was created by binning the spore ce ntroids into small bins (each bin 0.01 m2) according to position, followed by local smoot hing between histogram bins. Smoothing was performed with first and second order penalty smoothing, following the procedure outlined by Eilers and Goeman.107 Figure 6-8. Example of duplicating asymmetric unit and spore locations for presentation. 111

PAGE 112

Results and Discussion Each settled spore represents an individua l realization of the underlying settlement probability function, and each spore is therefore an experiment. Each map contained a total of between 44 and 71 individual mapped spores. Each spore centroid is represented by a single marker on the settlement map. The markers are approximately 0.1 to 0.2 m wide when drawn on the maps for the surfaces with 2 m wide feat ures, but represent the cen troids of spores that are approximately 5 m in diameter. The maps are not intended to de scribe how much of a region an individual spore covers; instead the maps are designed to show where the spores land. If the full 5 m width of the spore were included in the marker size, the markers would fill most of the map, and little useful information coul d be distinguished from the map. The spore settlement maps for the Gradient surfaces are shown in Figure 6-9. Figure 6-9. Spore settlement maps on Gradient surfaces. Each spore and asymmetric unit is represented 8 times (GR1, GR2, GR3, and GR4) or 16 times (GR0 and GR5). 112

PAGE 113

The spore settlement maps show se veral important aspects of the Ulva spore settlement on these surfaces. First, onl y two spores (out of 301) settled on th e protruding features. One of these spores was centered on the feature of the GR0 (+2.4GR0_2x2) surface. The other spore was at the very corner of the 12 m long feature of the GR4 (+2.5GR 4_2x2) surface. All other spores (>99%) settled centered on recessed regions. The spores were too larg e to fit between the features, but they settled bridging the features. The second important aspect shown by the maps is that most of the spores settled at the intersections of multiple features (Table 6-2). Both of these aspects were previously observed in previous reports, but here the preference is quantified. The spores on the GR0 (+2.4GR0_2x2) surface settled within a very tight area, with 96% settling w ithin the intersecti on of four adjacent features. Statistically, all of the preferential locations on the Grad ient surfaces are significantly different than if the spores were randomly distributed within the depressed regions. Table 6-2. Spore preference on Gradient su rfaces with 95% confidence intervals. Surface Spores in Intersections (%) S pores in Intersections/Total Spores +2.4GR0_2x2 96 7* 43/45 +2.5GR1_2x2 69 13* 35/51 +2.5GR2_2x2 76 12* 39/51 +2.4GR3_2x2 82 11* 36/44 +2.5GR4_2x2 56 13* 29/52 +2.3GR5_2x2 79 10* 46/58 Asterisks (*) denote proportion is statistically di fferent than expected if spores were evenly distributed in depressed regions ( = 0.05). The spore maps for the two surfaces with th e 4-element Sharklet AF pattern (Sharklet AF +2.5SK2x2 and Recessed Sharklet AF -2.0 SK2x2) are shown in Figure 6-10. These two surfaces have the same two-dimensional pattern, but differ in the nature of the features. The Sharklet AF (+2.5SK2x2) surface has features that protrude, while the Recessed Sharklet AF (-2.0SK2x2) surface has features that are recessed (i.e. pits). The Sharklet AF has two distinct intersections (labeled as the settlement sites 2 and 3 in Figure 6-10). The two sites 113

PAGE 114

denoted are the intersections of four distinct features (4, 8, 12, and 16 m lengths), while the site denoted is the intersecti on of two distinct feat ures (two 8 and two 12 m long features). Figure 6-10. Spore settlement maps for Shar klet AF and Recessed Sharklet AF surfaces. Each spore and asymmetric unit is represente d 8 times. Apparent preferential sites are circled and numbered. The spore settlement maps clearl y show that the spores settle within the depressed regions on both surfaces. On the +2.5SK2x2 surface, the s pores settle centered between the features, while on the -2.0SK2x2 surface, the spores settle centered on the recessed features. The spores are too large to fit within the depressed regions on both surfaces. On the Sharklet AF surface, 96% of the spores settled on the depressed regions, and on the Recessed Sharklet AF, 94% of the spores settled on the rece ssed regions. On the +2.5SK2x2 su rface, a large proportion of the spores settled within the preferential site 2, which is twice as large as the preferential sites 1 (due to the presence of two identical sites labeled 2). The probability density (% of spores in the s ite per the area the site encompasses) is a better descriptor of the probability a given spor e will settle in a specified location than the 114

