<%BANNER%>

Ulva Linza Zoospore Sensitivity to Systematic Variation of Surface Topography

Permanent Link: http://ufdc.ufl.edu/UFE0042438/00001

Material Information

Title: Ulva Linza Zoospore Sensitivity to Systematic Variation of Surface Topography
Physical Description: 1 online resource (134 p.)
Language: english
Creator: SHEATS,JULIAN TAYLOR
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ANTIFOULING -- ERI -- ROUGHNESS -- SPATIAL -- TOPOGRAPHY -- ULVA
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The use of surface topographical microstructure is abundant in nature. The lotus plant uses a fractal-like topography to create a highly non-wetting surface that self-cleans as water drops take dirt particles with them as they roll off. Analysis of how topography affects surface interactions offers a unique opportunity to attack a problem that affects our economy and societal health significantly. The attachment of biological material to manmade surfaces can be looked at as fouling or directed adhesion. Marine fouling on ship hulls costs the United States $600 million each year due to increased fuel usage caused by drag. Hospital-acquired methicillin-resistant Staphylococcus aureus infections cause thousands of deaths annually as a result of colonization of hospital surfaces. The lack of biocompatible synthetic surfaces for implants such as vascular grafts lead to restenosis as cells are unable to develop a natural interaction with the graft surface. In each circumstance there is much to learn about the complicated attachment process. This work expands the investigation of the role of topography in the attachment of the green fouling algae Ulva linza to poly(dimethylsiloxane) surfaces. Spore attachment density was correlated to the Wenzel roughness ratio on low surface energy, high-modulus poly(dimethylsiloxane)-grafted-silicon topographies. The role of topography on a scale less than the size of a spore was investigated on nano-roughened poly(dimethylsiloxane) elastomer surfaces. For a specific group of patterns, the spatial distribution of spores attached to topographies was quantitatively analyzed and shown to correlate with feature dimensions.
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 JULIAN TAYLOR SHEATS.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Brennan, Anthony B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-10-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042438:00001

Permanent Link: http://ufdc.ufl.edu/UFE0042438/00001

Material Information

Title: Ulva Linza Zoospore Sensitivity to Systematic Variation of Surface Topography
Physical Description: 1 online resource (134 p.)
Language: english
Creator: SHEATS,JULIAN TAYLOR
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2011

Subjects

Subjects / Keywords: ANTIFOULING -- ERI -- ROUGHNESS -- SPATIAL -- TOPOGRAPHY -- ULVA
Materials Science and Engineering -- Dissertations, Academic -- UF
Genre: Materials Science and Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: The use of surface topographical microstructure is abundant in nature. The lotus plant uses a fractal-like topography to create a highly non-wetting surface that self-cleans as water drops take dirt particles with them as they roll off. Analysis of how topography affects surface interactions offers a unique opportunity to attack a problem that affects our economy and societal health significantly. The attachment of biological material to manmade surfaces can be looked at as fouling or directed adhesion. Marine fouling on ship hulls costs the United States $600 million each year due to increased fuel usage caused by drag. Hospital-acquired methicillin-resistant Staphylococcus aureus infections cause thousands of deaths annually as a result of colonization of hospital surfaces. The lack of biocompatible synthetic surfaces for implants such as vascular grafts lead to restenosis as cells are unable to develop a natural interaction with the graft surface. In each circumstance there is much to learn about the complicated attachment process. This work expands the investigation of the role of topography in the attachment of the green fouling algae Ulva linza to poly(dimethylsiloxane) surfaces. Spore attachment density was correlated to the Wenzel roughness ratio on low surface energy, high-modulus poly(dimethylsiloxane)-grafted-silicon topographies. The role of topography on a scale less than the size of a spore was investigated on nano-roughened poly(dimethylsiloxane) elastomer surfaces. For a specific group of patterns, the spatial distribution of spores attached to topographies was quantitatively analyzed and shown to correlate with feature dimensions.
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 JULIAN TAYLOR SHEATS.
Thesis: Thesis (Ph.D.)--University of Florida, 2011.
Local: Adviser: Brennan, Anthony B.
Electronic Access: RESTRICTED TO UF STUDENTS, STAFF, FACULTY, AND ON-CAMPUS USE UNTIL 2011-10-31

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2011
System ID: UFE0042438:00001


This item has the following downloads:


Full Text

PAGE 1

1 ULVA LINZA ZOOSPORE SENSITIVITY TO SYSTEMATIC VARIATION OF SURFACE TOPOGRAPHY By JULIAN TAYLOR SHEATS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 201 1

PAGE 2

2 201 1 Julian Taylor Sheats

PAGE 3

3 To my loving Kim, supportive family and friends past, present and yet to come.

PAGE 4

4 ACKNOWLEDGMENTS I offer my sincere thanks to the people who have made this period of academic and personal growth for me possible. My advisor, Dr. Anthony Brennan has been a constant source of inspiration and guidance throughout this process. I would lik e to thank my committee members Dr. Christopher Batich, Dr. Tanmay Lele, Dr. Scott Perry, Dr. Susan Sinnott, and Dr. Gregory Sawyer for their counsel at each milestone on the path to my graduation. I gratefully thank Jennifer Wrighton for steadfast and th orough assistance to me throughout the years and for good times that have kept me relaxed through it all. I appreciate the wealth of knowledge I gained from Al Ogden, Bill Lewis, and David Hays at the UF Nanoscale Research Facility and thank them for it. It has been a great privilege to work with the many fellow students, at every different level, during my time. Both as good friends and as role models, they have taught me and been there as pillars of support. In turn, I have found particular joy from th e times I have been able to pass on my own knowledge and experiences or offer help. I have looked to Dr. Cliff Wilson, Dr, Leslie Hoipkemeier Wilson, and Dr. Christopher Long as both friends and mentors at different stages in my education and offer my tha nks for their guidance and friendship. I thank members past and present of my research group for assistance, concern, and friendship: D r. David Jackson, Dr. James Schumacher, Dr. Michelle Carman, Mr. Kenneth Chung, Mr. Jim Seliga, Dr Chelsea Magin, Mr. Scott Cooper, Mr. Jack Chen, Ms. Angel Ejiasi, Dr. Kevin Jin, Mr. Nick Sexson, Ms. Kim Struk, Ms. Adwoa Baah Dwomoh, and Mr. Joe Decker. Sean Royston is thanked greatly for his steadfast and diligent work and for his friendship. This work could not have been possible without the wonderful collaboration of Dr. Maure en Callow, Dr. Jim Callow, and M r. John Finlay.

PAGE 5

5 apparent ev ery day as new experiences show its lasting effe cts. I thank my parents, Henry and Lisa, for working hard to make sure my upbringing was both nurturing and filled with a drive to exceed their expectations. My grandparents, Hilda, George and the late Terese have always supported me and will always be in my heart. I especially thank my aunt Michele for her unwavering aid through all circumstances. My brother Remi, sister Jade, and cousin Noah, are constantly in my thoughts, and I will always be there for t hem. To Kim, the love of my life, I give thanks every day for having her in my life, and I sincerel y appreciate and cherish the support she has provided me.

PAGE 6

6 TABLE OF CONTENTS page ACKNOWLEDGMENTS ................................ ................................ ................................ ........... 4 LIST OF TABLES ................................ ................................ ................................ ...................... 9 LIST OF FIGURES ................................ ................................ ................................ .................. 10 LIST OF ABBREVIATIONS ................................ ................................ ................................ .... 14 ABSTRACT ................................ ................................ ................................ ............................. 15 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ............. 17 Scope of Research ................................ ................................ ................................ .............. 17 Specific Aims ................................ ................................ ................................ ..................... 19 Specific aim 1: Investigate Ulva linza Zoospore Attachment Response to High Modulus Engineered Topographies with High Aspect Ratios ................................ ... 19 Specific Aim 2: Investigate Ulva linza Zoospore Attachment to Combined Micro And Nano Scale Roughness ................................ ................................ ..................... 19 Specific Aim 3: Analyze Spatial Distribution of Ulva linza Zoospore Attachment to Engineered Topographies ................................ ................................ ......................... 20 2 BACKGROUND ................................ ................................ ................................ ............... 21 Introduction ................................ ................................ ................................ ........................ 21 General Biofouling Process ................................ ................................ ................................ 21 Adhesion ................................ ................................ ................................ ..................... 22 Surface Energy ................................ ................................ ................................ ............ 24 Ulva linza ................................ ................................ ................................ ........................... 26 Surface Chemistry and Surface Energy ................................ ................................ ........ 27 Modulus ................................ ................................ ................................ ...................... 29 Other Significant Fouling Organisms ................................ ................................ ................. 39 Marine Bacteria ................................ ................................ ................................ ........... 39 Diatoms ................................ ................................ ................................ ....................... 41 Barnacles ................................ ................................ ................................ .................... 41 Summary ................................ ................................ ................................ ............................ 42 3 ATTACHMENT OF ULVA LINZA ZOOSPORES TO HIGH MODULUS, HIGH ASPECT RATIO ENGINEERED TOPOGRAPHIES ................................ ........................ 43 Introduction ................................ ................................ ................................ ........................ 43 Materials and Methods ................................ ................................ ................................ ....... 47 Surface Patterns ................................ ................................ ................................ ........... 47

PAGE 7

7 Sample Fabrication ................................ ................................ ................................ ...... 47 Surface Characterization ................................ ................................ .............................. 48 Ulva Zoospore Attachment ................................ ................................ .......................... 49 Results ................................ ................................ ................................ ............................... 50 Surface Characterization ................................ ................................ .............................. 50 Ulva Zoospore Attachment ................................ ................................ .......................... 53 Discussion ................................ ................................ ................................ .......................... 57 Conclusions ................................ ................................ ................................ ........................ 60 4 ATTACHMENT TO NANO TOPOGRAPHICALLY MODIFIED LOW FOULING MICRO TOPOGRAPHIES FOR ULVA LINZA SPORE ATTACHMENT AND RELEASE ................................ ................................ ................................ .......................... 62 Introduction ................................ ................................ ................................ ........................ 62 Materials and Methods ................................ ................................ ................................ ....... 66 Sample Fabrication ................................ ................................ ................................ ...... 66 Master fabrication ................................ ................................ ................................ 66 Replication ................................ ................................ ................................ ........... 68 Surface Characterization ................................ ................................ .............................. 69 Ulva Attachment and Release Assay ................................ ................................ ........... 70 Results ................................ ................................ ................................ ............................... 70 Surface Characterization ................................ ................................ .............................. 70 Ulva Zoospore Attachment and Release ................................ ................................ ...... 73 Discussion ................................ ................................ ................................ .......................... 76 Conclusions ................................ ................................ ................................ ........................ 78 5 ANALYSIS OF SPATIAL DISTRIBUTION OF ULVA LINZA ZOOSPORE ATTACHMENT TO ENGINEERED TOPOGRAPHIES ................................ ................... 80 Introduction ................................ ................................ ................................ ........................ 8 0 Materials and Methods ................................ ................................ ................................ ....... 82 Sample Surfaces ................................ ................................ ................................ .......... 82 Ulva Spore Attachment ................................ ................................ ............................... 83 Image Acquisition ................................ ................................ ................................ ....... 85 Image Analysis ................................ ................................ ................................ ............ 86 Spatial Distribution ................................ ................................ ................................ ..... 87 Pair distribution functions ................................ ................................ ........................... 87 Finite size effect scale factor ................................ ................................ ....................... 89 Counting method ................................ ................................ ................................ ......... 89 Visual comparison ................................ ................................ ................................ ....... 91 Results and Discussion ................................ ................................ ................................ ....... 91 Image Analysis ................................ ................................ ................................ ............ 91 Spatial Distribution ................................ ................................ ................................ ..... 92 Random distribution ................................ ................................ ............................. 93 Effect of bin size ................................ ................................ ................................ .. 93 Image analysis bias ................................ ................................ ............................... 94 Polystyrene beads ................................ ................................ ................................ 96

PAGE 8

8 n series ................................ ................................ ................................ ................. 97 Pseudo mapped ................................ ................................ ................................ .. 103 Surface defects ................................ ................................ ................................ ... 105 Conclusions ................................ ................................ ................................ ...................... 110 6 CONCLUSIONS AND FUTURE WORK ................................ ................................ ........ 111 Conclusions ................................ ................................ ................................ ...................... 111 Future Work ................................ ................................ ................................ ..................... 112 Feature Height ................................ ................................ ................................ ........... 112 Feature Width and Spacing ................................ ................................ ........................ 113 Increased Roughness ................................ ................................ ................................ 113 Combined Topographies ................................ ................................ ........................... 113 Spatial Distribution ................................ ................................ ................................ ... 114 Summary ................................ ................................ ................................ ................... 114 APPENDIX A SILICON WET ETCH STUDY ................................ ................................ ........................ 116 B OVEREXPOSURE STUDY ................................ ................................ ............................. 121 C PRELIMINARY SPATIAL DISTRIBUTION ANALYSIS ................................ .............. 123 ................................ .. 123 Attachment Mapping: Triangle Pillars ................................ ................................ 124 Perfectly preferential attachment ................................ ................................ ........ 124 D HIGH ASPECT RATIO POLYSTYRENE TOPOGRAPHIES ................................ ....... 127 E MICROCONTACT PRINTING SHARKLET PATTERN ................................ ............... 128 F ULVA ATTACHMENT TO PDMS GRAFTED SILICON NANO PILLARS ................. 129 LIST OF REFERENCES ................................ ................................ ................................ ........ 131 BIOGRAPHICAL SKETCH ................................ ................................ ................................ ... 134

PAGE 9

9 LIST OF TABLES Table page 3 1 Dimensions and ERI II values for surfaces used to evaluate the effect of feature height on spore attachment. ................................ ................................ ................................ .... 47 3 2 Water contact angles on Si surfaces. ................................ ................................ ............ 51 4 1 Dimensions and ERI variables for surfaces used to evaluate the effect of nano scale roughness on spore settlement. ................................ ................................ ...................... 66 4 2 DRIE parameters for micron scale topography. ................................ ............................. 67 4 3 DRIE parameters for nano scale roughness parameters. ................................ ................. 68 4 4 Water contact angle measurements for smooth and nano roughened PDMSe surfaces. ................................ ................................ ................................ ........................ 73 5 1 Dimensions and ERI variables for PDMSe surfaces used to evaluate the spatial distribution of spore settlement. ................................ ................................ ................... 82 5 2 Characteristics of the RDF for manual and automated image analysis counting methods. ................................ ................................ ................................ ........................ 96 5 3 Characteristics of the RDF for beads and spores on +2.4SK5x2_n4. ............................ 97 5 4 Pearson correlation table comparing RDF characteristic values and n ......................... 100 5 5 Distribution of site site distances on n series patterns. ................................ ................. 101 5 6 Most preferential r max SD, for n series. ........................... 103 5 7 Spore density and total number of measured spores for image sets with and without defects for +2.9SK2x2_n4. ................................ ................................ .......................... 106 5 8 Characteristic values of the RDFs for +2.9SK2x2_n4 with and without defects. .......... 109 A 1 HNA etch rate study slow etch rate. ................................ ................................ .......... 117 A 2 HNA etch duration study fast etch rate. ................................ ................................ ..... 119 B 1 Topographies produced using over exposure (+SK2x2_n5). ................................ ........ 122

PAGE 10

10 LIST OF FIGURES Figure page 2 1 Empirical relationship between fouling retention and surface energy. ............................ 27 2 2 Attachment and release of Ulva spores to surface grafted PDMSe. ................................ 28 2 3 Attachment of Ulva spores to COOH/OH and CH3 capped alkanethiol coated silicon of varying advancing water contact angle. ................................ ........................ 29 2 4 Attachment of zoospores of Ulva linza to topographically modified PDMSe surfaces .... 31 2 5 Images of attached of zoospores of Ulva linza to topographically modified PDMSe surfaces. ................................ ................................ ................................ ........................ 31 2 6 Attachment of spores of Ulva tall original Sharklet pattern (Shark), in PDMSe. ................................ ................................ ................................ ....... 32 2 7 Pattern designs for study of Ulva ...... 33 2 8 Correlation of Ulva spore attachment to ERI. ................................ ................................ 34 2 9 Correlation of Ulva spore attachment to Sharklet feature height. ................................ .... 35 2 10 Image of Ulva attachment to hierarchical Sharklet Ridge topography. ........................... 35 2 11 Image Attachment of Ulva to force gradient topographies. ................................ ............ 36 2 12 Attachment of Ulva to force gradient topographies vs ERI. ................................ ........... 36 2 13 A Smoothed histograms of Ulva attachment to Sharklet patterns. ................................ .. 39 3 1 Ulva Overlay of microscope images of spores of Ulva on +1(B), +2(C) and +3(D)SK2x2.. ................................ ................................ ................................ ................ 45 3 2 Post fouling optical microscope images of Ulva on +2.8SK2x2.. ................................ ... 46 3 3 Pre surface modification state SEM images of r series surfaces.. ................................ ... 50 3 4 AFM images of Si surfaces. ................................ ................................ ........................... 52 3 5 Ulva zoospore attachment density on r series surfaces. ................................ ................ 53 3 6 Flourescent and white light microscopy images of Ulva zoospore attachment on r series surfaces. ................................ ................................ ................................ ............... 54 3 7 SEM images of spores on PDMS g Si surfaces. ................................ ........................... 55

PAGE 11

11 3 8 SEM (left) and fluorescent light (right) images of spores at the pattern smooth interface. ................................ ................................ ................................ ...................... 56 3 9 SEM images of spores infiltrating between features at the pattern smooth interface. .... 57 4 1 Mapping of spore attachment.. ................................ ................................ ....................... 64 4 2 Nano pillars formed on +15SK5x2 on a silicon wafer. ................................ ................... 65 4 3 Si nano pillars etched down to their nubs on +15SK5x2. ~45 tilt. ................................ 65 4 4 Configuration for etching micron the subsequent nano scale roughness step. ................................ ................................ ..... 67 4 5 Master and Kraton mold for the +7.4SK9.8x1.6_r1 sample. ................................ ........... 71 4 6 SEM images of PDMSe replicate of Kraton mold of +7.4SK9.8x1.6_r1 masters. ......... 71 4 7 AFM images in height mode of PDMSe samples. ................................ ........................ 72 4 8 Model to estimate contact area of spore adhesive on nano roughened PDMSe surface. ................................ ................................ ................................ .......................... 72 4 9 Attachment density of spores to nano roughened samples and controls. ...................... 74 4 10 Percent removal of spores from nano roughened samples and controls under water flow.. ................................ ................................ ................................ ............................. 75 4 11 Attachment of spores before and after exposure to water flow channel. ...................... 75 4 12 Optical images of spore attachment.. ................................ ................................ .............. 76 4 13 Tensile modulus measurements. ................................ ................................ .................... 77 4 14 Compressive modulus of regular cured and accelerated cured PDMSe. ....................... 77 5 1 Scanning electron micrographs of n series surfaces ................................ ...................... 83 5 2 White and/or fluorescent light images of spore or polystyrene bead attachment to topographies. ................................ ................................ ................................ ............... 84 5 3 Image selection scheme. ................................ ................................ .............................. 85 5 4 Automated image analysis counting method scheme selection scheme. .......................... 86 5 5 Scheme for A: radial and B: angular distribution functions. ................................ ........ 88 5 6 Scheme for four methods of spore spore distance counting. ................................ ......... 90

