UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

MEMS-based anti-biofouling - mechanism, devices and application Yeh, Po Ying 2009

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2009_fall_poying_yeh.pdf [ 3.65MB ]
Metadata
JSON: 24-1.0067165.json
JSON-LD: 24-1.0067165-ld.json
RDF/XML (Pretty): 24-1.0067165-rdf.xml
RDF/JSON: 24-1.0067165-rdf.json
Turtle: 24-1.0067165-turtle.txt
N-Triples: 24-1.0067165-rdf-ntriples.txt
Original Record: 24-1.0067165-source.json
Full Text
24-1.0067165-fulltext.txt
Citation
24-1.0067165.ris

Full Text

 MEMS-BASED ANTI-BIOFOULING - MECHANISM, DEVICES AND APPLICATION  by  Po-Ying Yeh   B.A.Sc., National Taiwan University, 1997 M.A.Sc., National Taiwan University, 1999   A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF   DOCTOR OF PHILOSOPHY  in  THE FACULTY OF GRADUATE STUDIES (MECHANICAL ENGINEERING)    THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  April 2009  © Po-Ying Yeh, 2009  ii ABSTRACT A novel anti-biofouling mechanism based on the combined effects of electric field and shear stress was reported. The mechanism was observed in millimeter-scale piezoelectric plates coated with different metal materials and microfabricated Micro-Electro-Mechanical Systems (MEMS) devices. Experimental observation on the quantities of protein desorption and theoretical calculations on surface interactions (van der Waals, electrostatic, hydrophobic, shear stress) have been carried out. This anti-fouling mechanism can also be activated by a vibrating micromachined Si/SiO2 membrane. The combined effect of polyethylene glycol (PEG) grafting and application of vibration on attenuation of protein adsorption was also investigated. Vibrating PEG-grafted surfaces significantly attenuate protein adsorption, especially at low PEG grafting densities. Polymer steric interaction dominates over vibration interaction with protein on surfaces with high PEG grafting densities. Monothiol-functionalized hyperbranched polyglycidols (HPG-SH) were synthesized and self-assembled on the gold surface. The characteristics of the polymer were studied and compared with linear PEG using various surface analysis techniques. This hyperbranched polyglycidol is more resistant to protein adsorption than is linear PEG of similar molecular weight. In addition, higher molecular weight HPG shows less protein adsorption than does lower molecular weight HPG. The hyperbranched polyglycidols (without a thiol group) were further modified to generate functionality for microchannel-based liquid chromatography applications. The microchannel surface was first amino modified by allylamine plasma, and amino groups then reacted with N- hydroxy succinimide-functionalized HPGs to form strong amide bonds. The grafted HPGs are resistant to nonspecific protein adsorption. The succinimidyl ester groups degrade in water to form carboxyl groups on HPGs. By giving extra carboxyl groups to each HPG, the HPG can selectively capture positive avidin from a mixture of avidin and bovine serum albumin (BSA). To increase the capture efficiency, the microchannel was integrated with micropillar arrays as the liquid chromatography column.   iii TABLE OF CONTENTS Abstract ........................................................................................................................................... ii Table of Contents........................................................................................................................... iii List of Tables ................................................................................................................................ vii List of Figures .............................................................................................................................. viii List of Illustrations.........................................................................................................................xv List of Symbols and Abbreviations.............................................................................................. xvi Acknowledgements..................................................................................................................... xxii Co-authorship Statement............................................................................................................ xxiii  Chapter 1   Introduction................................................................................................................1 1.1 Overview................................................................................................................................1 1.2 Literature Review...................................................................................................................2 1.3 Organization of the Thesis .....................................................................................................6 1.4 Bibliography ..........................................................................................................................8  Chapter 2  Electric Field and Vibration-Assisted