PAGE 115

percentage of spores that settle within the locatio n. When all of the locations are of similar sizes and are present in equal numbers, the two quanti ties are simply scaled. However, on the Sharklet AF surface, the intersection sites (2 and 3) are tw ice as large as site 1. There are two identical sites labeled 2, and combined they encompass twice the area of site 3 (four times the area of site 1). The spore preference was normalized with resp ect to area of the site to compare within a surface to compare the probability densities. Qu antification of the spore preference on the Sharklet AF and Recessed Sharklet AF is show n in Table 6-3. The settlement probability densities cannot, in general, be compared among different surfaces because the total number of settled spores (including clusters) is not factored into the analysis. Table 6-3. Spore preference on Sharklet AF and Recessed Sharklet AF surfaces with 95% confidence intervals Recessed Sharklet AF Sharklet AF (+2.5SK2x2) (-2.0SK2x2) Probability Density Probability Density Percent Percent 2m spores % 2m spores % (%) (%) Spores in intersections 56 12* 4.6* 0 0 Spores in preferential site 1 11 7 5.5 14 8 7.0 Spores in preferential site 2 39 11* 4.9* 25 10* 6.3* Spores in preferential site 3 17 9 4.2 NA NA Asterisks (*) denote proportion is statistically di fferent than expected if spores were evenly distributed in depresse d or recessed regions ( = 0.05) In terms of density, the all of the identified preferential settlement sites are similar on the Sharklet AF surface. However, only the settlement sites labeled 2 are statistically significant compared to a random distribution of spores with in the depressed regions. The spore density on site 1 is larger than on site 2, but the total area encompassed is much smaller and fewer total spores landed in site 1 than in site 2. This di fference in total spores leads to the statistical conclusion that the preferential site 2 is significant, while site 1 cannot be declared significant. 115

PAGE 116

The Recessed Sharklet AF surface has two preferen tial settlement sites, but only site 2 is statistically significant. Site 1 on the Recessed Sh arklet AF is only half of what it appears in Figure 6-10: the circled region comprises Site 1 and a mirrored version of Site 1. Site 2 is completely within the asymmetric unit and is of the apparent size in Figure 6-10. Site 2 encompasses twice the total area of site 1, which can also be explained by the fact that there are two 8 m features (on which Site 2 resides) in the design, and only one 16 m feature (on which Site 1 resides). Both preferential sites on the Recessed Sharklet AF ha ve similar probability densities, suggesting Ulva spores are similarly attracted to both of these sites. Smoothed histograms are a method of estim ating probability densities, and convey additional information. The smoothed histograms of the Ulva spore settlement maps on the Gradient surfaces are shown in Figure 6-11. These histograms indicate that the spore settlement probabilities are highest at the center of the intersection of th ree or four features. The high density areas on the +2.4GR0_2x2 and +2.3GR5 _2x2 surfaces are nearly circular, and the density between the preferential site s on these two surfaces is very low. The smoothed histograms of the Sharklet AF and Recessed Sharklet AF are shown in Figure 6-12. Some interesting th ings are revealed in the smoot hed histograms that were not distinguishable from the settlement maps. First, there is a small region within preferential settlement Site 2 on the Sharklet AF (+2.5SK 2x2) surface that has a high density of spores. The density of spores in Site 2 is on average lower than in Site 1 (see Table 6-1), but the smaller area identified by the dark red colo r within Site 2 has a higher density than the peak of Site 1. In the spore settlement maps, all of Site 2 appeared to a similar spore density. Second, the smoothed histogram for the Sharklet AF(+2 .5SK2x2) shows that the spores prefer the inte rsection of the 4 m, 8 m and 16 m features (Site 2a) to th e intersection of the 8 m, 12 m and 16 m 116

PAGE 117

features (Site 2b). The spore densities on the Sharklet AF surface are easier to compare between the various sites when using the smoothe d histogram, with Site 3 having a lower density than Sites 1 and 2. Figure 6-11. Smoothed histograms of Ulva spore settlement maps on Gradient surfaces. Intensities are not comparable among surfaces. The Recessed Sharklet AF (-2.0SK2x2) smoothed histogram shows two preferential settlement sites, similar to what was apparent from the spore settlement maps. However, the preferential sites are easier to distinguish from the rest of the pattern on the smoothed histogram than on the settlement map. The s pore density at the two preferential sites is similar, with lower spore density along the other rece ssed regions. The two preferentia l settlement sites stand out more clearly on the smoothed histogram than on the Ulva spore settlement map. The flat regions that are not recessed show low spore density which was also easily observed on the spore settlement maps. 117

PAGE 118

Figure 6-12. Smoothed histograms of Ulva spore settlement maps on Sharklet AF (+2.5SK2x2) and Recessed Sharklet AF (-2.0SK2x2) surf aces. Intensities are not comparable among surfaces. The spore settlement maps and smoothed hist ograms do not describe the overall settlement numbers. Groups of spores were omitted from the analysis, and images were taken from different settlement assays. The Ulva spores settle in lower numbers on the Sharklet AF (+2.5SK2x2) surface than on the Gradient surfaces, though this information is not directly available from the settlement maps and histograms. Preferential settlement Site 1 on the Sh arklet AF (+2.5SK2x2) surface was not previously observationally iden tified as having higher settleme nt than the other depressed regions. This site was not sta tistically signifi cant due to the small si ze, and should be reevaluated using a larger number of settled spores. Th e relative preference among intersections of multiple features has not been previously observed. Analysis of the Ulva settlement maps and the smoothed histograms reveal differences among the intersections. The inte rsection labeled 2a on the Sharklet AF surface has a higher density of spores than the other intersections (2b and the two identical intersections in Site 3). Diffe rences among the densities were not compared 118