PAGE 12

12 5 7 Size distribution of objects, single spore and spore groups, in the n series image sets.. .. 91 5 8 Pseudorandom test distribution ................................ ................................ ..................... 93 5 9 Pseudorandom RDF. ................................ ................................ ................................ .... 94 5 10 Effect of RDF bin size. ................................ ................................ ................................ 95 5 11 ADF of pseudo random distribution. ................................ ................................ ............ 95 5 12 RDF of an image analyzed using A: manual spore counting, and B: automated spore counting of pseudo random distribution. ................................ ................................ ........ 96 5 13 RDFs for +2.4SK5x2 PDMSe. ................................ ................................ ..................... 97 5 14 Comparison between mapped attachment of spores and polystyrene beads on +2.4SK5x2 PDMSe. ................................ ................................ ................................ .... 97 5 15 RDFs of n series. ................................ ................................ ................................ ......... 98 5 16 RDF characteristics for the n series ................................ ................................ ............... 99 5 17 Distances between preferential attachment sites (red stars) on A: +2.7SK2x2_n1, B: +2.7SK2x2_n2, C: +2.6SK2x2_n3, D: +2.9SK2x2_n4, E: +2.6SK2x2_n5.. ................. 101 5 18 ADF of n series for r = 0 80 m. ................................ ................................ ................. 1 02 5 19 ADF of n series for r = 160 ................................ ................................ ............ 103 5 20 n series. .............. 104 5 21 White+fluorescent light image of aligned spores on +2.6SK2x2_n4. ........................... 106 5 22 RDF of +2.9SK2x2_n4 with and without defects. ................................ ....................... 108 5 23 ADF of +2.9SK2x2_n4 with and without defects, r = 0 ................................ 109 6 1 Hierarchical surface with 2nd level of features present only in the recesses of the 1st level ................................ ................................ ................................ ........................... 114 6 2 Various combination topographies ................................ ................................ .............. 115 A 1 Wet acid Si etch of nano wide features ................................ .............. 116 A 2 Wet acid Si etch of nano wide features ................................ ........... 117 A 3 HNA slow etch rate study. ................................ ................................ ........................... 118 A 4 Wet acid Si etch at varied duration for rounding of recesses on +2 CH20x20. ............. 119

PAGE 13

13 B 1 Surfaces created using overexposure to increase feature width. ................................ .... 121 C 1 ................................ ... 123 C 2 series. A: Smooth, B: Channels, C: Pillars, D: Triangle Pillars, E: Sharklet (+2.8Sk2x2_n4). ................................ ................................ ............................ 124 C 3 Attachment mapping of +2.4TP10x2 21 177; C,D: all spores, including grouped spores, n = 239of ERI series .......................... 125 C 4 Overlay of distribution of perfectly preferential (right image) attachment to +2.6SK2x2_n5 (attachment map shown in left). ................................ ........................... 125 C 5 RDFs for perfectly preferential attachment to n series topographies. ........................... 126 D 1 High aspect ratio topographies in PS. ................................ ................................ ........... 127 E 1 Microcontact printing of Sharklet pattern. ................................ ................................ .. 128 F 1 Images of spores on PDMS g Si nano pillars. ................................ ............................ 130

PAGE 14

14 LIST OF ABBREVIATION S describes the risk of making a Type I error in statistical analysis ADF angular distribution function AFM atomic force microscopy ANOVA Analysis of Variance APTES 3 aminopropyl triethoxysilane DRIE D eep reactive ion etching ERI Engin eered Roughness Index p p value or observed significance level PDMS Poly(dimeth l ysiloxane) P DMSe P oly ( dimethylsiloxane ) elastomer Silastic T2 Dow Corning Corporation PEG Poly(ethylene glycol) RDF Radial distribution function Re Reynolds Number SD Screening distance SEM Scanning electron microscopy Si Silicon Spore Ulva linza zoospore Ulva Ulva linza

PAGE 15

15 Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy ULVA LINZA ZOOSPORE SENSITIVITY TO SYSTEMATIC VARIATION OF SURFACE TOPOGRAPHY By Julian Taylor Sheats May 201 1 Chair: Anthony B. Brennan Major: Material s Science and Engineering The use of surface topographical microstructure is abundant in nature. The lotus plant uses a fractal like topography to create a highly non wetting surface that self cleans as water drops take dirt particles with them as they roll off. Analysis of how t opography affects surface interactions offers a unique opportunity to attack a problem that affects our economy and societal health significantly. The attachment of biological material to manmade surfaces can be looked at as fouling or directed adhesion. Marine fouling on ship hulls costs the United States $600 million each year due to increased fuel usage caused by drag. Hospital acquired methicil l in resistant Staphylococcus aureus infections cause thousands of deaths annually as a result of colonization of hospital surfaces. The lack of biocompatible synthetic surfaces for implants such as vascular grafts lead to restenosis as cells are unable to develop a natural interaction with the graft surface. In each circumstance there is much to learn about the complicated attachment process. This work expands the investigatio n of the role of topography in the attachment of the green fouling algae Ulva linza to poly(dimethylsiloxane) surfaces Spore attachment density was c orrelated to the Wenzel roughness ratio on low surface energy, high modulus poly(dimethylsiloxane ) grafted silicon topographies. The role of topography on a scale less tha n

PAGE 16

16 the size of a spore was investigated on nano roughened poly(dimethylsiloxane) elastomer surfaces. For a specific group of patterns, the spatial distribution of spores attached to topographies was quantitatively analyzed and shown to correlate with feature dimensions.

PAGE 17

17 CHAPTER 1 INTRODUCTION Scope of Research Settlement and adhesion of microscopic organisms to substrata is a diverse area of research with significant implications on societal health, defense and economy. Such biological attachment is either categorized as a fouling process with negative implications for the substrata, or as a directed adhesion where a proliferation of certain cells is desired. In the marine environme fleet over $2.4 billion annually in terms of cleaning and increased fuel costs, reduces a military ecies from one environ to the next 1 Bacterial colonization transmitted to patients via hospital surfaces, known as nosocomial resistant staphylococcus aureus (MRSA) alone caused 18,650 deaths from July 2004 to December 2005 due to infections acquired during hospital stay 2 Medical implants also encounter a significant problem with bacterial infections, often due to the implant surface having a reduced resistance to colonization compared to actual tissue. A n improved engineered surface for such implants is desired both to target these infe ctions but also to create a more biocompatible joint with the implant site. In particular, the interface between the implant surface and immediately adjoining tissue should to adopt a natural confluence and morphology. Engineering such a surface requires knowledge of how surface morphology, topography, and energy effect how cells behave when in its proximity. From a materials science perspective, this problem is quite complicated. First, there is the diversity of biological objects at all scales, includi ng proteins, cells, micro and macro organisms that are present in the real environment in which a surface will be placed. Second, the

PAGE 18

18 statistical variation of biological responses within a given experiment and between experiments is quite large even when working with large numbers of organisms. This is especially true given the timescale of certain studies. Subsequent experime n t s can be separated by relatively long times, which allows a host of factors to be altered, if even slightly, in the interim. Third, even microorganisms have developed a number of settlement and attachment cues. While it is the aim of research such as this to investigate, isolate, and exploit these cues, care must be taken when making conclusions based upon experiments done with in carefully structured boundaries. The organism has evolved to evaluate a surface for favorable and unfavorable characteristics based upon environmental conditions it will have to endure over its lifetime. This is based upon the real environment in whic h a given organism will adhere to a surface and then grow and thrive. Most experiments of a reasonable duration will only encounter conditions set up by the interaction of the surface and adherent at early or other select stages of life on the surface. F urthermore, and finally, it would do well to remain abreast of the outlook that such work as this is done to engineer a surface which is, in essence, directly competing with nature. As mentioned bility to adapt to such competition. Given such complexity, isolating some of the factors involved in the attachment process of a given organism aids in investigating the attachment cues. An overall goal is to develop a set of physical relations that rela te these cues to basic thermodynamic and kinetic considerations. Here one complication that can arise is one cue having a greater preference assigned to it than others in certain conditions. One may arrive at equations that are very non monotonic. Terms may be turned on or off depending on such diverse factors as hydrodynamic conditions, surface energy, adherent concentration, and even time of year. In this way, while isolation of variables

PAGE 19

19 aids development by providing some clarity, interactions betwee n factors is of prime importance in a picture of the overall system of attachment. The aim of this work is to isolate selected factors in the attachment of the marine alga Ulva linza zoospores to engineered surfaces. Surfaces were designed and fabricated based upon an attachment model initially built upon a dimensionless ratio of topographical and surface wetting variables, called the engineered roughness index (ERI). Further develo pment of the model is investigated by extending the range over which it has been studied using these surfaces as well as an analysis of the spatial distribution of attachment. Specific Aims Specific aim 1: Investigate Ulva linza Zoospore Attachment Response to High Modulus Engineered Topographies with High Aspect Ratios This study investigates limitations of the attachment model by evaluating attachment to surfaces with much higher degrees of engineered roughness than that exami ned thus far. Increase in aspect ratio of the features is systematically studied while holding surface chemistry, surface area fraction, and feature bending moment constant. Attachment to these surfaces will be evaluated with respect to aspect ratio. Su rfaces are fabricated directly from silicon wafers which are topographically modified using photolithography and deep reactive ion etching. The surface energy is modified by surface grafting oligomeric poly(dimethyl siloxane). Specific Aim 2: Investigate Ulva linza Zoospore A ttachment t o C ombined M icro A nd N ano S cale R oughness Thus far, studies of attachment to surfaces with controlled roughness in the ERI attachment model have been confined to the micro scale. This is the same size scale for the organis ms studied. It is of interest to study the effect of high degrees of roughness on a scale less than that of the attaching organism. Roughness at this scale would have significant effects on adhesive wettability as well as surface mobility of the probing propagule. Exploiting some undesirable

PAGE 20

20 side effects of normal microfabrication techniques, surfaces having a superposition of controlled nano scale roughness upon engineered micro scale topography will be studied for Ulva linza attachment. Comparisons wi ll be made for attachment to engineered microtopographies as well as smooth surfaces, both with and without nano scale roughness. Specific Aim 3: Analyze Spatial Distribution of Ulva linza Zoospore Attachment to Engineered Topographies Among the factors that are affected by topography upon surfaces and in turn affect the attachment of settling organisms, the ability of topographical features to impede attachment of anti fouling performance. Observational evidence shows this to be the case for some of the strongly antifouling topographies developed based upon the ERI attachment model. The pair wise distribution function is used as quantitative descriptor for the spatial distribution of zoospores attached to one set of topographical surfaces. This work evaluated the spatial distributions of spores attached to smooth and patterned surfaces in a global nature (unrestricted by the unit cell). The pair wise distribution funct ion was used to measure the spatial relationship of spores attached to one group of topographies n that have been analyzed using the mapping algorithm. In order to gain insight into the correlation between the ERI II and attachment reduction, correlations between the output of the pair wise distribution function and topographical parameters were investigated.

PAGE 21

21 CHAPTER 2 BACKGROUND Introduction The attachment of cells and organisms to man made surfaces is a pro blem that is both very interesting due to the complexity of the variety of interactions and very costly in terms of its effect on general health and the economy. Bacterial infections plague medical implants 2 and hospital surfaces, costing tens of thousands of lives each year 3 In the marine environment, the focus of this work, b iofouling of ship hulls contributes to hundreds of millions of dollars of increased maintenance and fuel, annually 1 as well as a decrease in military performance Throughout history technology has been developed to reduce this cost on maintenance and performance of marine equipment of all kinds, water transport, containment and treatment vessels, as well as sewage systems. This has resulted in antifouling technologies for ship hulls of such varied complexity and success as simple wax and tar to toxic metal containing paints such a s the highly effective and recently banned tributyl tin (TBT) 4 Along the way, a wealth of literature on the attachment of aqua t ic microorganisms to surfaces under a plethora of conditions has been developed. There has been an international move tow ards non toxic antifouling strategies, and a shift has occurred with a focus on more inert surfaces that resist fouling. The goals of the following literature review are to glean some basic trends in the foulin g and antifouling processes, point out som e apparent ambiguities in the literature that seem to be a result of the aforementioned biological complexity and to provide an overview of the study of the target organism, Ulva linza as well as other marine microorganisms General Biofouling Process When a virgin surface is placed into marine water there are a number of steps that can be defined that comprise the way in which the surface becomes fouled with organisms First, water

PAGE 22

22 molecules come into equilibrium with the surface by setting up a ctric double nanometers thick 5 within the larger visc ous sublayer m in thicknes s 6 Next, within the first seconds marine proteins and other macromolecules diffuse to the surface through this sublayer and adhere through mainly van der Waals forces and ele ctrostatic interaction film that reaches equilibrium in a few hours 7 Then, marine bacteria come into contact with the surface via Brownian motion (non motile types) or swimming (motile). The bacteria develop a biofilm protective matrix formed of exo cellular polymeric substances comprising the second level of the fouling film. The final stage encompasses an interval of several days to weeks where unicellular (e.g., diatoms, some algae spores, protozoans) and multi cellular (e.g., barnacle cyprids, tubeworm larvae) eukaryotes reach the surface in a sessile or motile fashion. The components and structure of this last layer of the biofouling film will depend on the availability of species locally. Marine macromolecule s are ubiquitous and there is often a substantial concentration of marine bacteria in most areas, Additionally, the bacteria layer might be interspersed with this last layer due to localized density fluctuations. Adhesion At each step in the attachment process, there is a system of adhesion that is based upon of thermodynamic first principles When an object adheres to a surface, an interface is created. This interface supplants the interface that was present before adhesion, e.g. water is removed from the surface to allow contact. There are interfacial free energies associated with the interfaces before and after the object adheres. Interfaces with lower interfacial free energies are energetically more favorable than high ones, and this can be the overall drive for adhesion events The work of adhesion per unit area, W ij defines the energy required to for m or break the interface through equation 2 1:

PAGE 23

23 (2 1) i, j ij are the surface free energies of surfaces i and j, and interfacial free energy of the i j interface, respectively. While not explicitly denoted, such values are obtained for the case of specific third medium for example air, in the case of a drop of water on a solid surface, or water, in the case of a bacterium on an underwater solid surface. In terms of wettability the work of adhesion is related to the contact angle of a drop of liquid on a solid surface at equilibrium with a gas through the Young Dupr equation, (2 2) where ga s interface and the plane of the solid. However, in what might be seen as a suitable analogy for the complexities involved in the biofouling problem, an analysis was done on the practical issues with using these equations and standard wettability measureme nts. Schrader examined t he total free energy change when a drop of liquid contacts a surface and then comes to equilibrium 8 Equation 2 2 only deals with a drop of liquid of a hemispherical shape attaching or detaching from a flat, smooth solid surface. T he full process also involves the change in shape of the drop of liquid. Schrader considered a few interesting aspects of adhesion in a practical sense. First, the net free energy change during adhesion can be broken up into two parts. He used the surface energies of the drop in free spherical and hem ispherical sessile shapes to define 1) the change in free energy during formation of the spherical liquid drop from its saturated vapor, and 2) the change in free energy during formation of a sessile hemispherical drop on the solid surface. This net free energy showed stark differences compared to that given by equation 2 2. Second he examined the the case in practical situations. Finally, he pointed out that the above equations do not hold in

PAGE 24

24 the more general case of a drop of liquid that can undergo a chemical or ph ase change, such as a curable adhesive. This last point becomes relevant in the context of biofouling, when one considers that m any organism s secrete an adhesive compound upon a surface that changes in one or more ways during the anchoring process 5 10 Surface Energy The surface energy of a solid can be calculated from the contact angles of liquids at equilibrium on the surface. Equation 2 2 was derived from the Young equation : 11 (2 3) The forces that generate the cohesive bonds that create a surface tension and surface free energy can be separated into the categories of a) apolar dispersion force s, b) polar acid base forces, and c) polar hydrogen bonding 11 For this reason more than one type of liquid is needed to calculate a good estimate of the surface energy. The Young equation is valid for a drop on an ideally smooth and chemically homogenous surface. When a surface is rough, the contact area of the solid liquid interface will be affected by the roughness. Wenzel evaluated the contact angle that would occur from a surface with a given roughness 9 W that is related to the contact angle (2 4) where r is the degree of roughness of the surface and is defined as the total unit surface area divided by the projected planar unit surface area. Since r can only be equal to or greater than one hydrophilic surfaces with a contact angle less than 90 o will become more hydrophilic with roughness, and hydrophobic surfaces with a contact angle greater than 90 o will become more hydrophobic. The phenomenon of superhydrophilicity has been demonstrated using arrays of

PAGE 25

25 micropillars where the liquid extends beyond the actual outline of the drop and wicks between the micropillars 10 At the other end of the spectrum, superhydrophobicity can occur for rough surfaces when the wetting of the surface is actually not complete. Cassie and Baxter looked at a drop of water wetting a surface comprised of parallel fibers of a given diameter and spacing 11 They noted that depending upon the liquid solid, liquid air and solid air interfacial energies there is the condi tion where the drop only touches a fraction of the cylindrical surface of the fibers. Thus there remains an air liquid interface between the fiber tops which effectively presents to the liquid drop a n interface having a mixture of liquid LV and l iquid SL interfacial free energies. They showed that these three energies correlate with V S using the CB The apparent contact angle was defined as : (2 5) where f v and f s are the surface area fraction of air and solid, respectively. For water in air this can be simplified using (2 6) to give (2 7 ) Here we see that as f s decre CB approaches 180 o Stati c contact angles greater than 17 0 o have been observed on surfaces with very low values for f s This superhydrophobic

PAGE 26

26 effect, cleaning by water drops collecting dirt as they roll off its leaves, failing to stick at all. surface energy, that is different than that what would be observed for a perfectly smooth surface. Fu r thermore, a same way. As presented in equation 2 1, the work of adhesion is defined as a value per unit area. The contact pad of the drop will be greater or less than that defined by a disc of the same radius, for Wenzel and Cassie Baxter wetting, respectively. As the contact angle of a drop of liquid on a rough surface changes to enter one of the two wetting r egimes, Wenzel and Cassie Baxter, so too will the energy required to form or break that interface. Two points must be noted here, though. drop on a rose petal has a high contact angle yet remains adhered when the rose is upended. Second, as explained by Marmur, the scale of the roughness and the measurement drop must be taken into account when using the Wenzel and Cassie Baxter equations. 12 Usually, measurements of contact angle are obtained using drops that are orders of magnitude greater in size th an fouling organisms. Ulva l inza The green alga Ulva linza is one of the most common fouling organisms in the marine environment. Because of its ability to thrive in most types of marine conditions, it is considered the most widely spread surface fouler 13 It releases its colonizing zoospore propagules in vast quantities. The zoospores are ellipsoidal in shape with a length of about 10 m and width at the widest point of 5 m. They are flagellated with four flagella that allow them to swim at an average speed of 2 00 m/s 14 They have a very interesting attachment sequence involving

PAGE 27

27 distinct choice of attachment positions 13 that has led to their use as a standard organism for engineered surface selectivity experiments. Surface Chemistry and Surface Energy Based upon the large number of studies involving PMDS elastomer (PDMSe) 15 26 it i s perhaps one of the best understood materials with respect to attachment of zoospores of Ulva Initially, Baier performed adhesion experiments using human blood platelet rich plasma with a variety of polymeric surfaces with varied surface energies. He showed a min imum in adhesion at 25 mN/m 27 Further evidence for a variety of biofouling adherents and environments supported a trend wherein minimal fouling retention is exhibited for the region of 20 30 mN/m, as shown in Figure 2 1. PDMSe, with a surface energy of 23 mN/m, falls withi for the location of the minimum is that water has the dispersive component of its surface energy at a value of 22 mN/m. He states that, as a result, water finds the surface a thermodynamic minimum with regards to refor ming the water surface interface and displacing the adherent surface adhesion. Figure 2 1 Empirical relationship between fouling retention and surface energy. 28

PAGE 28

28 Brady, however, performed a study on similar surfaces which showed that the modulus may have been the real factor involved in fouling resistance 17 Wilson investigated the relationship between the attachment and fouling release of Ulva linza zoospores to modulus and surface energy by grafting a variety of endgroups to the surface of PDMSe. 17 In this way, the bulk modulus was held constant; the surface modulus was not measured. In Figure 2 2 we see that in this case the unmodified PDMSe with surface energy of 23 mN/m was no longer the minimum in fouling, either for initial attachment or after exposure to a shear force. Figure 2 2 Attachment and release of Ulva spores to surface grafted PDMSe 17 The attachment and release of spores of Ulva were also studied on self assembled monolayers (SAMs) of varying overall surface energy as measured by water contact angles. 29 31 The hydrophobicity was adjusted by varying the concentration of OH (or COOH) and CH 3 endgr oups. In this case as well, the bulk modulus was held constant since the substrate was gold coated silicon in each case. In Figure 2 3 we see no minimum, across three different experiments, and instead we observe a gradual reduction in attachment as th e surface becomes

PAGE 29

29 more hydrophilic. To resolve these apparent contradictions more work needs to be done, and some refinement is required in the investigation of the relationship between surface energy and attachment of spores of Ulva However, additional studies have observed Ulva attachment to be reduced on hydrophilic surfaces compared to hydrophobic ones where bulk modulus was held fixed. 32 33 Figure 2 3 Attachment of Ulva spores to COOH/OH and CH3 capped alkanethiol coated silicon of varying advancing water contact angle. TOPLEFT: [ 29 ]; TOPRIGHT: [ 31 ]; BOTTOM: [ 30 ]. Modulus Brady showed that modulus may be a contributing factor in biofouling for a subset of surfaces. However, in his work the surface energy was allowed to vary with the modulus. Chaudhury studied the effect of modulus on attachment and release of Ulva spores to PDMSe where the surface energy was held relatively fixed. 24 No effect was observed in attachment for values in the range 0.2 9.4 MPa, but release was substantially increased for the 0.2 MPa sample.