Nanomolecule Desorption and Anti- Fouling for Biosensor Applications ......................................................................................10 2.1 Introduction..........................................................................................................................10 2.2 Materials and Methods.........................................................................................................14 2.2.1 Materials .......................................................................................................................14 2.2.2 Methods.........................................................................................................................14 2.3 Results and Discussions.......................................................................................................17 2.3.1 Adsorption Isotherm and Fluorescence Quenching......................................................20 2.3.2 Electrochemical Impedance and Electrical Equivalent Circuits Fitting .......................31 2.3.3 Lateral Forces................................................................................................................36 2.3.4 Vertical Forces ..............................................................................................................43 2.4 Conclusions..........................................................................................................................49 2.5 Bibliography ........................................................................................................................50     iv Chapter 3  An Investigation on Vibration-Induced Protein Desorption Mechanism Using Micromachined Membrane and PZT Plate.........................................................................53 3.1 Introduction..........................................................................................................................53 3.2 Materials and Methods.........................................................................................................56 3.2.1 Materials .......................................................................................................................56 3.2.2 Device Fabrication and Characterization......................................................................56 3.2.3 Protein Adsorption and Vibration.................................................................................58 3.2.4 Scaning Electron Microscope .......................................................................................59 3.3 Results and Discussions.......................................................................................................59 3.4 Conclusions..........................................................................................................................68 3.5 Bibliography ........................................................................................................................70  Chapter 4  Attenuation of Protein Adsorption by using Self-Assembling Monothiol- Terminated Polyethylene Glycol and Vibration .................................................................72 4.1 Introduction..........................................................................................................................72 4.2 Materials and Methods.........................................................................................................74 4.2.1 Proteins and Polymers...................................................................................................74 4.2.2 Substrate Preparation ....................................................................................................74 4.2.3 Protein Polymer Modification.......................................................................................75 4.2.4 Protein Adsorption and Vibration Experiments............................................................76 4.2.5 Surface Characterization...............................................................................................77 4.2.5.1 X-ray Photoelectron Spectroscopy (XPS) Analysis ..............................................77 4.2.5.2 ATR-FTIR Spectroscopy.......................................................................................77 4.2.5.3 Fluorescence Intensity Measurement.....................................................................78 4.3 Results and Discussions.......................................................................................................78 4.3.1 PEG Coating and Surface Characterization..................................................................78 4.3.2 Protein Adsorption Studies ...........................................................................................85 4.4 Conclusions..........................................................................................................................91 4.5 Bibliography ........................................................................................................................93    