PAGE 119

statistically, and further work with a larger number of settled spores would be able to confirm or dispute the differences among intersections. The spore maps and histograms allow for the identification of preferential settlement sites. These sites can be used to help identify characteristics that the Ulva spores are detecting by comparing the characteristics at these sites to the characteristics at other locations. The mapping technique has the potential to be used with any regular obj ect on translationally symmetric patterns. Other fouling organisms, such as ba cteria or diatoms, c ould be mapped on these surfaces to examine preference of these organisms. This technique could be used in applications where certain cells are desired on a surface. Scaffo lds for tissue engineering, in which a specific cell type and morphology may be desired, could be aided by mapping positions of critical cell structure characteristics on pa tterns. The settlement mapping technique is not limited to topographical patterns; settlement preference on chemical patterns could be analyzed so long as the pattern can be distinguished in the images. Conclusions The process of mapping the locations of settled organisms (in this case Ulva linza spores) can aid in designing new surfaces and in investigating the role of topography in cell response. The settlement maps confirmed prev ious observational conclusions that Ulva spores settled preferentially at the intersection of several fe atures on the Gradient a nd Sharklet AF surfaces. A new potential preferential settlement site wa s identified on the Sharklet AF surface, though further investigation needs to be performed to confirm the significance of this site. The intersections of multiple feat ures on the Sharklet AF surface may not all have the same Ulva settlement density. Differences among the in tersections were observed in the smoothed histograms, but statistical significance was not determined. The identification of preferential settlement sites by mapping the settlement has the potential to drive the design of new surfaces 119

PAGE 120

that reduce or eliminate locations that are attractive to spores or other fouling organisms. Various other cells, including those desired in tissue engineering, and non-liv ing particles can be investigated using this technique Investigation into the properties of preferential settlement sites can aid in identifying localized cell-surface response and local particle-surface response to chemical, topographical, or other surface patterns. 120

PAGE 121

CHAPTER 7 CONCLUSIONS AND FUTURE WORK Conclusions Non-toxic surfaces that resist fouling by vari ous organisms are desired in many fields and industries including medicine, wate r treatment, and shipping. The use of engineered topographies has shown great promise for directing cell behavior and controlling fouling. Several models and analysis methods have been developed for the be havior of certain marine foulers. Engineered topographical surfaces were designed to evaluate and refine these models. The anisotropy in the static and dynamic c ontact angles on engin eered topographical surfaces was examined. By breaking channels into smaller lengths, such as the features found in some anti-fouling surfaces, the anisotropy is gr eatly reduced. The anisot ropy is not completely eliminated, and should be considered when performing contact angle analysis on surfaces with directional dependence. The receding contact angle was found to control the angle of tilt required to cause a liquid drop to slip off of the surfaces tested. Topography effectively alters the static and dynamic contact angles, and has the potential to create directiona lly dependent contact angles to be used in wallless microfluidic devices. A biosettlement model that shows predictive capability for Ulva linza spore settlement on engineered topographies was refined from previ ous work. The solid area fraction of the surface and the number of distinct featur es were shown to be important factors in the settlement of Ulva spores. The biosettlement mode l was shown to predict the Ulva settlement of three previouslyuntested engineered topographies. The surfaces inco rporated into the biosettlement model all had spacing between features of 2 m and were composed of poly(d imethyl siloxane) elastomer. This predictive model may be useful for other organisms, and will be useful in designing new antifouling surfaces. 121

PAGE 122

The response of a species of diatoms to engineered topographies was investigated. Attachment of diatoms was reduced on surfaces of the same size scale th at is effective for reducing Ulva settlement. The diatoms, which are not motile when approaching a surface, did not follow the biosettlement mode l. The Sharklet AF surface, which was previously shown to reduce Ulva linza settlement by approximately 75%, was shown to reduce the attachment of Navicula perminuta diatoms by approximately 35%. A technique to map the positions of Ulva spores and other ce lls on translationally symmetric topographies such as the Sharklet AF was developed. The technique was used to identify new preferential settlement sites and confirm previously observe d preferential sites. Smoothed histograms were used to show areas of higher spore density within preferential settlement sites. Spores tended to settle within recessed or depressed regi ons, specifically within intersections of multiple features. An additional pr eferential settlement site on the Sharklet AF was identified at the center of the region between the 16 m and 12 m long features. The Recessed Sharklet AF surface showed two prefer ential settlement site s within the recessed features. The mapping technique is an analysis method that may aid in the design of new antifouling surfaces and the predic tion of antifouling performance. The work performed pushes the boundaries of the scientific knowledge in biofouling and in surface characterization using contact angl e goniometry. Effective non-toxic antifouling coatings are greatly needed, and this work aids in the progress toward these coatings. The models and techniques developed for biofouling may be useful for developing an tifouling coatings in other fields and control ling desired cellular response for medi cal devices. Additional work can be used to further investigate the models and t echniques and add to the knowledge of cellular response to engineered topographies. 122