PAGE 30

30 Similarly, surface lubricity has been shown to effect the release but not attachment of Ulva spores 34 Higher modulus surfaces have yet to be systematically investigated in this way. Topography The role of topography, in particular engineered topography is the focus of studies of bioadhesion in our group. Thigmotaxis the settlement response to mechanical stimuli, has been observed to be a significant settlement cue for a number of fouling organisms. 35 Some are more sensitive to topography than others, and this may have to do with the scale of the topography or roughness relative to the size of the organis m. This leads to the question of what aspects of topography are most effective in reducing attachment from a geometrical perspective. The zoospore of Ulva linza perhaps provides an ideal candidate to study the effect of systematic changes to topography due to its behavior of actively searching a surface and choosing where to adhere. 13 This has been the focus of our group using PDMSe as the test material. One of the initial investigations observed enhancement to attachment. Spores exhibited a preference to attachment in valleys, and against circular pillars, 5 m high and wide, compa red to flat PDMSe. 15 Indeed, attachment was increased as much as tenfold. Furthermore, as shown in Figure 2 4, as the spacing of the channels or pillars was reduced from 20 m to 5 m, attachment increased. Figure 2 5 shows images of spores attached to the surface of 5 m spaced valleys and pillars. This attachment pattern led to the thermodynamic theory that by filling up recesses and forming a more flat spore surface in terface, thereby resulting in a lower surface area, the overall surface energy is lowered. The contact area for a spore that fits exactly in a recess (or against a pillar) is also higher, which likely leads to higher adhesion. This led to the hypothesis that if the spore were no longer able to fit in a recess, the surface area increase would be even higher than that of a flat surface, and the contact area would

PAGE 31

31 be reduced, leading to reduced attachment. Another likely possibility is that the spores are responding, perhaps concurrently, to the wettability differences amongst the surfaces. As described above by the theories of Wenzel and Cassie Baxter, topography will cause Figure 2 4 Attachment of zoospores of Ulva linza to topographically modified PDMSe wide circular pillars square packed with varied pitch. 15 Figure 2 5 Images of attached of zoospores of Ulva linza to topographically modified PDMSe surfaces. A: (Light tall,wide, and spaced valleys, B: tall,wide, and spaced circular pillars. 15 A B A B

PAGE 32

32 substantial differences in the way water behaves on these surfaces. To test these hypotheses, as well as investigate the effect of topography pattern geometry, a number of patterns were developed with feature spacing greater than and less than the size of a spore, ~5 m Some of the surfaces wet in the Wenzel state fully wetted while others wet in the Cassie Baxter state; all were made to be fully wetted for the Ulva assays. While no correlation to wettability was yet apparent, Figure 2 6 shows that the effect of feature spacing less than the size of spore reduced attachment substantially for the Sharklet pattern (86% reduction compared to smooth). This topography was biomimetically inspired by the denticles of the skin of certain sharks known to be resistant to spacing greater than the spore size the increase in attachment compared to smooth were much lower than in the previous experiment, and not all increased attachment; this exemp lifies biological variability inherent with this investigation. Figure 2 6 Attachment of spores of Ulva tall original Sharklet pattern wide ridges wide valleys. 26 Next, various patterns were designed that all had feature spacings of 2 m and feature widths of 2 m. In addition to the Sharklet pattern, the next experiment studied hexagonally packed pillars, ridges (also called valleys or ridges), and hexagonally packed 10 m equilateral

PAGE 33

33 triangles mixed with pillars, as shown in Figure 2 7. With the exception of the triangles, features were 2 m wide, and all were 3 m tall. To quantitatively characterize the topographical differences between these surfaces, a dimensionless ratio was defined. It involved three geometrical variables: 1) r the ratio of total surface area to projected plana r surface are, 2) f d the depressed (recessed) area fraction, and 3) df the degrees of freedom of movement when confined to move along the recesses of the topography. r is equivalent to the roughness ratio in f d is equivalent to 1 f S with f S being the solid liquid fraction in the Cassie Baxter equation. The variable was constructed based upon the trends observed in the attachment studies and named the Engineered Roughness Index (ERI). It is defined as, (2 8) All of the surfaces in Figure 2 7 reduced attachment compared to smooth. This is an indication that the scale of roughness is critical in determining whether roughness increases or reduces overall attachment. Moreo ver, as shown in Figure 2 8, there was a strong correlation to the ERI, with a correlation coefficient of R 2 = 0.69. Figure 2 7 Pattern designs for study of Ulva attachment to 21

PAGE 34

34 Using the Sharklet pattern, the feature height was also studied. 19 Feature heights of 1 m, 2 m, and 3 m were tested; note here that as feature height is increased, the bending moment of the features also decrease. Figure 2 9 shows that as feature height increased, a ttachment decreased. In this experiment spores attachment was also tested against hierarchical surfaces whereby the Sharklet topography was superimposed upon larger ridges. This surface was intended to target two species of fouling organism at once, Ulva and a species of barnacle. In Figure 2 8 Correlation of Ulva spore attachment to ERI. 21 this case, spore attachment was higher than smooth. As Figure 2 10 shows, spores preferentially settled against the taller ridges, seemingly ignoring the effect of the underlying Sharklet pattern. However, since attachment of spores to the tall ridges without a superimposed Sharklet was not tested, it cannot be certai n whether the spores were actually not affected by the addition of the Sharklet topography. Nevertheless, this result lends support to the aforementioned theory characterizing sites where a spore can rest against more than one surface as favorable. It al so offers some evidence towards the possibility that certain topographical attachment cues are weighted higher than others.

PAGE 35

35 Figure 2 9 Correlation of Ulva spore attachment to Sharklet feature height. 19 Figure 2 10 Image of Ulva attachment to hierarchical Sharklet Ridge topography. 19 Next, a first attempt was made to explain the reason for the correlation of attachment to the ERI shown in Figure 2 8, as well as for the superior antifouling performance of the Sharklet topography. It was hypothesized that as spores settled on the elastomeric features, the features would exert different forces, depending on the feature sizes, on the spore that influence attachment. In particular, when a spore rests on two rectangular features of different lengths and deflects them a finite amount, there is a force gradient across the spore body due to the different bending moment associated with each feature. A series of topographies based upon the constituent features of the Sharklet design were designed such that a linear increase in this force

PAGE 36

36 g radient, from 0 nN up to 374 nN, was obtained. However, no trend was observed that would correlate attachment with force gradient, as shown in Figure 2 11. Furthermore, when the attachment response of these surfaces wa s plotted against the ERI, the correlation wa s low 36 as shown in Figure 2 12. Two transformations of the data were performed to investigate refinements of the ERI variable towards a m odel predicting attachment of Ulva to engineered topographies. First, a natural log of attachment was chosen to show correlation with ERI due to the nearly equal variances among attachment values; for linear correlation, variance increased with attachment Figure 2 11 Image Attachment of Ulva to force gradient topographies. The force gradient across the features is displayed above the data columns 20 Figure 2 12 Attachment of Ulva to force gradient topographies vs ERI. 36

PAGE 37

37 Second, it was shown that a stronger correlation could be found by exchanging the factor df with n defined as the number of distinct features in a topographical pattern. For exampl e, in Figure 2 7 the Sharklet pattern has four distinct features, the mixed triangle/pillar pattern has two, and the ridges and pillars each have one. These transformations led to a better correlation across two different sets of experimental data and t en different topographies (R 2 = 0.86). 36 The new form, called ERI II and defined as (2 9) where is used to fit the the nomenclature of others 37 was examined with respect to the individual factors n and An attempt was made to is olate the effect each by varying one while holding the other factors constant. 36 The factor n was studied by varying the number of rectangular features in the Sharklet design while holding other parameters constant; was only held relatively constant here and varied from 0.38 to 0.49. The factor was evaluated by using positive and negative version of the Sharklet pattern, whereby in the negative version the rectangular sections ar e recessed rather than protruding. Both were shown to correlate well with attachment. Three new topographies designed based upon n and exhibited attachment that was predicted within ten percent of the actual value using the ERI II as a model for attachment. However, the model as it stands is still very much an empirical one. Thus far it only has been tested in a limited range of ERI II values and for a very small subset of patterns for those values For example, two different pat terns with the same ERI II value have not been studied. The goal is to develop the model using physical laws that would lead to a more universal predictive model. As mentioned previously, the wetting behaviors dictated by the factors r and do not apply to drops the size of spores. This does not rule out an effect on attachment that is

PAGE 38

38 related to wetting behavior; it only makes the description more complicated. And, interestingly, even in this context the attachment response to is oppo site to the effect it has on wetting. In the Cassie Baxter wetting regime, as is increased contact angle is increased and adhesion of the drop to the surface is decreased (usually recall the petal effect). In the ERI II model, attachment incre ases with The attachment response to the topographies where the feature spacing was greater than the spore size does not fit the model. However, this may be another instance where one competing attachment cue is valued much higher than another. Indeed, there is evidence that roughness greater than the size of a spore may be acting as a protection from hydrodynamic conditions. 38 These are just a few of the issues to be addressed. As such, there remain many possibilities for the actual parameters or combination of factors that act as cues for antifouling with spores of Ulva One hypothesis recently postulated is that topography on a scale less than the spore size can be effecting t he hydrodynamic conditions above the surface in an unfavorable manner, and some evidence for this has been shown. 39 Also, since the spores are known to attach gregariously 13 the geometrical patterns may be setting up impediments to this behavior. The preference of spores to attach at certain sites on a number of topographies has been studied quantitatively. 40 Figure 2 13 shows the relative attachment densities for three versions of the Sharklet design with three, four, and five distinct features. In each case, preferential attachment is observed at the intersections of the rectangular features relative to the areas on top of or between them. Certain topographical sites may be acting as attractors where a s mall number of spores can attach as normal, but larger numbers of spores are restricted at that site. Further modifications to the model relating specific topographical parameters to attachment trends are necessary to bring it to a point where concrete ties to physical relations can

PAGE 39

39 be made. This would lead to both a better understanding of the attachment behavior spores of Ulva and a more predictive set of surface design rules. Here, the goal is a minimally fouling surface, one where the energy required to attach is such that the spore decides to keep swimming. Figure 2 13 A Smoothed histograms of Ulva attachment to Sharklet patterns. 40 Other Significant Foul ing Organisms Marine Bacteria The ERI model has been shown to describe the attachment of the marine bacteria Cobetia marina when tested under flow. 39 The fact that all of the surfaces in Figure 2 7 reduced attachment is surprising for two reasons. First, the bacteria cells (1 2 m in width) are smaller than the feature spacing and can fit in the recesses. This is opposite to the behavior of Ulva where a feature spacing greater than the spore cell size increased attachment. 15 Second, C. marina is a non Ulva Settlement on a surface is dictated by gravity and fluid flow. The correlation to the ERI model, in comparison to the da ta for spores of Ulva linza was discussed in terms of hydrodynamics. As a first attempt, the Reynolds number was added as a prefactor to the ERI II This was done in order to investigate the slope of the plot of attachment vs ERI II Ulva and C. marina in two different growth phases each exhibited a different slope. It

PAGE 40

40 was hypothesized that the slope could be broken up into factors that are responded to differently by different organisms. When the estimated Reynolds numbers for Ulva and two growth sta ges of C. marina were put into the equation, the correlation for all the model to all three was high (R 2 = 0.77). 39 While this does not explain the correlation of each individual organism to the model explicitly, it suggests that hydrodynamic conditions created by the topographies may be a componen t in their antifouling performance, especially for non motile species. However, the assays for C. marina were performed under flow, at a rate of 1.24 cm/s, with an estimated flow velocity near the surface of 0.25 cm/s. 39 The Ulva assay is done in static conditions, and the spores swim at an average velocity of 20 0 m/s, almost one hundred times slower. Of course, this difference is the reason the Reynolds number was applied to the model. And yet, it could be that these results demonstrate two different attachment resistance modes, one dictated by hydrodynamic forces of a given magnitude, and one a result of spatial arrangement of attrac tion/inhibition sties. Another possibility for the attachment behavior of C. marina to topography is that the topographical features are inhibiting the initial stages of formation of a contiguous biofilm, which is the normal final mode of attachment and acts a protective layer promoting gregarious attachment and growth and inhibiting the effect of antibacterial compounds. This idea has been supported by a study on the biofilm formation of Staphylococcus aureus on the Sharklet topography in PDMSe. 41 Here formation of a biofilm was inhibited for 21 days on the Sharklet surface compared to smooth PDMSe where full biofilms formed. Interestingly, this study was done in static conditions, so the hydrodynamic component was like ly relatively reduced compared to that for C. marina

PAGE 41

41 Diatoms The attachment of the non swimming diatom Navicula perminuta was reduced compared to smooth PDMSe on topography with feature spacing less than 2 m and no space where the full body could not come into contact with a smooth area in a static assay 36 However, neither the ERI model nor attachment point theory 42 which relates attachment to the number of avail the various surfaces studied. As an example of an organism where topography with feature dimensions much less than that of its body fail to reduce attachment, we have the diatom Seminavis robusta This diatom is about 6 m wide and 30 50 m long and was tested against positive and inverse Sharklet topographies with 2 m feature dimensions. After a gentle rinse to remove unattached cells, cell density was not different from smooth PDMSe. 43 The different attachment behavior for these two similar organisms is interesting. Both Seminavis robusta and Navicula perminuta are raphid diatoms and move across a surface using grooves on the bottom of the cell surface called raphes, and neither swims. Combined, the results from these diatom studies indicate there may be a critical feature dimension for attachment resistance. This idea is supported by the attachment behavior of Ulva on 5 m x 5 m ri dges vs topographies with a feature spacing of 2 m. 15, 21, 26 Barnacles Barnacl e cyprids (final larval stage) of the species Balanus amphitrite have exhibited reduced attachment on both the Sharklet and ridges (channels) patterns. 19 The cyprids are approximately 500 m in size but have an attachment disc on the end of antennules that is 15 30 m in size that is used as a sensory appendage to search a surface. Therefore, a feature dimension of 20 m was chosen for th e width and spacing of the features in the two topographies

PAGE 42

42 to target this attachment disc. Two feature heights were tested, 20 m and 40 m to give aspect ratios of 1 and 2. For both patterns, attachment decreased with aspect ratio, resulting in an 84 % reduction for ridges and a 97% reduction for Sharklet, compared to smooth PDMSe. Summary The problem of biological fouling of aquatic surfaces is a complicated one. The diversity of the fouling organisms presents an opportunity as well as a need to st udy and attack this problem at a fundamental energetic level. The approach of this work is to add to a model that uses wetting and surface energy variables to compare the attachment performance of different topographical surfaces. Amidst the evolving des cription of how Ulva linza zoospores adhere to a variety of surface parameters, the topographical model described above has shown great promise in understanding the degree to which certain topographical factors affect attachment. Further work can be done to provide a physical basis for the model as well as produce new surfaces with improved antifouling performance.