v Chapter 5  Self-Assembled Monothiol-Terminated Hyperbranched Polyglycerols on Gold Surface: A Comparative Study on The Structure, Morphology and Protein Adsorption Characteristics with Linear PEG .......................................................................95 5.1 Introduction..........................................................................................................................95 5.2 Materials and Methods.........................................................................................................97 5.2.1 Materials .......................................................................................................................97 5.2.2 Polymer Synthesis.........................................................................................................98 5.2.3 Gold Substrate Preparation ...........................................................................................99 5.2.4 Preparation of Monolayer on Gold Substrates..............................................................99 5.2.5 Surface Characterization.............................................................................................100 5.2.5.1 Ellipsometry.........................................................................................................100 5.2.5.2 ATR-FTIR spectroscopy......................................................................................102 5.2.5.3 X-ray Photoelectron spectroscopy (XPS) analysis ..............................................102 5.2.5.4 Atomic Force Microscopy ...................................................................................102 5.2.6 Protein Adsorption Studies .........................................................................................103 5.3 Results and Discussions.....................................................................................................104 5.3.1 Polymer Synthesis and Characterization ....................................................................104 5.3.2 Surface Modification with Polymers ..........................................................................106 5.3.3 Surface Characterization: FTIR and XPS...................................................................107 5.3.4 Polymer Film Thickness and Graft Density................................................................110 5.3.5 Polymer Film Morphology .........................................................................................116 5.3.6 Protein Adsorption ......................................................................................................118 5.4 Conclusions........................................................................................................................122 5.5 Bibliography ......................................................................................................................123  Chapter 6   Covalently Bonding of Carboxyl Hyperbranched Polyglycerols on PDMS- Based Microfluidic Device for Liquid Chromatography Applications...........................126 6.1 Introduction........................................................................................................................126 6.2 Materials and Methods.......................................................................................................129 6.2.1 Materials .....................................................................................................................129 6.2.2 Microfabrication .........................................................................................................130 6.2.3 Polymer Synthesis and Grafting Protocol...................................................................130  vi 6.2.4 Contact Angle Measurements .....................................................................................132 6.2.5 FTIR Measurements....................................................................................................132 6.2.6 EOF Measurements.....................................................................................................132 6.2.7 Non-specific Protein Adsorption ................................................................................133 6.2.8 Ion-exchange Chromatography and Protein Separation in Microchannels ................133 6.2.9 Optical Profiler............................................................................................................134 6.2.10 Scanning Electron Microscope (SEM) .....................................................................134 6.3 Results and Discussions.....................................................................................................135 6.4 Conclusions........................................................................................................................151 6.5 Bibliography ......................................................................................................................152  Chapter 7   Conclusion ..............................................................................................................155 