PAGE 123

Future Work Feature Dimensions The biosettlement model for Ulva linza spore settlement is based on surfaces with 2 m feature widths and spacing of 2 m. Channels at least 5 m wide have been shown to increase spore settlement, and surfaces with 2 m spacing reduce settlement. Surfaces based on dimensions between 2 m and 5 m, and surfaces based on dimensions less than 2 m have not been extensively evaluated. There are several ap proaches to investigating feature dimensions within the gaps in knowledge for fouling response. This work w ould allow for the inclusion of the dimensional scale of the topogra phy into the biosettlement model. The first method to investigate additional feat ure dimensions is through the production of new photomasks. Photomasks with pa tterns of different dimensions and process silicon wafers as performed in many of the studies described. Th e Sharklet AF could be produced in different dimensions such as 3 m (SK3x3), 4 m (SK4x4), and 5 m (SK5x5). Photolithography may be able to produce a surface focused on 1 m (S K1x1), though experience ha s shown that as the dimensions of the pattern decrease, errors in the process become more pronounced. The SK5x5 surface would be expected to increase the settlement of Ulva spores versus a smooth surface, since the spores would be able to penetrate the recessed regions similar to 5 m wide channels. A second method to investigate additi onal feature dimensi ons is through the overexposure of photoresist during the photolith ography process. As the photoresist is overexposed, the photoresist degrades in the area s outside of the desi gned exposure area. The result of the overexposure is cas t surfaces with features that are wider and spacing between the features that is smaller than designed on the photomask. Several Sharklet AF surfaces were created in this way with feature spacing va rying from 1.1 m to 2.0 m (Figure 7-1). 123

PAGE 124

Figure 7-1. Sharklet AF surfaces overexposed to shri nk spacing between features. A third method to change the dimension scal e is to use other lithography processes. Electron-beam lithography is a method which ca n perform lithography to a very small length scale. Molds for Sharklet AF surfaces have b een made on very small areas (100 m x 100 m) down to feature widths of 200 nm. After perfor ming ebeam lithography, patterned photoresist is well defined down to 200 nm feature width (Figure 7-2). Figure 7-2. Sharklet AF patte rned ebeam resist using ebea m lithography. Feature width and spacing are A) 200 nm, B) 500 nm, and C) 1 m. D) is magnified view of surface with 200 nm feature width. Even after etching with deep reactive ion et ching, the silicon wafer with 200 nm feature width shows good fidelity in local areas (Figure 7-3). In some areas, the pattern was not transferred during etching. The l ack of pattern transfer could ha ve been due to overexposure of the ebeam resist or too thin a layer of ebeam re sist. Either of these cases would not sufficiently protect the silicon from etchi ng. More likely, the lack of pattern transfer was due to 124

PAGE 125

underexposure of the ebeam resist or too thick a layer of ebeam resist. If the resist was not fully developed the full depth to the si licon wafer, even a thin layer would shield the silicon wafer from being etched in areas in which etching wa s desired. These new designs, if produced over larger areas, would add to the available dime nsions of the Sharklet AF for probing cell response to topographies. Figure 7-3. Etched silicon wafer molds for Shar klet AF surfaces with 200 nm feature width and spacing. Waviness occurr ed during image capture. New Sharklet AF Designs Several new Sharklet AF designs could be cr eated to evaluate the order of the pattern and whether eliminating preferential sites will reduce settlement. A random Sharklet AF surface, in which the features of the Sharkl et AF are rearranged in pseudo-random manner (Figure 7-4A) would be useful for testing the arrangement of the Sharklet AF. This surface would have all of the same calculated biosettlement model characteristics ( n r, s) and same designed dimensions (height, widt h, and spacing) as the Sharklet AF. The arrangement of the features would be different between the Sharklet AF and Random Sharklet AF surfaces. Cell locations could not be mapped on the Rand om Sharklet AF in the same manner as on translationally symmetric designs (e.g. Sharklet AF. However, analysis of cell locations relative to the loca l features would provide comparison among a larger number of types of intersections between features. 125