PAGE 43

43 CHAPTER 3 ATTACHMENT OF ULVA LINZA ZOOSPORES TO HIGH MODULUS, HIGH ASPECT RATIO ENGINEERED TOPOGRAPHIES Introduction A model has been developed that focuses upon topographical factors and how they affect the initial attachment behavior of zoospores of the marine fouling alga Ulva linza 15, 21 It has shown the ability, w ithin a narrow range o f surface parameters, to be used as a predictive model for attachment to poly(dimethylsiloxane) elastomer (PDMSe) surfaces. 22 The model is an attempt to attack the problem of fouling from a fundamental en ergetics standpoint. Different topographies may have the propensity to present attachment sites that require greater energy expenditure to attach to them than others. The fouling organism capable of choice in attachment, such as Ulva linza or the barnacle Balanus amphritite 19 may choose to keep searching in hopes of a more favorable surface. Conversely, certain sites may be highly preferable and should be avoided in surface design. 15, 40 Carman et al. demonstrated that the modification of the surface water interface caused by topography ( via Wenzel 9 or Cassie Baxter 11 wetting theory ) may be related to the antifouling performance. 26 Schumacher et al. looked at the correlation between attachment and the forces differently sized features exert on a spore during attachment 19 There seems to be a threshold of topographical feature size and/or spacing, whereby attachment is reduced when this dimension is less than the size of a spore. 15, 19, 26 However, the relationship between feature dimensions and spore size still requires further examination to sufficiently describe the trend in attachment reduction Topography presents a number of attachment 15 Those that reduce attachment likely provide an energetically unfavorable surface The overall goal in this context, is to i dentify which p hysical forces correlate with attachment reduction This involves investigating

PAGE 44

44 the relationship between topographical parameters and attachment The present work investigates one of these topographical parameters. The main varia ble in the present attachment model is a dimensionless ratio of certain surface variables, here called factors. It has been named the engineered roughness index (ERI) and is defined in its second generation form 22 as (3 1) where r is the ratio of total surface area to planar projected surface area, n is the number of distinct topographical features in the pattern, and is the ratio of the surface area of feature tops to total planar projected surface area. In the typical use o f a dimensionless ratio, as the values of any combination of its factors are changed such that the value of the ratio is the same, the effect of the ratio should not change. This leads to the idea of isolation of factors, which was attempt ed for two of t he factors in the model, n and s by Long et al 22 Both were both shown to be effective at predicting attachment in the range of values tested. The third factor r was studied to some degree previously by Schumacher et al. 44 However, in that study the area the spore came in contact with may have v aried among the three surface types, such that the spores could have touched the depressed area. Three features heights As the image overlays in Figure 3 1 below show, the aspect ratio series possibly gave the spores the oppor tunity come in contact with the depressed area between features for the 1 and 2 micron tall topographies (B and C), but not for the 3 micron tall surface (D). Indeed, the matter is complicated further due to the fact that the spores are known to have the amoeboid ability to change their shape to fit against one another or surface objects. 13, 17 Two experiments have shown that the spores prefer locations where they can have a bottom surface

PAGE 45

45 intersecting a relatively tall vertical surface 15, 19 presumably to spread their adhesive to a greater extent, lower their interfacial surface energy, or protect themselves from hydrod ynamic forces 45 The relationship between attachment and height should be readdressed, inasmuch as removing the ability of the spores to touch the depressed area by going to a feature height range above 3 Using PDMSe this has not been possible due to its low modulus whic h leads to features bending and sticking together or to the depressed area. Higher modulus materials such as silicon facilitate the fabrication features with a higher aspect ratio that retain overall pattern fidelity. Figure 3 1 Ulva Overlay of microscope images of spores of Ulva 31 on +1(B), +2(C) and +3(D)SK2x2 44 Images rescaled based upon corresponding scale bars. Colored outlines drawn by hand: light green ( Ulva body), orange (features), darker green (level of Ulva lowest point), black (level of rec essed area). Another concern deals with the bending moment of the features As we see below in Figure 3 2 in the course of an attachment assay the features become deformed for PDMSe samples (modulus, E ~ 1 MPa) (left). It is not known for certain at what po int in the process this happens whether by the spores themselves in a pushing action or due to shrinkage from crosslinking of their adhesive via glutaraldehyde fixation One thing that is obvious from the image on the right is that surfaces with higher modulus (polystyrene, E ~ 3 GPa) do not become deformed even when spore numbers are high. The bending moment of the features will change as their height changes That is one factor that is not explicitly accounted for by the ERI II attachment model. In its construction ERI II is only a variable concerned with topography and not

PAGE 46

46 material factors Other variables could modify the model by affecting the slope 39 or intercept of a plot of attachment vs ERI II or as additional terms, altogether. However using ERI II in Figure 3 2 Post fouling optical microscope images of Ulva on +2.8SK2x2. A: PDMSe. B: PS (images kindly provided by Dr. Liwen Jin). Red outlines emphasize deformation of PDMSe features. comparing different surfaces should deal with surfaces that only differ topographically, to isolate the relationship between topography and attachment from the other variables The issue of feature bending moment can be dealt with by increasing the substrate modulus. The mold material for the polymer replication s single crystalline silicon (E ~ 100 GPa ) was selected for this study. Additionally, sil icon offers a surface in which the surface energy is easily tunable via silane chemistry The ability to modify the surface energy may be important for the following reason. It has been demonstrated that for PDMSe bulk surfaces grafted with polymer of v aried surface energy the Sharklet topography reduced attachment compared to smooth for all surface energies 17 However, in t hat case the bending moment of the surfaces were dominated by the bulk material, which did not change. Additionally, the reduction in attachment imparted by the Sharklet topography varied amongst the surface chemistries, with the lowest reduction exhibited for the hydroxyl and fluoropolymer grafted PDMSe. For more rigid features, it is possible that A B

PAGE 47

47 the fouling response of the spores may no longer be sensitive to the to pography at certain surface energies or surface chemistries For th is reason the silicon samples wer e grafted with short chain PDMS to lower the surface energy T he approach of this study b ased upon the se factors described above is to investigate the factor r in ERI II This is accomplished by using silicon which facilitates creation of features with aspect ratios that exceed those studied previously Additionally, a low surface energy PDMS was grafted to the silicon surface. Materials and Methods S urface Patterns Three values for the height are given in below in Table 3 1. The notation used to label the samples is as follows: +(feature height)[pattern](feature width)x(feature spacing)_[number of features]; SK is the Sharklet pattern. The first w ould allow the spore to come in contact with the bottom o f the surface. The other two w ould prevent this occurrence unless the spores greatly changed their shape during attachment. Table 3 1 Dimensions and ERI II values for surfaces used to evaluate the effect of feature height on spore attachment. When no unit is given, the quantity is unitless. Surface Height (m) Width (m) Spacing (m) r n S ERI II Smooth (SM) NA NA NA 1 0 0 0.0 +1.2SK1.6x2.4_n5 1.2 1.6 2.4 1.6 5 0.33 11.8 +5.0SK1.7x2.3_n5 5.0 1.7 2.3 3.4 5 0.36 26.4 +12.4SK1.6x2.4_n5 12.4 1.6 2.4 6.9 5 0.33 51.4 Sample Fabrication P type silicon wafers of <100> orientation approxima thick, were spin coated with Shipley S1813 photoresist and patterned using standard contact photolithography. Deep reactive ion etching (DRIE) was used to etch the topographies in a Bosch process using a STS DRIE system. An etch to passivate ratio of 7 s : 5 s was used to retain the relatively small

PAGE 48

48 considerably as height was increased leading to narrower features than desired. To compensate for this, the all widths would be the same regardless of feature height. The photoresist was removed with oxygen plasma. The wafers we re diced into 24 mm x 74 mm sections having a 24 mm x 25 mm patterned area off center. Then, the samples we 2 SO 4 : 30% H 2 O 2 1:1) (Fisher Scientific) at 80 o C for twe nty minutes. Next the samples we re sonicated in a 2% v/v solution of 3 aminopropyl triethoxysilane (APTES) (Sigma Aldrich) in toluene (Fisher Scientific) After removal of excess APTES, mono(2,3 epoxypropyl)propyl ether terminated PDMS, M n = 5000 6000 g/mol, (Gelest) wa s added dropwise neat to the sample surfaces and allowed to react fo r four hours at 80 o C. 46 Fi nally, excess short chain PDMS wa s removed by sonication sequentiall y in isopropanol chloroform and 95% ethanol (Fisher Scientific) Three replicates each of smooth PDMS grafted silicon and smooth PDMSe were used as controls. Smooth PDMSe was backed to glass slides. Surface Characterization Scanning electron microscopy (SEM) and light microscopy were used to evaluate the feature dimensions and pattern quality of the Si samples prior to surface modification SEM and fluorescent light images were taken of post assay samples. Advancing and receding w ater contact angle mea surements were taken using a Rame Hart goniometer. Ten microliter drops were placed upon the surfaces which were then tilted to the point just before the drop would roll off to determine advancing and receding angles as well as hysteresis and slip angle. Atomic force microscopy (AFM) was used to measure graft density with a NanoScope III Multimode Scanning Probe Microscope in tapping mode with a silicon cantilever 2 were

PAGE 49

49 scanned on both smooth PDMS g Si and smooth unmodified Si. Images we re evaluated qualitatively for surface morphology or graft density and quantitatively for surface graft chain dimensions and comparative surface roughness. Multiple spots on each surface were scanned to verify that a given image was representative of the surface. SEM and AFM imaging was performed at the Major Analytical Instrumentation Center (MAIC) at the University of Florida. Ulva Zoospore Attachment Three replicates of each surface type w ere sent to the University of Birmingham and assayed for Ulva sp ore attachment using methods described previously 47 48 Briefly, the samples were immersed in Tropic Marin artificial seawa ter (ASW) for 24 hr prior to the assay. Spores were obtained from fertile plants gathered from Wembury beach, UK ( 5 0 o o placed in sterile seawater suspension. Then the samples were placed at the bottom of individual wells in a Quadriperm assay dish (Fisher), and the wells were filled with 10 ml of spore suspension containing a concentration of 1.5 x 10 6 spores/ml. The spores were allowed to settle on the surfaces in the dark at ~20 o C for 45 min. Next, the surfaces were gently rinsed to r emove unattached spores, and attached spores were fixed to the surface using 2% glutaraldehyde in seawater for 15 min. The number of spores per field of view was counted using a image analysis system connected to a Zeiss epifluorescence microscope. Thirty fields of view were taken for each of three replicates per sample. Th e mean attachment densities were compared using a one comparisons. The spore attachment on the surfac es was compared to the predictions based on r and ERI II

PAGE 50

50 Results Surface Characterization The three types of surfaces in the pre PDMS graft state are shown i n Figure 3 3. In the 12.4 m sample it is seen most clearly that there is a decrease in feature width as the etch deepens. As reported in Table 3 1, the control of the width of the photoresist in the photolithography step successfully compensated for the feature w idening caused by etching At the bottom of the features of the 12.4 m sample in Figure 3 3, there are small pillar like features The se sub Figure 3 3 Pre surface modification state SEM images of r series surfaces. A) +1.2SK1.6x2.4_n5, edge on, B) +5.0SK1.7x2.3_n5, edge on, C) +12.4SK1.6x2.4_n5, edge on, D) +12.4SK1.6x2.4_n5, top down

PAGE 51

51 the etching chamber that are resistant to the etch for a given set of etch parameters. Since they are only present on the features at a depth greater than the spore length, it is expected that they do not influence attachment. Table 3 2 shows the results of water contact angle measurements on the Si surfaces. The increase in static contact angle from 23 o on the Piranha cleaned Si surface to 127 o on the PDMS grafted surface indicates a change from hydrophilic to hydrophobic behavior. The static contact angle of smooth PDMSe is 113 o The fact that the smooth PDMS g Si is 100 o indicates that there is possibly incomplete surface graft coverage. Further evidence for i ncomplete graft coverage is presented in the AFM data. Table 3 2 W ater contact angles on Si surfaces. Value is mean plus or minus standard deviation. Surface Static Advancing Receding Hysteresis Slip angle Smooth Si, Pira n h a 379 Smooth Si, with APTES 633 Smooth Si, with PDMS 1022 Smooth PDMSe 36 1131 1181 727 467 >90 Smooth PDMS g Si 1001 108 3 91 1 17 3 30.5 8.5 +1.2SK1.6x2.4_n5 Si 1207 141 8 85 3 56 8 >90 +5.0SK1.7x2.3_n5 Si 1453 161 6 126 2 35 6 29.5 1.3 +12.4SK1.6x2.4_n5 Si 1434 165 2 133 3 3 2 3 26.0 1.2 Figure 3 4 shows AFM tapping mode images of smooth PDMS g Si and Piranha cleaned samples in height and phase modes. Phase is the lag between tip oscillation drive signal and the response of the tip. For a very stiff surface with low adhesion interaction between tip and surface, the ph a se difference will be low Soft material and/or ones where there is a significant adhesive interaction between the tip and surface will exhibit phase difference quantified in degrees. The RMS roughness of the PDMS g Si sample is 0.56 nm compared to 0.15 nm for the Piranha cleaned sample. On the image s of one area the PDMS g Si sample (E and F) the re are relatively flat areas between the nodules. In the phase image (D and F), these nodules are further shown to have a different interaction with the AFM tip, being lighter regions

PAGE 52

52 compared to the surrounding darker area This suggests that these nodules are grafted chains, which are then separated by relatively large ungrafted spaces of either bare Si or Si bound with APTES One method that was planned for further characterization is a force volume map. This would Figure 3 4 AFM images of Si surfaces. A, B : Piran h a cleaned Si; C,D: PDMS g Si ; E,F: PDMS g Si (different spot); A,C,E: height mode; B,D,F: phase mode Yellow circles highlight relatively flat areas. A B C D E F

PAGE 53

53 measure the stiffness at each pixel in the AFM image to help determine which of the possible surface structures ha d been fabricated. Ulva Zoospore Attachment Figure 3 5 shows the attachment results for the PDMS g Si samples and PDMSe control. Multiple individual comparisons showed that each pair differed statistically except for smooth PDMS g Si and +12.4S K1.6x2.4_n5. The smooth PDMS g Si exhibited less than a third of the attachment of the smooth PDMSe control. The attachment on the +1.2 SK1.6x2.4_n5 sample was higher than smooth PDMS g Si by a factor of greater than five, and that of the +5.0SK 1.6x2.4_n5 sample was greater by a factor of approximately two. Figure 3 5 Ulva zoospore attachment density on r series surfaces. Error bars denote 95% confidence intervals. Asterisks denote statistically different groups. Figure 3 6 shows representative images of attachment to these surfaces. What wa s immediately apparent wa s that the typical gregarious attachment observed for smooth PDMSe wa s absent on the smooth PDMS g Si image. On the +1.2SK1.6x2.4_n5 sample, however, the attachment wa s even more gregarious than smooth PDMSe. The clumps of adjacently attached 0 200 400 600 800 1000 1200 spores/mm 2 Attachment Density to PDMSe and PDMS g Si Samples *** ** **** **

PAGE 54

54 spores seem ed to be of even greater size than that typical for smooth PDMSe. This wa s reflected in the mean spore densities. Figure 3 6 Flourescent and white light microscopy images of Ulva zoospore attachment on r series surfaces. A: Smooth PDMS g Si, B: +1.2SK1.6x2.4_n5, C: +5.0SK1.7x2.3_n5, D: +12.4SK1.6x2.4_n5 One of the main hypotheses of this would limit the ability of a spore to touch the recessed floor of the surface while this would not m tall features. This hypothes is wa s supported by the images in Figure 3 8. ed th e recessed floor whereas they d id not on the taller two samples. Images of the edges of the patterned area (pattern smooth interface) of each slide show (Fig ure 3 10) a patterns. Previously 15 it had been suggested that spores prefer a wall floor interfaces because they can A B C D

PAGE 55

55 Figure 3 7 SEM images of spores on PDMS g Si surfaces. A: +1.2SK1.6x2.4_n5, B: +5.0Sk1.7x2.3_n5, C: +12.4SK1.6x2.4_n5. Scale bar lower their overall surface area and energy by being in contact more of the sample surface. It has also been postulated that such locations offer protection against hydrodynamic forces. 38 The fact that the spores only pack ed against the pattern smooth interface for the tall patterns supports spore size (~4 5 ). An interesting behavior of the spores at the pattern smooth inter face is seen in Figure 3 11. This is the only region where the spores were observed to fit themselves between the features when attaching for the tall samples. Note that on the smooth sections of the patterned slides the spore attachment is quite grega rious with a much higher density than on the fully smooth slide (Figure 3 7 A). This was likely the result of pattern smooth interaction. Further work is necessary to investigate this difference. A B C

PAGE 56

56 Figure 3 8 SEM (left) and fluorescent light (right) images of spores at the pattern smooth interface. A,B: +5.0SK1.7x2.3_n5; C,D: +12.4SK1.6x2.4_n5; E,F: +1.2SK1.6x2.4_n5. The fluorescent images show an area of 1.4 mm x 1.1 mm. Discussion The attachment data display two results that were unexpected. First, the topography did not reduce attachment compared to smooth PDMS g Si. While the hypothesis that attachment would decrease as feature height increased is supported, the +1.2SK1.6x2.4_n5 and A B C D E F Smooth Pattern Smooth Pattern Smooth Pattern

PAGE 57

57 Figure 3 9 SEM images of spores infiltrating between features at the pattern smooth interface. A: +5.0SK1.7x2.3_n5, B: +12.4SK1.6x2.4_n5. Arrows denote spores that have fit themselves between features. +5.0SK1.7x2.3_n5 patterns significantly increased attachment, and t he +12.4SK1.6x2.4_n5 surface was not statistically different than the smooth PDMS g Si surface. Second, the smooth PDMS g Si sample exhibited a subst antially lower attachment compared to smooth PDMSe. This second result is likely due to the differences between the PDMSe and PDMS g Si surface characteristics. The contact angle characterization of the smooth PDMS g Si indicates that the surface energy was slightly higher than that of PDMSe. The surface heterogeneity shown in the AFM images indicate that the area coverage of the short chain PDMS grafts was lower than expected. Previous studies of surface grafted Si have reported attachment incre ases with contact angle for self assembled monolayers (SAMs) 29 31 A value around 100 has produced attachment densities above 1000 spores/mm 2 in each of the cases cited. The low value of attachment for the PDMS g Si in the present study, le ss than 200 spores/mm 2 indicate d that the morphology of the grafted oligomers may be the main contributor to the very low spore density. The SAMs studied form close packed homogeneous crystalline layers on the surface. This wa s in contrast to the heterogeneity exhibited by the PDMS g Si samples. Indeed, it may is due to the fact that the spore density wa s alre ady so low. Furthermore, the attachment pattern, A B

PAGE 58

58 at first glance, is more typical of a low fouling surface. In comparison to the typical attachment of spores to smooth PDMSe where they attach in groups the spores on the smooth PDMS g Si Based upon th is, o ne might hypothesize that the possibility for further attachment reduction wa s limited. However, spore densities for certain PDMSe topographies have been as low as 60 spores/mm 2 (+2.8SK2x2_n4 and +2. 5 SK10x2, unpublished). The performance of the smooth PDMS g Si leads to an idea that short chain oligomers have antifouling characteristics by virtue of their morphology alone The relatively unconstrained grafted chains may present to a settled spore small repulsive force s, such as those encountered in steric hindrance 49 50 These forces may be sensory mechanisms or may be interacting with their adhesive in an unfavorable way. The surface modulus difference between PDMSe and PDMS g Si may also be a factor here. The grafted chains here are much longer than those in SAMs but also attached at a much lower density, as suggested by the surface heterogeneity in Figure 3 4 As random coils rather than the crystalline structure found in SAMs, they produce a slightly softer surface as suggested in the AFM phase image. Even so, if we assume a simple volumetric r ule of mixtures for the extremely thin surface layer of PDMS on Si (E ~ 100 GPa), PDMSe (E~ 1 MPa) is much softer than PDMS g Si. While there has been some evidence that spores do not respond to changes in surface modulus, 24 this may be valuable direction to investigate. The attachment density on these patterned surfaces w as either higher than or equal to the smooth control However, t he feature spacing for these topograp by ca. 0.3 0.4 Previously, +5CH5x5 (channels) had greatly increased attachment compared to smooth in PDMSe. 15 The spores had fit in between the channels and packed in ordered lines.

PAGE 59

59 This led to the hypothesis that feature dimensions less than the size of the spore would decrease attachment. smooth. The results of the present study support this hypothesis and suggest that bound for attachment reduction Fe attachment compared to smooth for certain feature heights. Testing a series of surfaces with find the lower bound for attachment reduction. Such a study would also investigate the role of spore size relative to feature dimensions. The goal here would be to add a factor to the ERI model that takes size of the organism into account, which in its c urrent state it does not. Looking at Table 3 1 based upon ERI II the attachment for the +12.4SK1.6x2.4_n5 sample was predicted to be substantially lower than the +5.0SK1.7x2.3_n5 sample. However, from a purely mechanical contact perspective for the final attachment configuration, the two surfaces are identical under the assumption that the features are not b ent by the spores. The fact tall pattern suggests that the presence of the additional space beneath the feature tops renders the surface less favorable towards attach ment. Among the explanations for this, the change in hydrodynamic forces caused by the more open geometry wa s attractive. Yet, the fact that such conditions are constrained to remain in the laminar flow regime due to the length scales in question li mits the options for differences in flow characteristics between the surfaces. It is 31 probe d the recesses between features and distinguish ed between the surfaces in this way. One surface where a floor wa s encountered may be more favorable than one where no floor is found a less stable attachment site.