7.1 Summary ............................................................................................................................155 7.2 Future Work .......................................................................................................................158  Appendices..................................................................................................................................155 Appendix A: Experiment Details of Chapter 2.........................................................................155 Appendix B: Experiment Details of Chapter 3 .........................................................................163 Appendix C: Experiment Details of Chapter 4 .........................................................................167 Appendix D: Experiment Details of Chapter 5 ........................................................................170 Appendix E: Experiment Details of Chapter 6 .........................................................................179   vii LIST OF TABLES Table 2.1 Summary of fitting parameters of electrical equivalent circuits....................................34 Table 2.2 Shear stress applied on BSA molecules and related parameters ...................................41 Table 2.3 Protein adsorption on positively and negatively charged surfaces. ...............................44 Table 2.4 Variables used in Equ. (2.15)-(2.23)..............................................................................48 Table 3.1 Protein adsorption on surfaces with different conditions ..............................................66 Table 5.1 Characteristics of polymer-grafted surfaces obtained from AFM analysis .................115 Table 6.1 The properties of HPGs and number of COOH groups before and after grafting to PDMS surfaces.......................................................................................................................131   viii LIST OF FIGURES Figure 2.1 Schematic diagram of the proposed piezoelectric membrane as an implantable sensor coating. The membrane is composed of two driving electrodes with a piezoelectric material in between. Adsorbed protein can be desorbed by an electric field and carried away by acoustic streaming generated by the vibration.........................................................................13 Figure 2.2 (a) BSA adsorption isotherms on a PZT surface incubated at 37 oC for 1 hr and 30 min. The average fluorescence intensity was analyzed by Adobe Photoshop 5.0, and is proportional to the adsorbed BSA on the surface. The relationship between adsorbed BSA and the concentration of BSA in solution fits the Langmuir isotherm well. (b) Fluorescence microscope images of PZT plate surfaces with 4 vibrating conditions: (1) No vibration, (2) 10 Vpp at 16 kHz, (3) 10 Vpp at 1 kHz, and (4) 3 Vpp at 16 kHz for 5 min. (c) BSA adsorption measurement based on Fig. 2(b). (d) Protein adsorption characterization of BSA onto PZT with titanium surfaces, and anti-mouse IgG onto fired silver surfaces....................................19 Figure 2.3 Fluorescence images of (a) control Ti (250 nm)-PZT plate, (b) low concentration (2*10-4 mg/ml) BSA incubated 1 hr on Ti (250 nm)-PZT plate, (c) control Au (50 nm)-PZT plate, and (d) low concentration (2*10-4 mg/ml) BSA incubated 1 hr on Au (50 nm)-PZT plate. Stronger fluorescence intensities are observed in (b) and (d) compare to (a) and (c), respectively ..............................................................................................................................23 Figure 2.4 The dependence of modified fluorescence intensities (modified with fluorescence intensity of surface incubated in 2 mg/ml BSA for 1 hr) of (a) Ti (250 nm)-PZT plate and (b) Au (50 nm)-PZT plate. The data is fitted with the Langmuir adsorption isotherm.................24 Figure 2.5 The plot of 1/Fluorescence intensities of Au (50 nm)-PZT plates with 1/C (data from Fig. 2.4(b)) ...............................................................................................................................26 Figure 2.6 Illustration of BSA conformations on surfaces in a bimodal isotherm ........................27 Figure 2.7 Dependence of fluorescence intensities of Au (50 nm)-PZT plates with ln(1/C) (data from Fig. 2.4(b)) ......................................................................................................................29 Figure 2.8 Dependence of max max 1 1ln( ) FaF F F− + on 1ln C  (data from Fig. 2.4(b))........................29 Figure 2.9 Nyquist plot of three plates as working electrodes: bare PZT, BSA adsorbed, and BSA adsorbed after vibration surfaces in PBS. The curves can be modeled with electrical equivalent circuits ....................................................................................................................31  ix Figure 2.10 Three fitting equivalent electrical circuits, (a) one solution resistance and two parallel RCPE, (b) one solution resistance, one