PAGE 126

Other surfaces could be created which redu ce or eliminate the observed preferential settlement sites. Surfaces such as those in Figur e 7-4B and 7-4C attempt to prevent spores from settling at intersections between features. Th ese arrangements would be expected to reduce settlement by preventing pref erential settlement sites. Figure 7-4. New Sharkl et AF surfaces. A) Random Sharklet AF, B) Sharklet AF with pillars between rows, and C) Sharklet AF with pillars in intersections. New Sharklet AF designs with larger numbers of distinct features (e.g. 6, 7, etc) would extend the ability to probe the biosettlement model. New combinations of features with the same number of distinct featuressuch as a series of 3, 6, 9, 12, 15 m long featureswould complement the current designs. Settlement Mapping Additional surfaces and additional cells could be mapped with the mapping technique. In addition to the approach outlined above, in wh ich preferential locations are removed from current surface designs, an approach to design new surfaces based on local structures could be implemented. This approach would require exte nsive testing to identify settlement preference among various types of local conditions. A libra ry of settlement preference for localized topography could be created, and surfaces c ould be designed by compiling low-preference building blocks. 126

PAGE 127

The mapping process currently maps only the firs t cell to settle in a location by eliminating any groups of cells. However, Ulva linza spores tend to settle wher e other spores have already established. Mapping studies could be performed in which a single field of view is observed in a time-lapse study. This method would allow not only the first spores to settle to be mapped, but also to compare if there are settlement sites that tend to attract a second sp ore, third spore, etc. Summary This future work would further develop the bi osettlement model, exam ine the effectiveness of the mapping technique for developing new su rfaces, and improve the scope of biofouling knowledge. This proposed work would progress the field toward developi ng effective long-term non-toxic coatings that prevent f ouling. The knowledge gained from this future work may be a crucial step in the next generation of effective antifouling coatings. 127

PAGE 128

LIST OF REFERENCES 1. Wahl, M., In Fouling Organisms of the Indian O cean: Biology and Control Technology Nagabhushanam, R.; Thompson, M.-F., Eds. A.A. Balkema Publishers: Rotterdam, Netherlands, 1997; pp 33-84. 2. Yebra, D. M.; Kiil, S.; Dam-Johansen, K., Prog. Org. Coat. 2004, 50, 75-104. 3. Evans, S. M., Biofouling 1999, 14, 117-129. 4. Champ, M. A., Sci. Total Environ. 2000, 258, 21-71. 5. International Maritime Organization. Adoption of the Final Ac t of the Conference and Any Instruments, Recommendations and Resolu tions Resulting from the Work of the Conference International Convention on the Control of Harmful Anti-Fouling Systems on Ships, 2001. 6. Sousa, A.; Genio, L.; Mendo, S.; Barrosoi, C., Appl. Organomet. Chem. 2005, 19, 324328. 7. Callow, M. E.; Jennings, A. R.; Brennan, A. B.; Seegert, C. A.; Gibson, A.; Wilson, L. H.; Feinberg, A. W.; Baney, R.; Callow, J. A., Biofouling 2002, 18, 237-245. 8. Scardino, A. J.; Harvey, E.; De Nys, R., Biofouling 2006, 22, 55-60. 9. Schumacher, J. F.; Aldred, N.; Callow, M. E.; Finlay, J. A.; Callow, J. A.; Clare, A. S.; Brennan, A. B., Biofouling 2007, 23, 307-317. 10. Warren, D. K.; Quadir, W. W.; Hollenbeak, C. S.; Elward, A. M.; Cox, M. J.; Fraser, V. J., Crit. Care Med. 2006, 34, 2084-2089. 11. Liedl, B., Curr. Opin. Urol. 2001, 11 75-79. 12. Teughels, W.; Van Assche, N.; Sliepen, I.; Quirynen, M., Clin Oral Implan Res 2006, 17, 68-81. 13. Cloete, T. E.; Brozel, V. S.; Vonholy, A., Int. Biodeterior. Biodegrad. 1992, 29, 299-341. 14. Townsin, R. L., Biofouling 2003, 19, 9-15. 15. Schultz, M. P., Biofouling 2007, 23, 331-341. 16. Callow, M. E., In Biofilms: Recent Advances in Their Study and Control Evans, L. V., Ed. Harwood Academic Publishers: Amsterdam, Neth., 2000; pp 189-209. 17. Alzieu, C., Marine Environ. Res. 1991, 32, 7-17. 18. Gibbs, P.; Bryan, G., TBT Paints and the Demise of the Dog-Whelk, Nucella Lapillus (Gastropoda). In Oceans 1987; Vol. 19, pp 1482-1487. 128