PAGE 60

60 However, based upon ERI II these two surfaces were also predicted to reduce attachment compared to smooth. The fact that the +12.4SK1.6x2.4_n5 sample br ought the attachment back to that of smooth indicates that the cue that increased attachment for the shorter patterns was possibly offset by a different cue feature height. A natural extension of this study would be to reevaluate the effect of feature h eight using a pattern with feature spacing at the apparent critical would be less than smooth for all cases and follow the ERI II attachment model. Conclusion s This study show ed that feature height alone can be used to reduce attachment of Ulva zoospores. Surfaces with tightly controlled engineered roughness values were successfully fabricated. The comparison to smooth PDMS g Si suggests that a critical dimension for feature spacing wa The surface energy and surface morphology of the PDMS g Si samples deviated from that of PDMSe, a standard antifouling material and caused a substantial reduction in attachment for the smooth PD MS g Si It was possible that the presence of the grafted oligomers imparted a steric hindrance repulsion factor to the surface that caused this reduction. Future studies involving surface density or molecular oligomeric weight are a promising avenue to investigate this further. For the patterned surfaces, other material and surface parameters were held constant while feature height was varied between two regimes relevant to the size of the settling organism. Within the range studied, attachment both followed feature height and suggested two reduction levels. These levels may correspond to engineered roughness values that either allow ed the spore to contact the lowest surface point or constrain ed attachment to feature tops. In this latter configurat ion, it is possible that reduction in attachment is due in part to hydrodynamic factors or more in depth probing the spores Future studies

PAGE 61

61 performed at the critical feature spacing would need to investigate the potential of feature height to reduce attachment compared to smooth for high aspect ratio patterns.

PAGE 62

62 CHAPTER 4 ATTACHMENT TO NANO TOPOGRAPHICALLY MODIFIED LOW FOULING MICRO TOPOGRAPHIES FOR ULVA LINZA SPORE ATTACHMENT AND RELEASE Introduction A model has been developed that incorporates topog raphical factors into a description of the attachment behavior of zoospores of the green alga Ulva linza to poly(dimethylsiloxane) elastomer (PDMSe) engineered microtopographies 21 22 Thus far, studies of attachment to surfaces within this model have been confined to the micro scale which is the same scale for the size of the spores. Promising results have emerged using surfaces wit h nano scale roughness, especially when this nano scale roughness is combined with micro scale roughness. 51 The variation of surface contact area on two length scales likely presents a surface resistant to ad hesion, possibly related to the way roughness effects wetting via the models of Wenzel and Cassie Baxter. 9, 11 The factors used to construct the main variable of the present attachment model, the engineered roughness index (ERI II ), involve two factors from these two wetting models, r and s : 4 1 where r n is the number of distinct topographical features in the pattern, and is the ratio of the surface area of feature tops to total planar projected surface area which is equivalent to f 1 in Cassie 22 However, these models are only valid for explaining the wetting characteristics of a drop of water on a surface when the dimensions of the features that comprise the roughness are much less tha n the size of the drop. 12 While the correlation that the model exhibits of attachment to these topographica l features may yet be explained by wetting in one way or another, the relationship is not yet explicit. Thus far, surfaces that would be wet by a

PAGE 63

63 microliter sized drop with similar contact angles way can be either favorable 15 or unfavorable surface configurations 21 for spore attachment. The difference between these two cases seems to be a critical feature dimension. 26 When the protruding features are spaced a distance of approximately 2 m, which is less than the width of a spore, attachment is reduced compared to a smooth unmodified surface, for all topography geometries tested. 19, 21 22 Yet, the addition of features spaced much less than this, on the nanometer scale, have not yet been tested within this model. Recently, locations of preferential attachment, or lack thereof, upon certain topographies have been quantified 40 One surface in particular, +2.8SK10x2 (the notation used here is +[feature height][pattern][feature width]x[feature spacing], with values in microns) 22 exhibited attachment locations and overall fouling such that it would be a good candidate for modification using nano topography. First, it is a surface that is i nteresting in terms of its relative fouling behavior with regards to the attachment model Across two assays, it exhibited attachment equal to that of the original Sharklet, +2.8SK2x2 (data not yet published) This went against two alternative hypotheses : 1) since the ERI II value of +2.8SK10x2 is higher than +2.8SK2x2, its attachment was hypothesized to be lower, following the attachment model, or 2) since the wider feature width of +2.8SK10x2 should allow spores to attach to the feature tops more like a smooth surface, its attachment was hypothesized to be higher than +2.8SK2x2. The fact that neither of these occurred suggests a more complex interaction with spores during attachment. Second, along the lines of the second hypothesis above, more spores di d attach to the feature tops of the 10x2 surface compared to the 2x2 surface. Figure 4 1 shows the preferential attachment locations of these two surfaces along with the percentage of spores that attached to the feature tops on each. While it is hard to see from the se attachment maps ten times as many spores

PAGE 64

64 attached on the feature tops of the 10x2 surfa ce compared to t he 2x2 surface Investigating the 10x2 surface further may aid in understanding of the attachment model. Figure 4 1 Mapping of spore attachment. LEFT(and expanded): +2.5SK2x2 40 RIGHT: +2.8SK10x2. The 10x2 surface is a good candidate to explore the use of nano roughness by attempting to reduce the attachment of spores to its relatively wide feature tops through the addition of nano scale features. Exploiting some undesirable side effects of normal microfabrication techniques 52 54 surfaces having a superposition of disordered nano scale roughness upon engineered micro scale topography can be fabricated. This roughness will affect two factors in the ERI model: the W enzel roughness ratio factor r and the factor s Using the wide 10x2 feature tops as platforms to c reate submicron roughness, this provide s a sample to study both the submicron roughness itself as well as the performance of significan t roughness on scales both equal to and smaller than that of a spore. Figure 4 2 shows a preliminary surface using +15SK5x2. This surface employs silicon nano pillars extending up from the feature tops in a brush like manner to enhance overall surface roughness, and ERI II These nano p illars are too fragile to replicate using the typical PDMSe curing method However, they effectively roughen

PAGE 65

65 the feature tops such that when they are mostly removed, via a wet etch, the surface shown in Figure 4 3 is fabricated. It was hypothesized that t his roughened surface w ould reduce the attachment density and strength of Ulva spores. Figure 4 2 Nano pillars formed on +15SK5x2 on a silicon wafer. Top right: ~45o tilt. Bottom left and enlarged: top down. Figure 4 3 Si nano pillars etched down to their nubs on +15SK5x2. ~45 tilt.

PAGE 66

66 Materials and Methods Sample Fabrication Master fabrication Micron scale topography. Table 4 1 below lists the surface types to be tested and their characteristics For the nano roughened surfaces, the ERI II value is only said to be greater than the value for the non roughened surface. names to denote the addition of a first generation level of nano roughness to the given sur face type. It was found that the nano roughness step reduced the width of features. To compensate for this, the feature width was increased commensurately to produce the value in the table below in the micron scale topography step. Table 4 1 Dimensions and ERI variables for surfaces used to evaluate the effect of nano scale roughness on spore settlement. When no unit is given, the quantity is unitless. Surface Height (m) Width (m) Spacing (m) r n s ERI II Smooth (SM) NA NA NA 1 0 0 0.0 Smooth_r1 NA NA NA >1 0 0 >0 +7.4SK9.8x1.6 _r1 7.4 9.8 1.6 > 2.5 4 0.8 3 >60.0 type silicon wafers of <100> orientation. Deep reactive ion etching (DRIE) was used in a Bosch process to create the micron scale features. The DRIE parameters for this step are listed below in Table 4 2. The formation of nano pillars was found to be dependent on the pre existing micron scale topography as well the areal coverage of the wafer. The configuration used is shown in Figure 4 4 Four 1 in 2 sections were etched into the wafer adjacent to each other about its center. Oxygen plasma was used to remove the photoresist after etching. Nano scale roughness Next, the wafer was subjected to a second, different DRIE process to fabricate nano pillar s on the surface of the topography. The Bosch parameters are

PAGE 67

67 Table 4 2 DRIE parameters for micron scale topography. Parameter Value (average across all cycles) Etching cycle Passivation cycle Cycle time (s) 7 5 Number of cycles 11 (end on etch) 10 Coil power (W) 668 668 Platen power (W) 12 0 Chamber pressure (mTorr) 52 21 C 4 F 8 gas flow (sccm) 0 85 SF 6 gas flow (sccm) 130 0 O 2 gas flow (sccm) 13 0 Figure 4 4 Configuration for etching micron scale topography nano scale roughness step. Dark gray area is patterned withthe unit cell outl ined schematically in white spore attachment l isted below in Table 4 3. At the given values, the low etch to passivate 52 54 The number of cycles was found to increase area coverage of the nano grass. T hen, these tall nano pillars we re etched dow n to the nubs of their bases using a wet acid etch comprised of 47% hydrofluoric acid 97% nitric acid and glacial acetic acid (HNA etch) 55 Wet etch A series of compositions were tested to produce the desired morphology. The details of this study are in Appendix A. The selected composition was 10:9:80

PAGE 68

68 HF:HNO 3 :CH 3 COOH v/v/v. A duration of 12 min was used, with 3 parts of HF and HNO 3 added at 5 min to compensate for reaction produced dilution. Finally, the nano roughened wafer was cleaved leaving some unpatterned area as a border to the patterned sections. This cleaving was performed to aid in peeling the replications from the masters since the nano roughness increases surface area and, therefore, adhesion. One patterned section was used as the master for the +7 .4SK9.8x1.6_r1 sample, and an unpatterned section approximately 1 in 2 was used as the master for the Smooth_r1. Table 4 3 DRIE parameters for nano scale roughness parameters Parameter Value (average across all cycles) Etching cycle Passivation cycle Cycle time (s) 6 6 Number of cycles 21 (end on etch) 20 Coil power (W) 668 668 Platen power (W) 12 0 Chamber pressure (mTorr) 52 21 C 4 F 8 gas flow (sccm) 0 85 SF 6 gas flow (sccm) 130 0 O 2 gas flow (sccm) 13 0 Replication The Smooth_r1 sample was first replicated with PDMSe to form a transfer mold. Due to the absence of high aspect ratio features, it was possible to cure PDMSe on already cured PDMSe with the use of a hexamethyl disilazane (HMDS) anti adhesion layer. This was not possible with the micro topography sample. The replication would often not come apart from the transfer mold, and the relatively high aspect ratio fe atures of the transfer mold would bend over. To replicate the +7.4SK9.8x1.6_r1 master in its positive state, a negative transfer mold wa s prepared out of a thermoplastic elastomer, poly(styrene ethylene butadiaene styrene) ( Kraton G1657M ), for subsequent PDMSe replication. This material has a slightly higher modulus than PDMSe. This prevented the features present in the transfer mold from bending during

PAGE 69

69 replication. Kraton pellets were dissolved in toluene (Fisher Scientific) at a ratio of 29% : 71% v/v and mixed for 24 hr Then the mixture was poured into a well fixed onto the Si master to form a film via evaporative solution casting. After approximately 24 hr, this film was then removed from the master and well. Replication using PDMSe was performed using Silastic T2 (Dow Corning) base and curing agent mixed in a 10:1 weight ratio. This mixture was degassed and poured over the transfer mold and Smooth_r1 master which were affixed to a glass plate between spacers. Air bubbles were removed using a ne edle, a glass plate was pressed onto the spacers, and the assembly was placed in an oven at 60 o C for one hour. Finally, the cured PDMSe was peeled away. This curing procedure is different from the usual method used to produce PDMSe samples in our research group as reported previously. 15, 19 22, 26, 39 40, 43, 56 The usual method cures the PDMSe at room temperature for 22 hr. To study the effect of this ch ange in material preparation, smooth samples that were cured using the usual method were used as a baseline material. Surface Characterization Scanning electron microscopy (SEM) and light microscopy were used to evaluate the feature dimensions and pattern quality. Advancing and receding w ater contact angle measurements were taken using a Ram the surfaces, and then 0.1 was added or subtracted sequentially until the contact point moved to determine advancing and receding angles as well as hysteresis. The samples were also tilted to measure the slip angle. Atomic force microscopy (AFM) was used to measure graft density w ith a NanoScope III Multimode Scanning Probe Microscope in tapping mode with a silicon 2 Images were evaluated qualitatively for surface morphology and

PAGE 70

70 quantitatively for comparative surface roughness. Multiple spots on ea ch surface were scanned to verify that a given image was representative of the surface. Ulva Attachment and Release Assay Three replicate samples of each sample type w ere adhered to a glass slide using allyltrimethoxysilane coupling and assayed for both at tachment and release using Ulva linza zoo spores and methods described previously 47 48 Briefly, the samples were immersed in Tropic Marin artificial seawater (ASW) for 24 hr prior to the assay. Spores were obtained from fertile plants gathered from Wembury beach, UK ( 5 0 o o suspension. Then the sa mples were placed at the bottom of individual wells in a Quadriperm assay dish (Fisher), and the wells were filled with 10 ml of spore suspension containing a concentration of 1.5 x 10 6 spores/ml. The spores were allowed to settle on the surfaces in the d ark at ~20 o C for 45 min. Next, the surfaces were gently rinsed to remove unattached spores, and attached spores were fixed to the surface using 2% glutaraldehyde in seawater for 15 min. The number of spores per field of view was counted using a n image an alysis system connected to a Zeiss epifluorescence microscope. Thirty fields of view were taken for each of three replicates per sample. The mean settlement densitie s were compared among the surface types using a one factor analysis of variance (ANOVA) fo llowed by a test for individual comparison s Results Surface Characterization Figure 4 5 shows SEM images of the Si master and Kraton mold for +7.4SK9.8x1.6 _r1. Figure 4 6 shows SEM images of the resulting PDMSe replicates o f the Kraton mold Visually, there is a loss of roughness from the master to the Kraton transfer mold and possibly from the mold to the PDMSe replicate.

PAGE 71

71 Figure 4 5 Master and Kraton mold for the +7.4SK9.8x1.6_r1 sample. A: Si master, B: Kraton mold for etching micron scale topography Figure 4 6 SEM images of PDMSe replicate of Kraton mold of +7.4SK9.8x1.6_r1 masters. A: tilted, B: top down. Figure 4 7 s hows AFM images for the accelerated cured smooth and Smooth_r1 PDMSe sample s The morphology shown in the small area here is visually similar to the larger areas on the feature t ops in the SEM image (Figure 4 6). The Smooth_r1 sample had an RMS roughness value of 86.7 nm compared to a value of 0.9 nm for the unmodified smo oth sample. Beyond the mere change in roughness, the parameters of interest are r the Wenzel roughness, and s the contact area of spore on surface. Due to the lack of a well defined geometry, however, measuring these values is difficult here. As a very rough model or visual A B A B

PAGE 72

72 Figure 4 7 AFM images in height mode of PDMSe samples. A: Smooth accelerated cured, B: Smooth_r1 accelerated cured. aid c ompare this to a linescan of the AFM image. The black line in Figure 4 8 is the adhesive, and persistence while the bottom section portrays that for a 500 nm p ersistence length S ince the roughness of the non nano roughened PDMSe is nearly 100x less (the height scale on the figure tops at around 330 nm while it is about 5 nm for the non nano roughened PDMSe) in this model the adhesive contacts the surface at many more points and the space between is hard to distinguish at this length scale Figure 4 8 Model to estimate contact area of spore adhesive on nano roughened PDMSe surface. Black line is the adhesive modeled as a polymer with a given persistence length, and green line is an actual linescan of an AFM image of the SM_r1 sample. TOP: Persistence l ength = 1 Persistence length = 500 nm A B

PAGE 73

73 qualitatively, under these assumptions, there is likely less contact between the surface and the roughened surface, increasing r and reducing s Table 4 4 shows the water contact angle measurements for the different surfaces. A two Sidak test for pairwise comparisons. The effect of cure was only significant for advancing angle. The factor of surface (smooth vs nano roughened vs nano roughened+Sharklet) was significant for static, advancing and receding. Table 4 4 Water contact angle measurements for smooth and nano roughened PDMSe surfaces. Asterisks denote statistically different samples. Surface Static Advancing Receding Hysteresis Slip angle Smooth Reg. cure 1122 1174 781 39 4 882 Smooth Accel. cure 1113 1124* 783 345 >90 Smooth_r1 (Accel. cure) 1144* 1194** 786 417 >90 SK_r1 (Accel. cure) 1105 1274*** 853* 425 5115 Ulva Zoospore Attachment and Release The results of the Ulva spore attachment are shown in Figure 4 8 The accelerated cured smooth samples exhibited less than half the att achment of the regular cured smooth sample. Adding nano roughness to a smooth accelerated cured sample increased the attachment almost threefold The combination of topography and nano roughness lowered the attachment somewhat. Figure 4 9 shows the release performance of the different surfaces. The spore density before and after exposure to shear flow is displayed in Figur e 4 10 Compared to previous studies of PDMSe, 17 the release from all samples was quite high. The addition of the nano roughness increased the removal, with a value over 90% for the Accel. Smooth_r1 sample compared to 54% for the Accel. Smooth sample. Figur e 4 11 shows light microscopy images of spores attached to these surfaces. What is evident is that at some point during the assay and before imaging, a substantial amount of the

PAGE 74

74 Figure 4 9 Attachment density of spores to nano roughened samples and controls. Error bars represent 95% confidence intervals. Asterisks denote statistically different groups. features of +7.4SK9.8x1.6_r1 became stuck together, reducing pattern fidelity. At this feature aspect ratio, the modulus of PDMSe was expected to be suffici ent to prevent the features sticking together. Therefore, based upon this occurrence and the unexpected attachment reduction displayed for the accelerated cured smooth sample, further characterization was performed on the accelerated cured PDMSe. Tensile modulus was measured using an Instron system equipped of L. Wilson, 17 the linear portion of the low strain region between 0.2 and 0.5 lbf. was used to calculate the elastic tensile modulus using TestWorks 4 software. A Texture Technologies Corp. TA XT plus Texture Analyzer was used to measure the compressive modulus of th e two types of PDMSe. A sample roughly 1 mm thick and 6 to 8 mm in width and length was placed under a 0.5 in. diameter Delrin compression cylinder. A crosshead displacement rate of 0.1 mm/s and a target load of 2 kg were used The linear strain region b etween 5 and 10 % was used to calculate the modul us 0 500 1000 1500 2000 2500 Reg. Smooth Accel. Smooth Accel. Smooth_r1 Accel. +7.4SK9.8x1.6_r1 spores/mm 2 Attachment Density ** ***

PAGE 75

75 Figure 4 10 Percent removal of spores from nano roughened samples and controls under water flow. Error bars represent 95% confidence intervals of arcsine transformed data. Asterisks denote statistically different groups. Figure 4 11 Attachment of spores before and after exposure to water flow channel. R emoval of spores from nano roughened samples and controls under water flow. Error bars represent 95% confidence interval s of arcsine transformed data. 0 10 20 30 40 50 60 70 80 90 100 Reg. Smooth Accel. Smooth Accel. Smooth_r1 Accel. +7.4SK9.8x1.6_r1 % Removal Removal Under Shear Flow ** *** 0 500 1000 1500 2000 2500 Reg. Smooth Accel. Smooth Accel. Smooth_r1 Accel. +7.4SK9.8x1.6_r1 spores/mm 2 Attachment and Release Before flow After flow