parallel RC and one parallel RCPE in serials, and (c) one parallel RCPE........................................................................................................32 Figure 2.11 Bode plots of BSA adsorbed surfaces: (a) Z’ v.s. log(frequency) and (b) phase angle v.s. log(frequency) ...................................................................................................................34 Figure 2.12 Fluorescence image of PZT plate after vibration 5 min under 10 Vpp at resonance. The inset is the simulation of a PZT plate vibrating at resonance frequency. The highest vibration amplitude is located at the center of PZT plate ........................................................38 Figure 2.13 The dependence of BSA desorption and modified flow velocity...............................38 Figure 2.14 Modified fluorescence intensities of surfaces before and after AC application. The error bar is the standard deviation of 6 measured data ............................................................39 Figure 2.15 Schematic diagram of rotating plate system...............................................................40 Figure 2.16 Fluorescence image of a copper plate after rotating for 5 min. The dark region in the center is connecting to a motor for rotation .............................................................................40 Figure 2.17 Fluorescence intensity of Ti coated plate surfaces at different applied shear stresses on one BSA molecule...............................................................................................................42 Figure 2.18 Calculated protein-surface interaction energy between a BSA molecule and a negatively charged PZT surface (pH = 7.4). In the proximity of the surface, the total attractive energy of VDW and hydrophobic interaction can be compensated by the repelling electrostatic interaction ............................................................................................................47 Figure 3.1 Schematic diagram of the proposed piezoelectric membrane as an implantable sensor coating. Proteins can be desorbed by surface charge and acoustic streaming force generated by vibration. .............................................................................................................................55 Figure 3.2 (a) The experimental setup (b) Front side view of a micromachined PZT plate (3000μm x 1000μm x 500μm) on membrane with an electrical wire bonded onto the plate surface (c) The backside view of the Si membrane. A thin SiO2 (1μm) layer remained on the surface. In the following experiments, the proteins are adsorbed onto the backside of the Si membrane (with a SiO2 surface). .............................................................................................60 Figure 3.3 The simulation and experimental vibration spectrum of the PZT plate/silicon membrane. The vibration at 308, 320, 500, and 575 kHz correspond to 1.5, 2.5, 3.5 wavelength bending mode, and longitudinal mode, respectively. ...........................................61  x Figure 3.4 The simulation and experimental vibration amplitude distribution across the Si/SiO2 membrane along the membrane center line. X axis is given as the distance from the center of membrane to the left periphery. The half length of the membrane is 1000μm........................62 Figure 3.5 Kinetic adsorption isotherms of BSA and IgG on a SiO2 surface. The protein solution concentrations were 2 mg/ml for BSA and 0.1 mg/ml for IgG. After each different protein incubation time, the surface was washed with PBS 5 times before the fluorescence intensity measurements. Equation (A.1) was used to fit the data...........................................................63 Figure 3.6 Comparison of relative fluorescence intensities, which linearly depend on protein adsorption quantity, with and without PZT vibration. The vibration was initiated by a 10 Vpp AC signal at 308 kHz for 5 min. The proteins were incubated in 2 mg/ml BSA and 0.1 mg/ml IgG for 1 hr. Control 1 and 2 are the surfaces incubated in BSA and IgG solution, respectively. For Vib 1 and Vib 2, BSA and IgG adsorbed surfaces respectively were used. The fluorescence intensity at every condition was measured after washing the surface after that condition by PBS for 5 times. ...........................................................................................64 Figure 3.7 Relative fluorescence intensities, which linearly depend on protein adsorption quantity, show a strong dependence on the vibration amplitude of the membrane. Protein desorption cannot be effective if the surface vibrated at off-resonance frequency or under low voltage AC signals application. In all the experiments, samples were incubated in BSA (2 mg/ml) for 1 hr and vibration conditions were applied for 5 min. Conditions for Vib 1, Vib 2, and Vib 3 were 308 kHz 10 Vpp, 308 kHz 1 Vpp, and 400 kHz 10 Vpp, respectively. The fluorescence intensity at every condition was measured after washing the surface after that condition by PBS for 