PAGE 129

19. Antizar-Ladislao, B., Environ. Int. 2008, 34, 292-308. 20. Appel, K. E., Drug Metab. Rev. 2004, 36, 763-786. 21. Omae, M., Chem. Rev. 2003, 103, 3431-3448. 22. Srinivasan, M.; Swain, G. W., Environ. Manage. 2007, 39, 423-441. 23. Bellas, J.; Granmo, A.; Beiras, R., Mar. Pollut. Bull. 2005, 50, 1382-1385. 24. Munari, C.; Mistri, M., Mar. Environ. Res. 2007, 64, 347-357. 25. Paradas, W. C.; Filho, G. A. A., Braz. J. Oceanogr. 2007, 51-56. 26. Bao, V. W. W.; Leung, K. M. Y.; Kwok, K. W. H.; Zhang, A. Q.; Lui, G. C. S., Mar. Pollut. Bull. 2008, 616-623. 27. Karlsson, J.; Eklund, B., Mar. Pollut. Bull. 2004, 49, 456-464. 28. Showalter, S.; Savarese, J. Restrictions on the Use of Marine Antifouling Paints Containing Tributyltin and Copper ; MASGP 04-052; Sea Grant Law Center: October 2004 (Updated August 2005), 2004. 29. Voulvoulis, N.; Scrimshaw, M. D.; Lester, J. N., Appl. Organometal. Chem. 1999, 13, 135-143. 30. Chambers, L. D.; Stokes, K. R.; Walsh, F. C.; Wood, R. J. K., Surf. Coat. Tech. 2006, 201, 3642-3652. 31. Holland, R.; Dugdale, T. M.; Wetherbee, R.; Bre nnan, A. B.; Finlay, J. A.; Callow, J. A.; Callow, M. E., Biofouling 2004, 20, 323-329. 32. Finlay, J. A.; Callow, M. E.; Ista, L. K.; Lopez, G. P.; Callow, J. A., Integr. Comp. Biol. 2002, 42, 1116-1122. 33. Krishnan, S.; Ayothi, R.; Hexemer, A.; Finlay, J. A.; Sohn, K. E.; Perry, R.; Ober, C. K.; Kramer, E. J.; Callow, M. E.; Callow, J. A.; Fischer, D. A., Langmuir 2006, 22, 50755086. 34. Martinelli, E.; Agostini, S.; Galli, G.; Chiellini, E.; Glisenti, A.; Pettitt, M. E.; Callow, M. E.; Callow, J. A.; Graf, K.; Bartels, F. W., Langmuir 2008, 24, 13138-13147. 35. Ista, L. K.; Perez-Luna, V. H.; Lopez, G. P., Appl. Environ. Microbiol. 1999, 65, 16031609. 36. Majumdar, P.; Lee, E.; Patel, N.; Stafslien, S. J.; Daniels, J.; Chisholm, B. J., J. Coat. Technol. Res. 2008, 405-417. 129

PAGE 130

37. McMaster, D.; Bennett, S.; Tang, Y.; Finlay, J.; Kowalke, G.; Nedved, B.; Bright, F.; Callow, M.; Callow, J.; Wendt, D.; Hadfield, M.; Detty, M., Biofouling 2009, 25, 21-33. 38. Ista, L. K.; Callow, M. E.; Finlay, J. A.; Coleman, S. E.; Nolasco, A. C.; Simons, R. H.; Callow, J. A.; Lopez, G. P., Appl. Environ. Microbiol. 2004, 70 4151-4157. 39. Krishnan, S.; Weinman, C. J.; Ober, C. K., J. Mater. Chem. 2008, 18, 3405-3413. 40. Dobretsov, S.; Dahms, H. U.; Qian, P. Y., Biofouling 2006, 22, 43-54. 41. Holmstrm, C.; Egan, S.; Franks, A.; McCloy, S.; Kjelleberg, S., FEMS Microbiol. Ecol. 2002, 41, 47-58. 42. Yee, L. H.; Holmstrom, C.; Fuary, E. T.; Le win, N. C.; Kjelleberg, S.; Steinberg, P. D., Biofouling 2007, 23, 287-294. 43. Bers, A. V.; Wahl, M., Biofouling 2004, 20, 43-51. 44. Scardino, A. J.; de Nys, R., Biofouling 2004, 20, 249-257. 45. Scardino, A. J.; Hudleston, D.; Peng, Z.; Paul, N. A.; de Nys, R., Biofouling 2009, 25, 8393. 46. Carman, M. L.; Estes, T. G.; Feinberg, A. W.; Schumacher, J. F.; Wilkerson, W.; Wilson, L. H.; Callow, M. E.; Callow, J. A.; Brennan, A. B., Biofouling 2006, 22, 11-21. 47. Schumacher, J. F.; Carman, M. L.; Estes, T. G.; Feinberg, A. W.; Wilson, L. H.; Callow, M. E.; Callow, J. A.; Finlay, J. A.; Brennan, A. B., Biofouling 2007, 23, 55-62. 48. Schumacher, J. F.; Long, C. J.; Callow, M. E. ; Finlay, J. A.; Callow, J. A.; Brennan, A. B., Langmuir 2008, 24, 4931-4937. 49. Scardino, A. J.; Guenther, J.; de Nys, R., Biofouling 2008, 24, 45-53. 50. Petronis, S.; Berntsson, K. ; Gold, J.; Gatenholm, P., J. Biomat. Sci-Polym. E. 2000, 11, 1051-1072. 51. Andersson, M.; Berntsson, K.; Jonsson, P.; Gatenholm, P., Biofouling 1999, 14, 167-178. 52. Berntsson, K. M.; Jonsson, P. R.; Lejhall, M.; Gatenholm, P., J. Exp. Mar. Biol. Ecol. 2000, 251, 59-83. 53. Efimenko, K.; Finlay, J.; Callow, M. E.; Callow, J. A.; Genzer, J., ACS Appl. Mater. Interfaces 2009, 1 1031-1040. 54. Hoipkemeier-Wilson, L.; Schumacher, J.; Carman, M.; Gibson, A.; Feinberg, A.; Callow, M.; Finlay, J.; Callow, J.; Brennan, A., Biofouling 2004, 20, 53-63. 130