PAGE 76

76 Figure 4 12 Optical images of spore attachment. A: Accel. Smooth, B: Accel. Smooth_r1, C: +7.4SK9.8x1.6_r1. Figure 4 12 displays the results of tensile tests on regular and accelerated cured PDMSe. To check if additional curing occurred at room temperature after the initial heated duration, tests were performed on samples one hour after removing from the oven as well as 24 hr aft er removal. A loss in tensile modulus of over 20% occurred as a result of accelerated curing. Figure 4 13 shows the results of measurements of the compressive modulus. A loss of 30% in modulus is shown for the accelerated cured sample. Discussion The hypothesis that the spores would be less likely to attach to a surface with nano roughness was not supported by this study. However, the strength of the adhesion to such a surface wa s shown to be substantially weaker at this scale of nano roughness. Base d upon the C B A

PAGE 77

77 Figure 4 13 Tensile modulus measurements. Error bars denote standard deviations. Gray bar groups together s tatistically similar samples. Figure 4 14 Compressive modulus of regular cured and accelerated cured PDMSe. Asterisk denotes statistical significance final attachment density after exposure to flow, the nano roughened surface exhibits superior performance compared to the unmodified PMDSe. Micro scale topograp hy has shown the propensity to both reduce 20 22 and enhance 15 attachment to PDMSe surfaces. The results of this 0.927 0.704 0.722 0 0.2 0.4 0.6 0.8 1 1.2 Regular cure Accel. cure 1 hr Accel. cure 24 hr Tensile Modulus (low strain) [MPa] PDMSe Tensile Modulus 2.77 2.12 0 0.5 1 1.5 2 2.5 3 3.5 Regular cure Accel. Cure Compressive Modulus (5 10% strain) [MPa] PDMSe Compressive Modulus

PAGE 78

78 study, along with that of Cao et al. 51 demonstrate nano scale roughness has a similar ability. H igher levels of nano roughness similar to that shown in the Si masters in Figure 4 2 could further explore whether attachment reduction can scale with degree of nano roughness, given certain critical values of feature dimensions such as aspect ratio, or certain morphologies Indeed, the depressions created by the nano roughness mi spore. This may present favorable attachment sites similar to that of +5CH5x5 In such a configuration the adhesive pad area would be increased whilst overall surface energy is decreased, following the argument of Callow et al. 15 If this is the case, increasing the aspect ratio of these nano features, and possibly their areal density, may impart the attachment reduction observed for the micron scale topographies. The difference between regular and accelerated cured PDMSe was shown to have an unexpected significant effect on attachment of spo res, as well as on adhesion strength to a lesser extent. This is surprising considering the small degree of difference between the two samples. The slightly lower advancing contact angle of the accelerated cured PDMSe possibly indicates a less crosslinke d surface in which PDMS chains are more able to change their conformation and orient the more polar oxygen backbone groups towards the water droplet. The lower modulus also indicates a lower degree of crosslinking. If the attachment response is a result of these differences, then this demonstrates that the spores are p articularly sensitive to a) sub t l e changes in surface properties and b) modulus. This latter possibility is in contrast to the findings of Chaudhury et al. 24 w h o also investigated PDMSe, although using different fabrication methods. Conclusions Surfaces were fabricated with a combination of engineered micro topography and nano scale surface roughness in PDMSe. The addition of nano roughness increased spore removal under shear flow compared to smooth for these surfaces. At this level of roughnes s, attachment

PAGE 79

79 was increased for the nano roughened surface compared to the non roughened surfaces. A dependence on PDMSe modulus and/or wettability was dem onstrated. These results offer the opportunity to further explore the effect of higher degrees of n ano roughness and PDMSe modulus and surface energy.

PAGE 80

80 CHAPTER 5 ANALYSIS OF SPATIAL DISTRIBUTION OF ULVA LINZA ZOOSPORE ATTACHMENT TO ENGINEERED TOPOGRAPHIES Introduction Among the factors that affect the attachment of micro organisms to surfaces the ability of topographical features to impede attachment of propagules in close proximity to one another may be a central component to a topographical Many species settle gregariously, with propagules pac ked adjacent to one another, in order to facilitate various forms of protection from the environment. Bacteria do this to create an encompassing protective biofilm layer. 2 Ulva linza zoospores pack into groups presumably to lower their overall surface energy and/or create a stronger bond to the surface that is also more resilient against hydrodynamic forces. 38 Topographical features in PDMSe have exhibited either enhancement or reduction of atta chment of depending on certain feature dimensions. An empirical model has been developed that relates attachment to a dimensionless ratio of topographical variables, the ERI II 21 22, 39 However, the range of topographical variations that have been investigated thus far is narrow. While the correlation between topographical parameters and attachment in this model is strong, there remains the basic question as to how these parameters are exactly inhibiting attachment within this range. Observational evidence suggests that this reduction may be a result of the topography presenting a terrain where spores are less likely to settle close to one another. This can be verified by quantifying the spatial distribution of attached spores amongst different surfaces Moreover, underlying spatial trends may be identifi ed that correlate with the trends in attachment behavior seen thus far. The pair wise distribution function is a standard material s science descriptor that is often used to describe short range order in amorphous materials and scattering data. It also offers a n

PAGE 81

81 algorithm to describe the spatial distribution of objects in space By quantifying the numbe r of objects attached at increasing distance from a given objects the distances to nearest neighbors as well as their quantity are averaged and any short range order is shown Busscher look ed at the spatial distribution of bacteria attached to surfaces It was postulated that p arameters of the resulting distribution function could be correlated with surface 57 a nd inter bacteria interaction forces 58 59 such as the van der Waals and double layer forces involved in the DLVO theory of charged objects in solution. 60 In the case of spores attaching to topographically modified surfaces, the interactions are perhaps dominated by mechanical forces arising from the feature dimensions The pair wise distribution function provides a quant itative analysis for identify ing spatial trends in attachment behavior. Correlations between these trends and topographically derived attachment cues can then be made. Recently, the positions of attached spores have been mapped to the unit cell of several topographical patterns 40 Both preferential and inhibitory sites were id entified on the microtopographical patterns. However, as of now this mapping technique is purely qualitative. Applying the pair wise distribution function to the attachment of spores to topographies would introduce a level of quantitative statistical sig nificance. An analysis of the spatial distribution of the attached spores should corroborate the results of the attachment mapping. Moreover, the pair wise distribution function is sensitive to length scales greater than that of the mapping algorithm, w hich is limited to the topography pattern unit cell. As the ERI II is increased, the spore density on a surface is decreased significantly. R elatively large distances between individual spores are created a s the spore density is reduced Furthermore the gregarious attachment behavior of the spores necessitates studying a large area in order to analyze the

PAGE 82

82 response of the spores to multiple unit cells within a topograph y The intent of this study is to build upon the mapping analysis and contribute to the overall goal of understand ing the underlying mechanisms involved in the inhibitory ability of certain topographies. Th e long range, spatial distributions of spores attached to smooth and patterned surfaces w ere investigated The pair wise distribution function was used to analyze the spatial relationship of spores attached to one group of topographies n series The output of the long range pair wise distribution function is compared to attachment preferences identified using the local mapping algorithm. C orrelations between the output of the pair wise distribution function and topographical parameters were investigated in order to gain insight into the relationship between the ERI II and attachment reduction Materials and Methods S ample Surfaces The topography parameters, ERI II value, and reduction in attachment ( compared to smooth ) for the surfaces analyzed in this st udy are listed in Table 5 1 These s urfaces were produced using a silicon wafer master fabrication and PDMSe cast replication method ; SEM images of these surfaces are shown in Figure 5 1 Table 5 1 Dimensions and ERI variables for PDM Se surfaces used to evaluate the spatial distribution of spore settlement. When no unit is given, the quantity is unitless Surface Height (m) Width (m) Spacing (m) r n s ERI II % Reduction n series Smooth (SM) NA NA NA 1 0 0 0.0 N/A +2.7SK2x2_n1 2.7 2 2 2.4 1 0.38 3.9 2 +2.7SK2x2_n2 2.7 2 2 2.5 2 0.43 8.4 45 +2.6SK2x2_n3 2.6 2 2 2.3 3 0.46 13 68 +2.9SK2x2_n4 2.9 2 2 2.3 4 0.48 19 73 +2.6SK2x2_n5 2.6 2 2 2.5 5 0.49 23 77 +2.4SK5x2 _n4 2.4 5 2 1.8 4 0.66 21 63

PAGE 83

83 Figure 5 1 Sc anning electron micrographs of n series surfaces 22 (A E) (A)+2.7SK2x2_n1, (B)+2.7SK2x2_n2, (C)+2.6SK2x2_n3, (D)+2.9SK2x2_n4, (E)+2.6SK2x2_n5 and (F) +2.4SK5x2 _n4 Scale marker bars are 20 m in length for A E, and 50 m for F Ulva S pore Attachment Three replicate samples of each sample type w ere fabricated as 25 mm x 25 mm patterned areas centered between two smooth sections of the same area. This PDMSe film was then adhered to a 25 mm x 75 mm glass microscope slide using allyltrimethoxysilane coupling Samples were sent to the University of Birmingham and assayed for both attachment and release using Ulva linza spores and methods described previously 13 Briefly, the samples were immersed in Tropic Marin artificial seawater (ASW) fo r 24 hr prior to the assay. Spores were obtained from fertile plants gathered from Wembury beach, UK ( 5 0 o o in sterile seawater suspension. Then the samples were placed at the bottom of individual wells in a Quadriperm assay dish ( Fisher), and the wells were filled with 10 ml of spore suspension containing a concentration of 1.5 x 10 6 spores/ml. The spores were allowed to settle on the surfaces in the dark at ~20 o C for 45 min. Next, the surfaces were gently rinsed to remove unatta ched spores, and attached spores were fixed to the surface using 2% glutaraldehyde in A B C D E F

PAGE 84

84 seawater for 15 min. Images of spore attachment were taken for each surface for spore counting using white light, fluorescent light, or a combination of both. Figure 5 2 White and/or fluorescent light images of spore or polystyrene bead attachment to topographies. A: Smooth, B: +2.7SK2x2_n1, C: +2.7SK2x2_n2, D: +2.9SK2x2_n3, E: +2.9SK2x2_n4, F: +2.7SK2x2_n5, G: Spores on +2.4SK5x2_n4, H: Beads on +2.4SK5x2_n4. A B C D E F G H

PAGE 85

85 These images were used to calculate the attachment maps by C. Long. To act as a control, settlement of polystyrene beads on PDMSe surface topography was analyzed. Beads of the same size as spores were deposited in a solution of artificial seawater at the same concentration as that of a spore assay. White light images were taken to analyze the attachment of the beads. Examples images of s pores (beads for H) on each surface using both white and fluorescent light (fluorescent only for A, and white light only for H) are shown in Figure 5 2 Image A cquisition The scheme for imaging the sample slides was designed to sample a large portion of the patterned area. Fluorescent light only images were taken using green light, a red filter, and a 10x objective on a Zeiss microscope. The image placements shown in the scheme of Figure 5 3 started 5 mm from each patterned area edge to avoid the bias caus ed by the pattern/smooth interface. Corresponding to the density of each sample set, different numbers of images were taken for each sample set to arrive at roughly the same number of spores per sample. These ranged from 11 to 44 and were evenly spaced a cross each slide Figure 5 3 Image selection scheme. A 25 mm x 75 mm sample slide containing a 25 mm x 25 mm patterned area is shown in top left. The square below shows the positions of the images taken, and an example image of spores on +2.9SK2x2_n4 is expanded, to the right.

PAGE 86

86 Image Analysis image. T he contr ast was enhanced such that the image wa s converted into a binary one (B in Figure 5 4 ) where the spores are pure black and the background is pure white Grouped spores, which contact each other, are labeled as one large object of varied size. Examples of these large objects are circled below in red. Furthermore, the green circles show spores that actually contact each other but were erroneously counted as single spores. Finally, the spore positions were found using the Analyze Particles function, which outlines each object in red and reports its position as the centroid of outline. Spore positions were converted from pixels into microns and stored for each image. Figure 5 4 Automated image analysis counting method scheme selection scheme. B A C

PAGE 87

87 Spatial Distribution Pair d istribution functions The spatial distr ibution of spores on a surface wa s anal yzed through the use of two pair wise distribution functions: 1) a radial distribution function (RDF) and 2) an angular distribution function (ADF). Equation s 5 1 and 5 2 below are the general form of a n areal RDF and ADF, respectively 59 : ( 5 1 ) (5 2 ) The RDF calculates the ratio of the spore density within a ring of size dr that is a distance r away from a given spore to the average overall spore density, as shown in Figure 5 5 ; the ADF slices this distribution into angular sections The actual result is a histogram of inter spore distances across the whole area; each spore pair is counted twice. A value above 1 indicates clustering of objects at that distance bin, whereas when below 1 it indicates that at that distance there is a lower density of objects than the average overall density. It is by definition zero up to the minimum object diameter for hard sphere objects. Beyond this, the distance at which it separation distance characteristic of spore spore and spore topog raphy interactions. An algorithm for calculating the RDF and ADF An array was created for each image of n i objects representing its n i x n i symmetric inter object distance array A number of bins of size dr were created up to the limits of the image size ; this lengths object object distance fell

PAGE 88

88 Figure 5 5 Scheme for A : radial and B: angular distribution functions. Adapted from [ 57 ] into a given r+dr bin the value dn i (r+dr) was incremented by one. This is then divided by the area da (r+dr) i and the number of objects in the image, N i This value was then averaged over all the I number of images : (5 3 ) In order to check the algorithm, a pseudo random spatial distribution of object was simulated. A random number generator in Matlab was incorporated into a program to create a random distribution generator. The size of the objects was set to that of a ty pical spore, ca 5 15 Objects that overlapped were excluded. Twenty nine iterations of the distribution were used to calculate an d check the RDF and ADF Three characteristics were chosen to act as measureables (characteristics) and quantify the differences between RDFs for different samples. The screening distance, SD, is the point at which the RDF crosses 1; below this distance r max A B

PAGE 89

89 the distance at which the density of spores is the highest above the overall average spore density. Area ab o ut 1 is the area of the g dr (r) curve above and below 1. This quantifies the degree to which clustering of objects occurs for a sample across all r v a vg. min. dist. distance, across all spores, to a nearest neighbor. This value was calculated separately and concurrently to the RDF, but does not explicitly show up in the RDF. Finite size effect scale factor The finite image size results in a steady reduction in g dr (r) with r due to particles near the image boundaries having fewer neighbors than those at the center. This can be compensated for by adding a scale factor based upon the density of particles in an infinitely large image. This factor has been calculated for rectangular shaped images by R.J. Bell as: 61 (5 4 ) where L x and L y are the two side lengths of the rectangular image. The g dr (r) values in this study have been scaled by this factor. Counting method As mentioned above, the spores are known to settle gregariously. The aim of this analysis is to study the effect topograph y has on attaching spores investigate any negatively thigmotactic cues. For this reason, it wa s here assumed that spores that have attached in groups were less affected by the topography and more affected by the (possibly) stronger cue of gregarious at tachment. Group attachment affects the spatial distribution by shifting the screening distance towards one spore diameter. Four cases for single spores can be used for the analysis to take this into account and to reveal the effect topography has on non grouped vs grouped attachment: A ) counting all spores B ) counting only those spores that had attached at

PAGE 90

90 some measurabl e distance from other spores single spores C) counting only single spores, but also the distances between these spores and the near est spore in each grouped spore group, and D) counting only the distances between single spores and grouped spore groups. Each method, shown schematically in Figure 5 6 will generate different aspects of the total spatial distribution and has different amounts of bias associated with it. The automated image analysis used in this study results in a modifica tion of method B: in addition to those separations shown, the distances between each single spore and the center of each group were also counted. Note here that for smooth many of the spores are grouped spores, and thus its distribution w ould be heavi ly biased compared to the topographies and shifted towards higher values of SD. As such, smooth was not included in the statistical analysis of the data. Figure 5 6 Scheme for four methods of spore spore distance counting. ngle spores + Single group A B C D

PAGE 91

91 Visual comparison There is a large amount of information contained in the whole spatial distribution plots. Therefore, a visual a id was produced to compare qualitatively, in addition to quantitative data on the mapping analysis, and the RDF and ADF were plotted. This is presented in Appendix C. Results and Discussion Image An a lysis As mentioned above, the automated method used here to measure the spore positions contains a bias for grouped spores. This results in groups of spores being counted as large Figure 5 7 Size distribution of objects, single spore and spore groups, in the n series image sets. The size of 6 or more 4 5 2 3 1 0 10 20 30 40 50 60 70 80 90 SM n1 n2 n3 n4 n5 % of total number of objects measured Sample type Size distribution of "objects" single and grouped spores

PAGE 92

92 spores. A size histogram was calculated for each sample using the measured area for each object (single spore or spore group) and converting i t to a diameter, assuming a spherical object (Figure 5 7). An example is that an area encompassing a group of roughly four spores of a given size spore bin size for the histogram was based upon the typical spore diameter and encompassed values from 3.64 to 5.75 m. Spatial Distribution previously for attachment density. 21 It was found that the low number of spores this produced led to substantial noise within the spatial distribution plots. This data is shown in Appendix C. n n series sample set showed excellent correlation to both the ERI II and the factor n 22 Furthermore, the change in n from one sample to the next results in a linear change in the length largest feature. Since spores have been found to preferentially attach at the intersections of the features, the probability of a spore setting between features is low. This phenomenon shoul d show up in the radial distribution as a corresponding increase of screening distance, SD, or the position at which g(r) (the subscript dr has been removed for simplicity) reaches a maximum. As such, this analysis is intended to offer complementary data to the mapping investigations. In comparison to the image analysis performed for mapping of these surfaces (image size = 0.13 mm 2 ), these images are much larger, being 1.547 mm 2 As mentioned above, grouped spores were labeled as single large objects. This will contribute to a bias in the spatial distribution functions that effectively acts as an error. The large number of spores ( ca. 1 0,000 ) obtained with these larger images is intended to compensate for this to some degree.