5 times. ..................................................................................................65 Figure 3.8 Comparisons of relative fluorescence intensities, which linearly depend on plasma adsorption quantity at (a) different times without and with PZT vibration, (b) 4 hr without and with PZT vibration in different media and time intervals. The conditions in (a) are Control- incubation in plasma protein solution for 1 hr; Vib 1-vibration application in PBS for 5 min after incubation; Adsorption 1-second plasma protein incubation for 20 min after Vib 1 on the sample sample; Vib 2- vibration application in PBS for 5 min after Adsorption 1 on the sample sample. The conditions in (b) are Control- Incubation in plasma protein solution for 4 hr; Vib 1- vibration application in plasma prtein solution for 5 min; Vib 2- vibration application in PBS for 5 min after Vib 1; Vib 3- vibration application in PBS for 10 min after Vib 2; Vib 4- vibration application in PBS for 20 min after Vib 3. All vibrations are  xi applied by a 10 Vpp AC signal at 308 kHz. The plasma protein concentration was 0.3 mg/ml. The fluorescence intensity at every condition was measured after washing the surface after that condition by PBS for 5 times. ...........................................................................................68 Figure 4.1 (a) XPS spectra (binding energy from 0 to 600 eV) for (1) control (bare surface) and surfaces incubated in (2) 2x10-7 mg/ml, (3) 2x10-5 mg/ml, (4) 2x10-3 mg/ml, (5) 1, (6) 6 mg/ml PEG initial concentration solutions. (b) Variant atomic percentage of elements C, O and Au on surfaces incubated in different PEG initial concentration solutions. Increases of elements C and O and decrease of element Au indicate increase of PEG moieties on the surface. .....................................................................................................................................79 Figure 4.2 High resolution C 1s peak scan of (1) control (bare surface) and surfaces incubated in (2) 2x10-7 mg/ml, (3) 2x10-5 mg/ml, (4) 2x10-3 mg/ml, (5) 1, (6) 6 mg/ml PEG initial concentration solutions. The increase of C-O and decrease of C-C peak intensity is shown with the increase of PEG initial concentration ........................................................................82 Figure 4.3 (a) The log-log plot of integral intensity of peak Au4f1/2 with PEG initial concentration, (b) The linear-log plot of the corresponding PEG grafted thickness and PEG concentration. The inset in Figure 4.3(b) is the linear-linear plot of PEG thickness and PEG concentration, which shows the PEG adsorption behavior and can be fitted by Langmuir isotherm with R2=0.9387. ........................................................................................................83 Figure 4.4 The infrared spectra of PEG grafted surfaces (with gold-coated PZT substrate), incubated in (1) 6 mg/ml, (2) 0.1 mg/ml, and (3) 2x10-2 mg/ml PEG initial concentration solutions, measured by ATR-FTIR. The C-O-C and –CH2 peaks are characteristic peaks of PEG. The intensity of these peaks increases with the surface incubated in increased PEG initial concentration .................................................................................................................84 Figure 4.5 Relative fluorescence intensities, which linearly depend on BSA adsorption quantity, of surfaces incubated in different PEG initial concentration solutions with or without application of vibration. * indicates significant statistic differences (p<0.05) ........................86 Figure 4.6 Fluorescent photos of patterned surfaces (a) before vibration and (b) after vibration. The brighter area is gold surface without grafted PEG and the darker area is the surface with grafted PEG incubated in 6 mg/ml PEG solution for 16 hr .....................................................88  Figure 4.7 Relative fluorescence intensities, which linearly depend on IgG adsorption quantity, of surfaces incubated in different PEG initial concentration solutions with or without application of vibration. * indicates significant statistic differences (p<0.05). .......................90  xii Figure 4.8 Relative fluorescence intensities, which linearly depend on plasma protein adsorption quantity, of surfaces incubated in different PEG initial concentration solutions with or without application of vibration. * indicates significant statistic differences (p<0.05)...........91 Figure 5.1 Representation of the structures of polymeric films on gold surface (a) linear mPEG monothiol linear and (b) HPG monothiol. .............................................................................107 Figure 5.2 ATR-FTIR spectra of gold surfaces coated with (a) monothiol-functionalized linear mPEG-5000 at (1) 6 mg/ml, (2) 2 x 10-2 mg/ml, (3) 2 x 10-3 mg/ml, and (4) 2 x 10-5 mg/ml polymer incubation concentration and (b) monothiol-functionalized HPG-SH-L at (1) 6 mg/ml, (2) 1 mg/ml, (3) 2 x 10-2 mg/ml, and (4) 2 x 10-3 mg/ml polymer concentration. (c) Comparison of linear