PAGE 131

55. Granhag, L. M.; Finlay, J. A.; Jonsson, P. R.; Callow, J. A.; Callow, M. E., Biofouling 2004, 20, 117-122. 56. Kohler, J.; Hansen, P. D.; Wahl, M., Biofouling 1999, 14, 237-248. 57. Diaz, C.; Cortizo, M. C.; Schilardi, P. L.; Saravia, S. G. G. d.; Mele, M. A. F. L. d., Mat. Res. 2007, 10, 11-14. 58. Medilanski, E.; Kaufmann, K.; Wick L. Y.; Wanner, O.; Harms, H., Biofouling 2002, 18, 193-203. 59. Finlay, J. A.; Krishnan, S.; Ca llow, M. E.; Callow, J. A.; Dong, R.; Asgill, N.; Wong, K.; Kramer, E. J.; Ober, C. K., Langmuir 2008, 24, 503-510. 60. Baier, R. E., In Adsorption of Microorganisms to Surfaces Bitton, G.; Marshall, K. C., Eds. John Wiley and Sons: New York, 1980; pp 59-104. 61. Callow, J. A.; Callow, M. E.; Ista, L. K.; Lopez, G.; Chaudhury, M. K., J. R. Soc. Interface 2005, 2 319-325. 62. Chen, Y.; He, B.; Lee, J. H.; Patankar, N. A., J. Colloid Interf. Sci. 2005, 281, 458-464. 63. Nakae, H.; Inui, R.; Hirata, Y.; Saito, H., Acta Mater. 1998, 46 2313-2318. 64. Zhao, Y.; Lu, Q. H.; Li, M.; Li, X., Langmuir 2007, 23, 6212-6217. 65. Bico, J.; Marzolin, C.; Quere, D., Europhys. Lett. 1999, 47, 220-226. 66. Sommers, A. D.; Jacobi, A. M., J. Micromech. Microeng. 2006, 16, 1571-1578. 67. Oner, D.; McCarthy, T. J., Langmuir 2000, 7777-7782. 68. Wenzel, R. N., Ind. Eng. Chem. 1936, 28, 988-994. 69. Cassie, B.; Baxter, S., T. Faraday Soc. 1944, 40, 546-551. 70. Bico, J.; Thiele, U.; Quere, D., Colloid Surf. A-Physicochem. Eng. Asp. 2002, 206, 41-46. 71. Yoshimitsu, Z.; Nakajima, A.; Watanabe, T.; Hashimoto, K., Langmuir 2002, 18, 58185822. 72. Extrand, C. W., Langmuir 2002, 18, 7991-7999. 73. Furstner, R.; Barthlott, W. ; Neinhuis, C.; Walzel, P., Langmuir 2005, 21 956-961. 74. Bico, J.; Tordeux, C.; Quere, D., Europhys. Lett. 2001, 55, 214-220. 75. Martines, E.; Seunarine, K.; Morgan, H.; Gadegaard, N.; Wilkinson, C. D. W.; Riehle, M. O., Nano Lett. 2005, 5 2097-2103. 131