PAGE 93

93 Random distribution One thousand objects were distributed in each of twenty nine iterations of the random distribution generator, one of which is shown in Figure 5 8 The RDF and ADF for a dilute Figure 5 8 Pseudorandom test distribution gas like, completely random distribution with an infinite number of objects will be equal to 1 for all angles and equal to 1 for r values equal to and greater than the object diameter and zero below 62 The pse udorandom distributions created here as well as the actual spore distributions, are more dense, like a liquid, and finite. Thus, the SD will be somewhat above the object diameter, and there will be noise about 1, as shown in Figure 5 9 Effect of b in size The effect of bin size on SD can be seen in Figure 5 10 Ideally, an infinitesimal value for dr would be used, but the trade off here is noise, even for the large number of objects in this study. F our values of the bin size dr were used to calculate the RDF for a pseudorandom distribution. Each plot is labeled with its SD bin The SD bins overlap for A, B, and C, but not for D This indicates that while noise is reduced by increasing the bin size, the SD bin can vary as the bin size is changed. An r value near the lowest value here, 1.15 m of 1.6 m was

PAGE 94

94 selected as the bin size to use for this study; this value is the result of the constraint of the discrete pixel size. Figure 5 9 Pseudorandom RDF random, dilute gas like distribution. Image analysis bias An method described above, as well as using manual counting of only single spores. This was done to estimate the extent to which the automated image analysis adds error to the RDF. We can see in Figure 5 12 Table 5 2 that the RDF characteristics of the two differ; statistical significa nce for SD, r max and area above 1 wa s not possible since only one image was used. This is intended to be a rough gauge for the error imparted by the automated method on the data presented in this study. 0 0.2 0.4 0.6 0.8 1 1.2 0 20 40 60 80 100 g(r) r ( m) RDF Randdist iter = 29, dr = 1.15 m SD= (5.8,6.9]

PAGE 95

95 Figure 5 10 Effect of RDF bin size 5. Figure 5 11 ADF of pseudo random distribution. A: short r values, B: longer range r values. 0 0.2 0.4 0.6 0.8 1 1.2 0 50 100 g(r) r ( m) RDF dr = 1.15 m SD= (5.78,6.94] 0 0.2 0.4 0.6 0.8 1 1.2 0 50 100 g(r) r ( m) RDF dr = 2.31 m SD= (4.63,6.94] 0 0.2 0.4 0.6 0.8 1 1.2 0 50 100 g(r) r ( m) RDF dr = 3.24 m SD= (6.48,9.72] 0 0.2 0.4 0.6 0.8 1 1.2 0 50 100 g(r) r ( m) RDF dr = 3.47 m SD= (6.94,10.42] 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 10 20 30 40 50 60 70 80 90 g(r=0 100, ) (degrees) ADF Rand dist r = 0 100 0 0.2 0.4 0.6 0.8 1 1.2 1.4 0 10 20 30 40 50 60 70 80 90 g(r=200 300, ) (degrees) ADF Rand dist r = 200 300 A B C D A B

PAGE 96

96 Figure 5 12 RDF of an image analyzed using A: manual spore counting, and B: automated spore counting of pseudo random distribution. Table 5 2 Characteristics of the RDF for manual and automated image analysis counting methods Image Analysis Method Avg. Min. Dist. (m) Area abo ut 1 r max (m) SD (m) Manual 35.920.1 0. 10 2 4.6 13.7 Automated 32.418.9 0. 11 1 3.7 13.7 Polystyrene b eads Figure 5 13 shows the RDF for spores and beads settled on +2.4 SK5x2 PDMSe. Qualitatively the bead RDF behaves more like a random distribution than the spore RDF. The spores tend to preferentially attac h at the corners of the features, whereas the beads are distributed more evenly across the pattern, as demonstrated by the attachment maps in Figure 5 14 It was hypothesized that analysis of the spatial distribution would complement the mapping results and reveal similar differences. This is supported qualitatively by the presence of more p ronounced peaks in the RDF of the spores compared to that of the beads and by the difference in the SD (Table 5 3). 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 200 400 g(r) r ( m) RDF n5 slide1 manual counts single spores 0 0.2 0.4 0.6 0.8 1 1.2 1.4 1.6 1.8 2 0 100 200 300 400 500 g(r) r ( m) RDF n5 slide1 automated single spores A B # spores = 223 # spores = 249

PAGE 97

97 Figure 5 13 RDFs for +2.4SK5x2 PDMSe A: Polystyrene beads B: Spore s. Table 5 3 Characteristics of the RDF for beads and spores on +2.4SK5x2_n4 Asterisks denote Objects avg. min. d ist. (m) Area abo ut 1 r max (m) SD (m) Beads 16420 0. 63 0. 23 91.7 96.8 8.4 6.1 Spores 10 5 21 1.78 0.74 22. 8 13.8 14 7 7.5 Figure 5 14 Comparison between mapped attachment of spores and polystyrene beads on +2.4SK5x2 PDMSe. A : Polystyrene bead settlement, B : Spore attachment Colors scheme is relevant to spore density only within an image and not between them n series The characteristic values ( Figure 5 16 ) of the RDFs detect the differences in attachment responses to these topographies. The contrast between smooth and the patterned surfaces is due to the fact that the majority of the spores on smooth are grouped. The SD, r max and avg. min. 0 2 4 6 8 10 12 0 50 100 g(r) r ( m) RDF +2.4SK5x2_n4, beads 0 2 4 6 8 10 12 0 50 100 g(r) r ( m) RDF +2.4SK5x2_n4, spores A A B B *

PAGE 98

98 dist f or smooth are all biased towards r values larger than that for randomly distributed single spores. As such, these values have been highlighted in yellow to distinguish them from that of the topographies, and recall that the smooth data were not added to t he sample sets when calculating statistical differences. Figure 5 15 RDFs of n series. Each plot is the average of multiple images from 2 (n1 and n2) or 3 slides (n3, n4, n5, and SM). The red horizontal lines show the position of unity for each plot. 0 2 4 6 8 10 12 14 16 0 200 400 600 800 1000 g (r) [stacked by arbitrary amount] r ( m) RDF n series, Avg of slides n5 n4 n3 n2 n1 SM 0 2 4 6 8 10 12 14 16 0 20 40 60 80 100 g (r) [stacked by arbitrary amount] r ( m) RDF n series, Avg of slides n5 n4 n3 n2 n1 SM B A

PAGE 99

99 Figure 5 16 RDF characteristics for the n series. The asterisks denote statistically different groups. A Pearson correlation table was calculated for these four characteristic values comparing them to n ( Table 5 4) Only r max the SD, and avg. min. dist. t end to increase with n though the correlation is weak for all three (maximum correlation coefficient = 0.41) This is due to the high variance (COV up to 0.91) of these values ; the means correlate very well, with values above 0.9 Surface defects, described at length below, are likely the cause of this variance. 0 10 20 30 40 50 60 SM n1 n2 n3 n4 n5 r (m) r max *** *** **/*** */** 0 10 20 30 40 50 60 70 SM n1 n2 n3 n4 n5 r ( m) Avg. min. dist. ** *** **/*** 0 2 4 6 8 10 12 14 16 18 SM n1 n2 n3 n4 n5 r ( m) SD */** **/*** *** **/*** 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 SM n1 n2 n3 n4 n5 Area About 1 */*** **** */** *** */** A B C D

PAGE 100

100 Table 5 4 Pearson correlation table comparing RDF characteristic values and n Characteristic Correlation coefficient P value r max 0.235 1.46x10 4 a vg. min. dist. 0.410 8.75x10 12 SD 0.265 1.68x10 5 Area About 1 0.027 0.669 It was hypothesized that the r max SD, or avg. min. dist. values for these samples should relate to certain distances defined by any existing preferred attachment sites. Such attachment sites were previously identified by the mapping algorithm. The distances between neighboring sites (red and yellow stars) are shown in Figure 5 17 All of the patterned sample ADFs ( Figure 5 18 ) exhibit a peak at the feature orientation (long axis), alt hough there is some variance about 0 that is likely caused by uncertainty in the orientation of the features (~3) N o such peak occurs on the smooth slide ADFs This will be further discussed below. preferential along the feature long axis, the minimum distance between neighboring sites is 8 m for each surface except for +2.7SK2x2_n1, for which both the maximum and minimum is 6 m. The maximum separation distance for the other surfaces is simply the length of the longest feature: n2 8 m, n3 12 m, n4 16 m, n5 20 m. The SD and r max seem to follow the minimum separation distances ( ; the Pearson correlation coefficients are rough ly the same as for n Again, the variance of these characteristic values result ed in low er coefficients. In addition to the contribution from surface defects t he variance could be related to the other less likely preferential sites There is a distribution of these distances within each pattern The mean distance and standard deviation between all the sites shown in Figure 5 17 are given in Table 5 5 It stands to reason that the variation in SD and r max is may partially be a result of the variation in separation distance between preferred settlement sites. N ext nearest neighbor distances might also be a factor.

PAGE 101

101 Table 5 5 Distribution of site site distances on n series patterns Surface Mean (m) Stand. Dev. +2.7SK2x2_n1 5.7 3.6 +2.7SK2x2_n1 7.0 4.6 +2. 6 SK2x2_n1 9.7 5.0 +2. 9 SK2x2_n1 13.9 6.5 +2. 6 SK2x2_n1 15.9 7.7 Figure 5 17 Distances between preferential attachment sites (red stars) on A: +2.7SK2x2_n1, B: +2.7SK2x2_n2, C: +2.6SK2x2_n3, D: +2.9SK2x2_n4, E: +2.6SK2x2_n5. Red stars are 1st most preferential and yellow starts are 2nd most preferential. preferential nd most preferential site on +2.6SK2x2_n5 (denoted in yellow in E in Figure 5 17 ) is 31% and 27%, A B C D E

PAGE 102

102 respectively; recall that these values are not statistically significant. Therefore, many spores were attached to the ends of the both the 12 and 16 m long features. That fact that the SD may follow the minimum separation of the preferred attachment sites is intuitive and supports the mapping analysis. The fact that the r max values are larger than the SD and are somewhat greater than the maximum preferred site separation indicate that a more complex interaction with the topography is presenting itself. Figure 5 18 ADF of n series for r = 0 80 m Each plot is the average of multiple images from individual slides. The red horizontal lines show the position of unity for each plot. 10 0 10 20 30 40 50 60 70 80 90 g(r =0 80 m, ) [stacked by arbitrary amount] (degrees) ADF By slide, r = 0 80 m n5 1 n5 2 n5 3 n4 3 n4 2 n4 1 n3 1 n3 2 n3 3 n2 1 n2 2 n1 1 n1 2 SM 1 SM 2 SM 3

PAGE 103

103 Table 5 6 Most preferential r max SD, for n series (+SK2x2_ ) n1 n2 n3 n4 n5 Minimum site separation distance (m) 6 8 8 8 8 < SD > (m) 6.6 7.1 7.8 8.5 8.4 st and. dev 1 3 2 3 1.5 2 7 2 7 Maximum site separation distance ( m ) 6 8 12 16 20 < r max > ( m ) 10.7 11.8 15.6 18.7 19.3 st and. dev 7.9 5.6 13.2 16.9 11.8 preferential is based upon qualitative mapping data and does not imply a statistical significance. Figure 5 19 ADF of n series for r = 160 individual slides. Pseudo mapped A set of fake distributions was constructed based upon the perc entage of spores that attached to th e preferred sites identified by the mapping algorithm. This was done to study the 10 10 30 50 70 90 g( r= 160 240 m, ) [stacked by arbitrary amount] ADF By slide, r = 160 240 m n5 3 n5 2 n5 1 n4 3 n4 2 n4 1 n3 1 n3 2 n3 3 n2 c n2 a n1 c n1 b SMa SMb SM c

PAGE 104

104 differences between the spatial distribution analysis and the local information gained from the mapping, which also cannot study inter spore relationships. These were mapped Each 1 um 2 location in an area representing an image was given a probability of containing a spore. This probability was a product of two probabilities. The first is the overall probability of a spore residing in any random 1 um 2 spot which was given by the density of objects calculated by the spatial distribution algorithm for each pattern The second is the probability associated with the various positions within the unit cell of each pattern given by mapping 40 These distribution were then run through the spatial distribution algorithm and the outputs compared to th ose of the actual samples (Figure 5 20 ). Figure 5 20 R DF of n series. 0 5 10 15 20 25 30 35 40 45 n3 n4 n5 r ( m) r max Actual Pseudo mapped 0 10 20 30 40 50 60 70 n3 n4 n5 r ( m) Avg. min. dist. Actual Pseudo mapped 0 2 4 6 8 10 12 14 n3 n4 n5 r ( m) SD Actual Pseudo mapped

PAGE 105

105 Surface defect s The peaks in the ADFs ( Figure 5 18 ) are near the feature orientation and we re quite large compared to the rest of the plot especially for +2.9SK2x2_n4. This was, however, evident to the author during the image acquisition. Many images showed long lines of spores aligned with the f eature orientation ( Figure 5 21) It appears that there are areas where the features are stuck together, or deformed in some other way. This is especially evident in the region where the spores are aligned with the feature orientation. The question is, did these occur befo re attachment, during, or afte r ? T he spores release an adhesive when they attach that becomes crosslinked during glutaraldehyde fixation and shrinks, pulling features together. T here are areas where features are deformed in the absence of spores T he re we re indications in the literature 13 that some spores exude a temporary adhesive prior to final attachment at another location Therefore, one cannot conclude that the features were deformed prior to spore attach ment during the assay, or another unknown mechanism However, based upon the appearance and size of the defects, it is most likely that the defects were formed prior to attachment. Since the goal of this anal ysis is to study the attachment to topographical geometry alone, these defective regions will act as an error or bias in the calculations. In order to determine the degree to which attachment to the defective regions biases the spatial distribution, a set of images for a single sample were acquired that contain only areas that are nearly completely free from defects. The images will have some defects since the attachment of the spores themselves usually creates defects, but these will be isolated to the attachment site alone. The +2.9SK2x2_n4 sample was chosen since it is the most studi ed surface. The images that are nearly completely free of defects had spore densities less than the sets of images that contain defects for slides 2 and 3 ( Table 5 7) For slide 1 only 12 images could be found that were

PAGE 106

106 Figure 5 21 White+fluorescent light image of aligned spores on +2.6SK2x2_n4. The white arrow designates the measured feature orientation along their long axis. The red circle highlights those spores aligning with the feature orientation. Some are grouped spores and some are containing images that were used to calculate the RDF and ADF are less than those for the set containing defects ( Table 5 7 ). This results in the Figure 5 22 and Figure 5 23 ). Table 5 7 Spore density and total number of measured spores for image sets with and without defects for +2.9SK2x2_ n4 Image set Avg. # s pores per image (meanstdev) Total number of spores in RDF/ADF Slide 1 with defects 11422 5035 Slide 2 with defects 427144 5121 Slide 3 with defects 278 81 5004 S lide 1 without defects 138 30 1652 Slide 2 with out defects 11626 1743 Slide 3 with out defects 86 19 1289 Feature orientation

PAGE 107

107 A s expected, the removal of the defects increases the average minimum distance between spores ( avg. min. dist. ). It also makes the peak in the ADF near the feature orientation disappear. ( Figure 5 1 ) The fact that the degree of clustering ( A rea abo ut 1 ) is increased was not expected, but this is most likely due to the degree of noise in the RDF. Neither the SD nor the r max were statistically different. The lack of a peak near the feature orientation indicates that there is no preference for the feature direction across the range of values r = 0 80 m. This suggests that the preference for Figu r e 5 23 ) is likely only due the pre sence of the defects. As such, the discussion above concerning correlations of the SD and r max must be readdressed. The range of r values above does not exclude the possibility that within a shorter regions. As shown in Figure 5 18 as the r value range increases, the ADF peaks diminish. Suppose that spores separated by either the SD or r max showed a preference with the feature direction. Then, the correlations postulated above would still appear to exist. However, at these low r values the numbers of spores is low, often zero. This low number of spores prevents conclusions. A larger number of spores is necessary to further investigate this point. The image analysis described here suggested the following idea. Previously, it was shown that spores align with 5 m tall, wide and spaced channels (+5CH5x5). 15 The defects on the n series creates 4 m wide channels. The spore attachment in these defects correlates with the alignment observed on the +5CH5x5. However, spore attachment on the n series was reduced compared to smooth. It was increased on +5CH5x5 by a factor of four. The two differences between here are 1) the surfaces are not completely covered with the defect regions where the 4 m wide channels occur and 2) the 4 m wide channels are not complete, being

PAGE 108

108 broken up by the intersections of the features. Visually, the packing density of spores in the channels of +5CH5x5 is less than that within the defect streaks. It is likely that if these defects are removed, then the attachment reduction will incr ease. Figure 5 22 RDF of +2.9SK2x2_n4 with and without defects. A: all r values, B: short r values. The red lines indicate the value of 1 for each plot. 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 200 400 600 800 1000 g (r) [stacked by arbitrary amount] r ( m) RDF +2SK2x2_n4 with and without defects, Avg of slides, all r values n4 without defects n4 with defects 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 0 20 40 60 80 100 g (r) [stacked by arbitrary amount] r ( m) RDF +2SK2x2_n4 with and without defects, Avg of slides, short r values n4 without defects n4 with defects A B

PAGE 109

109 Table 5 8 Characteristic values of the RDFs for +2.9SK2x2_n4 with and without defects Images Avg. min. d ist. (m) Area abo ut 1 r at m ax (m) SD (m) With defects 48.213.8 0.20 0.09 18.716.9 8.52.7 Without d efects 61.410.2 0.34 0.10 16.010.7 9.54.0 Figure 5 23 ADF of +2.9SK2x2_n4 with and without defects, r = 0 for each plot. It has been thought in our group that it is the intersections themselves that produce the reduction in attachment. It seems that first, the spores are d rawn to the intersections to attach, and then further attachment near those sites around the first spore is inhibited. Now, as the distance between those sites is increased (the longest feature length), it would follow that attachment should go down. The correlation between r max and the separation distance appears to support this hypothesis. The lack of a difference between these values for images with and tion to the SD and r max at least for the +2.9SK2x2_n4 sample. This could be extended by analyzing the defect bias for the other samples. This could also be addressed by testing nearly defect free 0 5 10 15 20 25 30 10 10 30 50 70 90 g(r=0 80 m, ) [stacked by arbitrary amount] (degrees) ADF +2.9SK2x2_n4 with and with out defects, r = 0 80 m n4 without defects 3 n4 without defects 2 n4 without defects 1 n4 with defects 3 n4 with defects 2 n4 with defects 1

PAGE 110

110 samples which our group is currently attempting t o fabricate. Furthermore, idea that spores tend to avoid attaching near spores that have attached to the preferential settlement sites could be studied with a spatial distribution that is localized upon each site. Conclusions This analysis made progress towards quantifying the trends observed for spores attaching to topographically patterned PDMSe surfaces. A relationship between the SD and separation distance between preferred attachment sites has been suggested T he SD for the n series was s h own to be between 6.6 and 8.5 m, and the r max values were between 10.7 and 19.3 m The effect of topography defects on the RDF and ADF has been addressed for one sample. Finally, importance of the intersections between features has been emphasized as a key parameter for the reduction in attachment compared to smooth and topographies such as channels.