mPEG-5000, HPG-SH-L, and HPG-SH-H at 6 mg/ml solution concentration. Polymer films were produced by incubating the gold surface in polymer solution for 16 h .....................................................................................................................109 Figure 5.3  High resolution C1s scan from XPS spectra of surface grafted HPG-SH-L at (1) 6 mg/ml, (2) 1 mg/ml, (3) 2x10-3 mg/ml polymer incubation concentration, and (4) bare gold- coated substrate......................................................................................................................110 Figure 5.4 Effect of incubation time on the dry thickness of linear mPEG-5000 film measured by ellipsometry on the gold surface at 6 mg/ml polymer concentration.....................................113 Figure 5.5 Effect of incubation concentration and type of polymers on the (a) thickness and (b) graft density of polymer films on gold surface. Polymer films were produced by incubating the gold surface in polymer solution for 16 hr.......................................................................114 Figure 5.6 AFM images showing the surface morphology of bare gold surface (a), linear mPEG- 5000 grafted surfaces (b), low molecular weight HPG-SH-L grafted surfaces (c), and high molecular weight HPG-SH-H grafted surfaces (d) at 6 mg/ml incubation concentration. For a given sample the 3D image (upper) and section plot (lower) are shown...............................117 Figure 5.7 Fluorescence photographs of BSA-adsorbed (a) bare gold and (b) linear mPEG-2000-, (c) linear mPEG-5000-, (d) low molecular weight HPG-SH-L-, and (e) high molecular weight HPG-SH-H-grafted surfaces. Polymer films were produced by incubating the gold surface in polymer solution at 6 mg/ml for 16 h. (f) Effect of the graft density on the BSA adsorption of mPEG-, HPG-SH-L-, and HPG-SH-H-grafted surfaces..................................120 Figure 5.8 Fluorescence photographs of IgG-adsorbed (a) bare gold and (b) linear mPEG-5000-, (c) low molecular weight HPG-SHL-, and (d) high molecular weight HPG-SH-H-grafted surfaces. Polymer films were produced by incubating the gold surface in polymer solution at  xiii 6 mg/ml for 16 h. (e) Effect of the graft density on the IgG adsorption of mPEG-, HPG-SH- L-, and HPG-SH-H-grafted surfaces......................................................................................121 Figure 6.1 Illustration of a PDMS based microfluidic device for the selective capture of relevant proteins...................................................................................................................................129 Figure 6.2 (a) Representation of the chemical structure of HPG containing functional succinimidyl ester groups (molecular weights 2.5 kDa or 8 kDa) and (b) the reaction scheme involving the grafting process of functionalized HPG to amine modified PDMS surfaces ..137 Figure 6.3 (a) Contact angles of native, amino (NH2), HPG-P1, HPG-P2, and HPG-P3 modified PDMS surfaces (error bars extrapolated from standard deviations from total of 8 measurements), and (b) FTIR spectra of HPG-P1, HPG-P2, and HPG-P3 modified PDMS surfaces. The surfaces were dried before contact and FTIR measurements. .........................138 Figure 6.4 Values of measured μEOF of native, amino (NH2), HPG-P2, and HPG-P3 modified PDMS microchannels. The applied electric fields for native, HPG-P2, and HPG-P3 modified microchannels are 4*103 V/m. The applied electric field for amino modified microchannels is 16*103 V. The average velocity and error bars are from 3 independent measurements .......140 Figure 6.5 (a) Fluorescence intensities of native, amino (NH2), HPG-P1, HPG-P2, and HPG-P3 modified PDMS surfaces and microchannels after FITC labeled BSA adsorption (the error bars in the figure are the standard deviation of 6 measured data points). The BSA solution was incubated with PDMS surfaces for 1 hr at room temperature. After adsorption, PDMS surfaces and microchannels were washed with PBS buffer 5 and 3 times, respectively. (b) Fluorescence photos of the native, amino (NH2), HPG-P1, HPG-P2, and HPG-P3 modified PDMS microchannels shown in Fig. 6.5 (a). The white dash lines are the boundaries of the microchannels ........................................................................................................................142 Figure 6.6 (a) The topography SU8 mold on a silicon substrate measured by an optical profiler (part of the ion exchange column, scale is shown in the figure). The SEM photos of (b) x 400, and (c) x 1000 PDMS pillar arrays. The dimensions of the cross section of the pillar arrays are 15 μm x 15 μm .................................................................................................................144 Figure 6.7 Fluorescence images of columns (a) filled with avidin, (b) incubated with avidin for 3 min and washed 3 times  with PBS buffer (the fluorescent intensity is 21.17 ± 2.78 % compared to (a)), (c) filled with BSA, (d) incubated with BSA for 3 min and washed 3 times with PBS