PAGE 132

76. Burton, Z.; Bhushan, B., Nano Lett. 2005, 5 1607-1613. 77. Suh, K. Y.; Jon, S., Langmuir 2005, 21, 6836-6841. 78. He, B.; Patankar, N. A.; Lee, J., Langmuir 2003, 19, 4999-5003. 79. Abdelsalam, M. E.; Bartlett, P. N.; Kelf, T.; Baumberg, J., Langmuir 2005, 21, 17531757. 80. Gao, L.; McCarthy, T. J., Langmuir 2007. 81. McHale, G., Langmuir 2007, 23, 8200-8205. 82. Nosonovsky, M., Langmuir 2007, 23 9919-9920. 83. Marmur, A.; Bittoun, E., Langmuir 2009, 25, 1277-1281. 84. Miwa, M.; Nakajima, A.; Fujishima, A.; Hashimoto, K.; Watanabe, T., Langmuir 2000, 16, 5754-5760. 85. Wolansky, G.; Marmur, A., Colloid Surf. A-Physicochem. Eng. Asp. 1999, 156, 381-388. 86. Feng, L.; Li, S. H.; Li, Y. S.; Li, H. J.; Zhang, L. J.; Zhai, J.; Song, Y. L.; Liu, B. Q.; Jiang, L.; Zhu, D. B., Adv. Mater. 2002, 14, 1857-1860. 87. Bormashenko, E.; Stein, T.; Whyman, G.; Bormashenko, Y.; Pogreb, R., Langmuir 2006, 22, 9982-9985. 88. Zheng, Y. M.; Gao, X. F.; Jiang, L., Soft Matter 2007, 3 178-182. 89. Morita, M.; Koga, T.; Otsuka, H.; Takahara, A., Langmuir 2005, 21, 911-918. 90. Drelich, J.; Wilbur, J. L.; Miller, J. D.; Whitesides, G. M., Langmuir 1996, 12, 19131922. 91. Chaudhury, M. K.; Whitesides, G. M., Science 1992, 256, 1539-1541. 92. Rasaband, W. S. ImageJ U.S. National Institutes of Health: Bethesda, Maryland, 19972007. 93. Furmidge, C. G., J. Coll. Sci. Imp. U. Tok. 1962, 17, 309-324. 94. Mougin, K.; Castelein, G.; Haidara, H., Tribol. Lett. 2004, 17 11-17. 95. Aldred, N.; Clare, A. S., Biofouling 2008, 24, 351-363. 96. Callow, J. A.; Callow, M. E., In Biological Adhesives Smith, A. M.; Callow, J. A., Eds. Springer-Verlag: Berlin, 2006; pp 63-78. 132

PAGE 133

97. Molino, P. J.; Wetherbee, R., Biofouling 2008, 24, 365-379. 98. Callow, M. E.; Callow, J. A.; Ista, L. K.; Coleman, S. E.; Nolasco, A. C.; Lopez, G. P., Appl. Environ. Microbiol. 2000, 66, 3249-3254. 99. Schilp, S.; Kueller, A.; Rosenhahn, A.; Grunze, M.; Pettitt, M. E.; Callow, M. E.; Callow, J. A., Biointerphases 2007, 2 143-150. 100. Ederth, T.; Nygren, P.; Pettitt, M. E.; Ostblom, M.; Du, C. X.; Broo, K.; Callow, M. E.; Callow, J.; Liedberg, B., Biofouling 2008, 24, 303-312. 101. Long, C. J.; Schumacher, J. F.; Brennan, A. B., 2009, Accepted 102. Callow, M. E.; Callow, J. A.; Pick ett-Heaps, J. D.; Wetherbee, R., J. Phycol. 1997, 33, 938-947. 103. Wetherbee, R.; Lind, J. L.; Burke, J.; Quatrano, R. S., 1998, 34, 9-15. 104. Lind, J. L.; Heimann, K.; Miller, E. A.; va nVliet, C.; Hoogenraad, N. J.; Wetherbee, R., Planta 1997, 203, 213-221. 105. Fraley, C.; Raftery, A. E., J Am Stat Assoc 2002, 97, 611-631. 106. Vieira, V.; Webster, T.; Weinberg, J.; Aschengrau, A., Environ. Health 2009, 8 13. 107. Eilers, P. H. C.; Goeman, J. J., Bioinformatics 2004, 20, 623-628. 133

PAGE 134

BIOGRAPHICAL SKETCH Christopher James Long, son of R. Allen and Karen Long, was born in Salt Lake City, Utah. Christopher has one brother, Scott, and a sister, Jennifer. When Christopher was five years old, the Long family moved to Huntsville, Alabama, where he attended Roger B. Chaffee Elementary School, Weatherly Elementary Scho ol, Mountain Gap Middle School, and Virgil I. Grissom High School, where he met his future wife, Vict oria Lynn Salazar. Upon graduating high school, Christopher attended Auburn University where he earned a dual major in materials engineering and physics in 2004. While working towards his undergraduate degree, Christopher worked for Dr. Jeffrey Fergus as an undergraduate research assistant and spent two summers at the Kennedy Sp ace Center in the Failure Analysis Laboratory working for Mr. Donald Parker as part of the National Aeronautics and Space Administration Undergraduate Summer Research Program. Christopher received the Alumni Fellowship Award to study at th e University of Florida in the Materials Science and Engineering Departme nt. Christopher performed his graduate work under the advisement of Professor Anthony Brennan and specialized in biomaterials. He received his masters degree in 2007 and his doctorate in 2009.