PAGE 111

111 CHAPTER 6 CONCLUSIONS AND FUTU RE WORK Conclusions This work extended the development of the e mpirical model for attachment of Ulva linza zoospore to topographically modified surf aces. The r factor in the ERI II variable has been shown to correlate directly with attachment for aspect ratios below and above 1 The ability of tall fe atures to prevent spores from contacting the recessed surface floor was demonstrated to correlate with decreased attachment. This reduction was greater than that produced by the value of r alone. The fact that attachment was reduced, successively, across two feature heights where spores could not contact the recessed floor suggests a hydrodynamic component to the performance of antifouling topographies. Additionally, t he fact that features spaced 2.3 m did not reduce attachment compared to smooth sup port ed the hypothesis that there wa s a critical feature spacing near 2 enhanced our understand ing of the complex interaction between spore attachment and the particular geometrical parameters that define topographies. This also shows that potential for further understanding by isolating feature spacing at values near and below s also substantially reduced ( 69%) on smooth PDMS g Si surfaces compared to smooth PDMSe. This suggest ed the antifouling potential of surface grafted oligomers or high modulus surfaces. The surface of smooth and patter ned PDMSe was modified by the additions of a nano scale roughness to both increase r and s the area fraction of surface a spore contacts. Attachment density was surprisingly increased to this nano roughness which could be the result of spores conforming their bodies to the nano features. However, the strength of the attachment was reduced compared to the smooth control. This su ggest ed that the contact area was indeed lowered as intended. After removal of spores under shear water flow, the nano roughened

PAGE 112

112 surface had the lowest attachment density of all samples tested. Additionally, spore sensitivity to small changes in bulk modulus and/or wettability of PDMSe was demonstrated. Attachment was reduced ( 58 %) on PDMSe cured at an accelerated rate at elevated temperature. This material had a lower modulus (20%) and advancing water contact angle (112 vs 117) compared to PDMSe cured at room temperature. The spatial distribution of spores attached on a set of topographies was analyzed quantitatively. The aver age distance between nearest neighbors of non grouped spores was shown to correlate with the ERI model factor n The screening distance of the pairwise distribution function and the distance where it obtains a maximum were shown to correlate with the minimum and maximum feature length, respectively. This suggests a route for pattern design for enhanced antifouling performance: increase the lengths of th e shortest and longest feature s Finally, the role of defects was isolated in the spatial distribution of spores. Spores were shown to align with the long axis of the features only when the features had stuck together in long lines, thereby creating wide r channels for the spores settle. This demonstrates the potential for improvement in antifouling performance by slightly increasing the modulus of PDMSe in order to prevent defects. Future Work Feature Height The study of the effect of feature height alon e on attachment to topographies suggests that at higher values than studied in this work, attachment to PDMS g Si will be reduced even further. Fabrication of such tall features has since been made possible while controlling feature width. It wa s hypoth esized that when feature width is near 2 m and feature spacing is slightly below 2 m, feature heights of 10 m and above w ould reduce attachment similar to that for PDMSe, i.e., greater than 90% compared to smooth. Furthermore, this study would deter mine

PAGE 113

113 the degree to which the bulk modulus of the substrate affects attachment since the surface chemistry and energy are similar to that of PDMSe. Feature Width and Spacing Based upon the results of Chapter 3, the Ulva spores appear to exhibit sensitivity to feature spacing on the nanometer scale The surfaces shown in Appendix B would provide a sample set to test this hypothesis. Of note is the fact that no additional photomask design or fabrication is necessary as a result. Increased Roughness The effect of nano scale roughness has only been explored at a low level in this work. Moreover, the combination of nano scale roughness and micro scale topography was not targeted directly. Sur faces with higher values of nan o like nanopillars ( Figure 4 2 ) and nano/micro combination surfaces would be possible using the Si masters as attachment assay samples. Combined Topographies A number of surfaces containing two levels of topography using different geometries have been fabricated. These surfaces would allow for an analysis of the role of more complex topographies that greatly extend the ERI II variable. of interest that would investigate the effect of the 2 nd level features being present only within the recesses ( Figure 6 1 ). Figure 6 2 displays some of these surfaces ; the fabrication of a surface like that in Figure 6 1 is shown in F. This last surface would open up the potential to study changes in hydrodynamic conditions, isolated from the 1 st level topography.

PAGE 114

114 Figure 6 1 Hierarchical surface with 2nd level of features present only in the recesses of the 1st level Spatial Distribution The spatial distribution analysis in this study only focused on single spores. It would be interesting to look at the spatial distribution of groups of spores (those in gregarious contact with one another) as well as the interaction between single spores and groups. This involves using the counting methods C and D in Figure 5 6 Additional parameters in the RDF would be useful to gain a better picture of how the spores are organizing themselves when a ttaching to these surfaces. This could include focusing on places on the RDF plot where the value dips below one, so called inhibitory distances. Finally, one would want to increase the number of topographies within the sample set to cover more of the ERI II model range. Summary Th ese proposed studies would extend our investigation of the interaction between spores and topography. In such a complicated system, broad scale analysis is necessary to gain confidence in the results that are presented The goal of formulating a thermodynamic model for this interaction is an ambitious one, and cautious, thorough steps offer the tools to reach this goal.

PAGE 115

115 Figure 6 2 Various combination topographies A B C D E F

PAGE 116

116 APPENDIX A SILICON WET ETCH STUDY A variety of wet acid etch compositions and durations were used to develop a controlled etch tuned to the size of the nano scale roughened surfaces. A solution composed of h ydrofluoric, ni tric, and acetic acid (HNA) was used as the wet isotropic etchant. This solution is usually used as a very fast etch wit h a relatively high concentration of the nitric acid, which is the oxidative species; the hydrofluoric acid is present to etch away the silicon dioxide product of this oxidation. At high concentrations of nitric and/or hydrofluoric acid, this solution has been well characterized 63 For our purpose this needed to be slowed down considerably by increasing the concentration of the acetic acid. However, these compositions have not been studied extensi vely in the literature. An initial attempt is shown in Figure A 1 This show s a sample where the features were Figure A 1 Wet acid Si etch of nano pilars on 5 m wide features A: near vapor solution interface, B D: in solution A B C D

PAGE 117

117 Figure A 2 Wet acid Si etch of nano wide features A:prior to etch, B: near vapor solution interface, C D: in solution. initially 5 m wide. Half of the sample (ca. 2 mm 2 in size) was dipped into the solution for 10 s. The diluent, acetic acid, was increased in conce ntration, and a much more slow, controlled etch was produced. The series of compositions studied is sh own in Table A 1, with results in Figure A 3. The composition that created the desired morphology is G (shown in Chapter 4 as Figure 4 5 ), and the images before and after etching are labeled G 1 and G 2, respectively. Table A 1 HNA etch rate study slow etch rate. Compostition Acetic acid Hydrofluoric acid Nitric acid Duration A 75 15 16 1 min B 71.4 14.3 14.3 1 min C 68.2 13.6 18.2 1 min D 65.3 21 13.7 1 min E 62 20 18 1 min F 59 23 17.2 1 min G 80.6 9.2 10.2 12 min A B C D

PAGE 118

118 Figure A 3 H NA slow etch rate study. An attempt to round off the recessed regions of +20CH20x20 was made using the HNA wet etch solution. The aim here was to attempt to form a surface that would address the problem species ( Ulva spores and Balanus A B C D E F G 1 G 2

PAGE 119

119 barnacles). 19 There the Ulva spores seemed to preferentially attach to the intersection of the recess bottom and the feature walls. After forming the features in the DRIE, the photoresist was left on the feature tops to act a shield against the wet etch in order to prevent rounding o f the feature tops. The compositions used are listed in Table A 2. Images of the results are in Figure A 4; the photoresist was removed before imaging. Table A 2 HNA etch duration study fast etch rate Compos ition Acetic acid Hydrofluoric acid Nitric acid Duration A 50 18 32 0 s B 50 18 32 1 s C 50 18 32 5 s D 50 18 32 10 s E 50 18 32 20 s F 50 18 32 60 s Figure A 4 Wet acid Si etch at varied duration for rounding of recesses on +2 CH20x20 A B C D E F

PAGE 120

120 The result was that while the recesses did become somewhat rounded, the feature width became correspondingly narrower. This prevented the subsequent formation of the Ulva specific Sharklet pattern to be transferred to the channel feature tops. One way to deal with this would be to increase the initial feature width to compensate for the narrowing during wet etch.

PAGE 121

121 APPENDIX B OVEREXPOSURE STUDY The exposure parameter of the photolithography process step was investigated for use in extending the use of the photolithography masks to produce different feature dimensions. By exposing the photoresist (PR) for extended durations, ultraviolet light esc apes the confines of the mask dark fields at an intensity that makes the PR soluble in it s developer. Figure B 1 displays surfaces fabricated using this method. The goal here is to study the effect of feature width and spacing at a submicron level. The photomask feature width and spacing were nominally both 2 mJ/cm 2 performed a Karl Suss MA6 mask aligner in vacuum contact mode with a lamp intensity of 7.3 mW/cm 2 Figure B 1 Surfaces created using overexposure to increase feature width A B C D

PAGE 122

122 Table B 1 Topographies produced using over exposure (+SK2x2_n5). Value is shown as the mean of multiple features and one standard deviation Exposure Dose Width Spacing (mJ/cm 2 ) 197 2.680.03 1.500.10 292 2.960.02 1.240.04 438 3.490.06 0.640.06 584 3.770.08 0.500.04

PAGE 123

123 APPENDIX C PRELIMINARY SPATIAL DISTRIBUTION ANALYSIS Figure C 1 21 A B C D E

PAGE 124

124 Figure C 2 series. A: Smooth, B: Channels, C: Pillars, D: Triangle Pillars, E: Sharklet (+2.8Sk2x2_n4) Attachment Mapping: Triangle Pillars rectangular asymmetric unit and unit cell. This necessitated a slight modification to the mapping algorithm template created by C. Long 40 to take into account the triangular geometry. The results of m apping spore attachment to this surface are shown as a Figure C 3 (A,B: single spores only, C,D: all spores, including grouped spores). Perfectly preferential attachment Figure C 4 shows, as a partial overlay of a portion of the distribution over the at tachment map for +2.6SK2x2_n5, the scheme for creating the visual aid. Every preferential attachment site has been filled, and no other sites on the surface are occupied. This is similar to a crystalline lattice, and the RDFs are correspondingly sharply p eaked, within the limitations of the binsize dr Figure C 6 displays the RDFs for the n series. A B C D E

PAGE 125

125 Figure C 3 Attachment mapping of +2.4TP10x2 21 includin g grouped spores, n = 239 of ERI series Figure C 4 Overlay of distribution of perfectly preferential (right image) attachment to +2.6SK2x2_n5 (attachment map shown in left). A B C D

PAGE 126

126 Figure C 5 RDFs for perfectly preferential attachment to n series topographies. 0 1 2 3 0 10 20 g(r) r ( m) RDF +SK2x2_n1 PerfDist 0 1 2 3 4 0 10 20 g(r) r ( m) RDF +SK2x2_n2 PerfDist 0 1 2 3 4 0 10 20 30 g(r) r ( m) RDF +SK2x2_n3 PerfDist 0 1 2 3 4 5 0 10 20 g(r) r ( m) RDF +SK2x2_n4 PerfDist 0 1 2 3 4 5 0 5 10 15 20 25 g(r) r ( m) RDF +SK2x2_n5 PerfDist A B C E D

PAGE 127

127 APPENDIX D HIGH ASPECT RATIO POLYSTYRENE TOPOGRAPHIES High aspect ratio topographies are not possible using PDMSe They are possible using silicon wafers, but this is an expensive process. We can use an inverse PDMSe mold for thermoplastic replication to form high aspect ratio features in materials like polystyrene (PS). Aspect ratios as high as 20 have been fabri cated in PS using the Triangle Pillar topography ( Figure D 1 ). Additionally, features as small as 100 nm are possible using this technique; these features are the result of micromasking during the wafer DRIE step. Figure D 1 High aspect ratio topographies in PS

PAGE 128

128 APPENDIX E MICROCONTACT PRINTING SHARKLET PATTERN It is possible that the topographical patterns that have shown antifouling characteristics may exhibit similar reduction in attachment when present only in a 2 D fas h ion. Recently, such a s tudy was performed using channels of hydrophilic and hydrophobic s urface grafts. 32 Surfaces were fabricated using the microcontact printing method 64 that have a nanometer thick, amphiphilic chemical graft layer in the Sharklet pat tern Glass microscope slides were coated with a thin gold palladium layer using sputtering. Th thin layer of octadecanethiol ( ODT, hydrophobic thiol) and pressed onto the slide s; the thiol group bonds with the gold palladium layer to form a self assembled monolayer (SAM) Finally, the slide is immersed in a solution of mercaptohexadecanoic acid ( MHDA, hydrophilic thiol). The scheme for this and an image of a patterned slide are shown belo w Figure E 1 Microcontact printing of Sharklet pattern. The scheme for the fabrication is shown on the left. and a light microscope image of the a sample near the edge of the patterned area is shown to the right; the contrast has been enhanced

PAGE 129

129 APPENDIX F ULVA ATTACHMENT TO PDMS GRAFTED SILICON NANO PILLARS A preliminary sample was studied for s pore attachment to PDMS g Si nano pillars A high areal density of nano pillars was fabricated on the smooth section of the +12.4SK1.6x2.4_n5 sample using the DRIE etch to passivate ratio of 7s:5s This area was imaged using fluorescent light and a 5x ob jectiv e for an image area of 6.1 mm 2 A counting method was devised for this larger area and lower resolution. Fluorescent objects that appeared to be groups of spores were labeled. Using the objects that appeared to be single spores, an average value f or the area of a spore was calculated. This value was used to designate the number of spores in the labeled groups, which then gave the number of spore in each image. The images were also evaluated in white light to look for defects, and the image s containing defects higher density of spore at the defects were removed from the calculation. From 12 images, a n average spore density of 18 spores/mm 2 ( standard deviation 15 spores/mm 2 ) was observed. This wa s a reduction of 89% compared to smooth PD MS g Si, and a reduction of 97% compared to smooth PDMSe. Moreover, the adhesion of the spores appear ed to be quite low, as shown in the SEM images below where groups of spore have peeled away from the tips of the nano pillars. This peeling wa s likely a result of the crosslinking during glutaraldehyde fixation or the imaging conditions.

PAGE 130

130 Figure F 1 Images of spores on PDMS g Si nano pillars. A C: SEM images, D: fluorescent light image, A B C D 1000 m

PAGE 131

131 LIST OF REFERENCES 1. Champ, M.A., Science of the Total Environment 2000 258 21 71. 2. Schierholz, J.M. and J. Beuth, J. Hospital Infection 2001 49 87 93. 3. Weinstein, R.A., Emerging Infectious Disease 1998 4 416 420. 4. Yebra, D.M., S. Kiil, and K. Dam Johansen, Progress in Organic Coatings 2004 50 75 104. 5. Lyklema, J., Fundamen tals of Interface and Colloid Science Vol II : solid liquid interface Vol. 2. 1995, London: Academic Press. 6. Abelson, A., D. Weihs, and Y. Loya, Linmology and Oceanography 1994 39 164 169. 7. Wahl, M., Marine Ecology Progress Series 1989 58 14. 8. Schrader, M.E., Langmuir 1995 11 2585 2589. 9. Wenzel, R.N., Industrial and Engineering Chemistry 1936* 28 988 994. 10. Ding, C., et al., J. Microelectromech. Syst. 2010 19 878 884. 11. Cassie, A.B.D. and S. Baxter, Transactions of the Faraday So ciety 1944 40 546 551. 12. Marmur, A. and E. Bittoun, Langmuir 2009 25 1277 1281. 13. Callow, M., Callow, J., Pickett Heaps, J., Wetherbee, R., Journal of Phycology 1997* 33 938 947. 14. Granhag, L.M., A.I. Larsson, and P.R. Jonsson, Marine ecolog y progress series 2007 344 63 69. 15. Callow, M.E., et al., Biofouling 2002 18 237 245. 16. Hoipemeier Wilson, L., et al., Biofouling 2004 20 53 63. 17. Wilson, L.H., Materials Science and Engineering 2005 PhD 18. Estes, T.G., et al., Polymer P reprints 2004* 45 610 611. 19. Schumacher, J.F., et al., Biofouling 2007 23 307 317. 20. Schumacher, J.F., et al., Langmuir 2008 24 4931 4937. 21. Schumacher, J.F., et al., Biofouling 2007 23 55 62. 22. Long, C.J., et al., Biofouling 2010 26 411 419.

PAGE 132

132 23. Gudipati, C.S., et al., Langmuir 2005 21 3044 3053. 24. Chaudhury, M.K., et al., Biofouling 2005 21 41 48. 25. Fang, J., et al., Polymer 2010 51 2636 2642. 26. Carman, M.L., et al., Biofouling 2006 22 11 21. 27. Baier, R.E., et al ., Journal of Biomedical Materials Research 1985* 19 1157 1167. 28. Baier, R.E., J. Mater. Sci.: Mater. Med. 2006 17 1057 1062. 29. Ista, L.K., et al., Applied and Environmental Microbiology 2004 70 4151 4157. 30. Callow, M.E., et al., Applied and Environmental Microbiology 2000* 66 3249 3254. 31. Finlay, J.A., et al., Integr. Comp. Biol. 2002 42 1116 1122. 32. Finlay, J.A., et al., Langmuir 2008 24 503 510. 33. Schilp, S., et al., Biointerphases 2007 2 143 150. 34. Bowen, J., et al., J R. Soc. Interface 2007 4 473 477. 35. Fletcher, R.L. and M.E. Callow, European Journal of Phycology 1992 27 303 329. 36. Long, C.J., Materials Science and Engineering 2009 Ph.D. 133. 37. Quere, D., Annual Review of Materials Research 2008 38 71 99. 38. Granhag, L.M., et al., Biofouling 2004 20 117 122. 39. Magin, C.M., et al., Biofouling 2010 26 719 727. 40. Long, C.J., Biofouling 2010 26 941 952. 41. Chung, K.K., et al., Biointerphases 2007 2 89 94. 42. Scardino, A.J., E. Harvey, and R. De Nys, Biofouling: The Journal of Bioadhesion and Biofilm Research 2006 22 55 60. 43. Magin, C.M., Biomedical Engineering 2010 PhD 183. 44. Schumacher, J.F., et al., Biofouling 2007 1 45. Granhag, L.M., et al., Biofouling 2004 20 5. 4 6. Lee, M.J., et al., Advanced Materials 2006 18 3115 3119. 47. Callow, M.E., et al., Journal of Phycology 1997 33 938 947.

PAGE 133

133 48. Finlay, J.A., et al., Integrative and Comparative Biology 2002 42 1116 1122. 49. Jeon, S.I., et al., J. Colloid and Int erface Sci. 1991 142 149 158. 50. Kleshchanok, D. and P.R. Lang, Langmuir 2007 23 4332 4339. 51. Cao, X., et al., Advanced Functional Materials 2010 20 1984 1993. 52. Jansen, H., et al., J. Micromech. Microeng. 1995 5 53. Stubenrauch, M., et al., J. Micromech. Microeng. 2006 16 S82 S87. 54. Jung, K., et al., J. Vac. Sci. Technol. B 2010 28 143 148. 55. Hull, R., ed. Properties of Crystalline Silicon 1999, The Institution of Electrical Engineers: London. 1003. 56. Hoi pkemeier Wilson, L., et al., Biofouling 2004 20 53 63. 57. Busscher, H. and J. Sjollema, Colloids and Surfaces 1990 47 337 352. 58. Busscher, H., et al., FEMS Microbiology Letters 1990 69 263 270. 59. Busscher, H.J., J. Biomater. Sci. Polymer Edn. 1992 3 85 94. 60. Busscher, H.J., A.T. Poortinga, and R. Bos, Current Microbiology 1998 37 319 323. 61. Bell, R.J., Nature 1969 221 50 51. 62. Allen, S.M. and E.L. Thomas, The Structure of Materials 1999, New York, NY: John Wiley & Sons. 63. Gad el Hak, M., ed. MEMS: Design and Fabrication 2005, Taylor & Francis: New York, NY. 64. Whitesides, G.M., et al., Nanotechnology 1992 7 452 457.

PAGE 134

134 BIOGRAPHICAL SKETCH Julian Taylor Sheats was born to William Henry Sheats and Marie Anne Lisa Frame on June 1, 1980, in Ft. Lauderdale, Florida. He has a younger brother, Remi, and a younger sister, Jade. His childhood is filled with warm memories of playing football, basketball, racquetball, rollerblading, and videogames with his siblings and friends. After eight years of Catholic primary school at St. Ambrose in Deerfield Beach, FL, he attended Dillard High School in their Performing Arts magnet program for visual arts in Ft. Lauderdale. Julian spent one and a half years at Rochester Institute of Technology in Rochester, NY where he tried out Industrial Design and c hemistry as majors before deciding upon p hysics. He thereupon transferred to the University of Florida where he earned a Ba chelor of Science in p hysics in 2003. He received a teaching assistantship to attend the University of Southern California where he received a Master of Arts in p hysics in 2005. After spending one year teaching learning disable d elementary and high school students, Julian returned to the University of Florida where he received the Alumni Fellowship Award to study Materials Science and Engineering. He received a Master of Science in 2009 and his doctorate in 201 1 under the advisement of Professor Anthony B. Brennan