buffer (the fluorescent intensity is 8.84 ± 1.23 % compared to (c)). The  xiv concentration of BSA and Avidin was 1 mg/ml, and the measurement was performed at room temperature ............................................................................................................................146 Figure 6.8 Fluorescence images of columns at (a) 0 min after incubation with protein mixture (0.1 mg/ml avidin and 0.1 mg/ml BSA), (b) After 3 min incubation and 3 times wash with PBS buffer and (c) fluorescence intensity ratio of avidin/BSA at the onset and after 3 min of incubation (followed by washing 3 times with PBS buffer) with the protein mixture ..........147 Figure 6.9 Selectivity (left) of columns and capture efficiency (right) of avidin with various concentrations (mg/ml) of avidin and BSA mixtures (wt ratio 1:1). The selectivity was defined in terms of the relative quantities of avidin to BSA based on the fluorescent intensity. The capture efficiency of avidin is relative total capture of proteins after 3 min..................149 Figure 6.10 In-situ selectivity of columns vs. time at two protein concentrations (0.02 and 0.066 mg/ml) of avidin and BSA mixtures (wt ratio 1:1) with or without pillar arrays. Incubations of columns with pillar arrays with similar concentrations of the avidin and BSA mixture showed higher selectivity.......................................................................................................150     xv LIST OF ILLUSTRATIONS Illustration 4.1 The depiction of a cross-section of Au/Cr/SiO2/Ti/PZT/Ti plates, and the scheme of grafting PEG on plain or patterned plates ...........................................................................75 Illustration 5.1 Synthetic route for mono thiol funtionalized hyperbranched polyglycidols.......105    xvi LIST OF SYMBOLS AND ABBREVIATIONS x g oC o A 3D [c] [εr] [eT] Α β2 = ρf/2η φ φ γS γL γi γ δ ε0 ε θ θl κS 1/к μl η μEOF ν ξ ρ σ ωτ Gravity (m/s2) Centigrade Angtrong (10-10 m) Three dimension Stiffness matrix (1010 N/m2) Dielectric matrix Piezoelectric matrixT (C/m2) κs(1-( Pυ / Fυ )2)1/2 ρf/2η Quantum efficiency Interaction energy of one molecule with its neighbor (m2*kg/s2) Surface tension of solid-air  (N/m = kg/s2) Surface tension of liquid-air Surface tension of solid-liquid Wavelength (nm) Evanescent decay length (nm) Relative permittivity of free space (F/m = s4A2/m3kg) Permittivity Angle (o) Surface coverage ratio Propagation constant Debye length (nm) Microliter Dynamic viscosity (cPose = 10-2 Pose = 10-3 Pa·s = 10-3 kg·m−1·s−1) Mobility of EOF (10-4 cm2/ V·s) Kinematic viscosity (cStoke = 10-2 stokes = 1 mm²/s) Zeta potential (mV) Density (kg/m3) Surface charge density (C/m2) Wall shear stress (dyn/cm2)  xvii Pυ Fυ ψo ω Γ Γl Γlmax Σ Ω A Amp AC AFM ATR-FTIR Ag Al Au Β BE BOE BSA C0 C (unit) C CPE c Cu cm Do D DAPI DC Phase velocity of sound in the membrane (m/s) Velocity of sound in the fluid (m/s) Surface potential (mV) Radial velocity of the rotating disk (rpm) Polymer surface coverage (mg/m2) Quantity of adsorbed particles Maximum adsorbed amount of particles Chain density (chains/nm2) Ohms (V/A = m2·kg/s3·A2) Area (m2) Amplitude of Vibration (nm) Alternating Current Atomic Force Microscopy Attenuated Total Reflectance - Fourier Transform Infrared Spectroscopy Silver Aluminum Gold Bending stiffness of membrane (N/m2) Electron Binding energy (eV) Buffered Oxide Etch Bovine Serum Albumin PZT capacitance at resonance frequency (F) Concentration (mole/l) Carbon Constant Phase Element (F) Capacitance Copper Centimeter Characteristic decay constant of short-range hydrophobic interaction (nm) Cut-off distance (nm) 4',6-diamidino-2-phenylindole (C16H15N5) Direct Current  xviii DES DI DHB DMF DNA DTNB Da d d0 dyn E Eo Es Eso EsP Ee Eh ET EV EDL EOF eV Fe FV F dF F∞ maxF∞ FITC FPW f g Diethylsilane (C2H5)2SiH2 ) De-Ion water 2,5-dihydroxybenzoic acid ((HO)2C6H3COOH) Dimethylformamide ((CH3)2NCOH) Deoxyribonucleuic acid 5,5 ′ -dithiobis (2-nitrobenzoic acid) ([-SC6H3(NO2)CO2H]2) Atomic mass unit Thickness or distance (nm) The distance of dye to metal surface with 50 % quantum efficiency Force unit (10-5 N) Electric field (V/m) Standard potential of the half reaction Empty sites of the surface Total sites of the surface Adsorbed complex on the surface Electrostatic interaction energy (J) Hydrophobic interaction energy (J) Total interaction energy (J) VDW interaction energy (J) Electrostatic Double Layer Electroosmotic flow Electron volt (1.6x10-19 V) Electrostatic interaction force (N) VDW interaction force (N) Farady   (C/V = s4·A2/m2·kg) Fluorescence intensity of dye at distance d from the metal surface Fluorescence intensity of dye at infinite from the metal surface Monolayer Fluorescence intensity of dye at infinite from the metal surface Fluorescein isothiocyanate Flexural Plate Wave Frequency (Hz) Gram  xix 1H NMR H HMDS HPG HPG-SH HPG-SH-H HPG-SH-L HSA Hz h hr I IgG J K KF Kα KBr K+CH-Ph2 K2HPO4 k k1 kHz kV LC LDV l lbs M MALDI-TOF MEMS MW Hydrogen-1 Nuclear Magnetic Resonance Hamaker constant Hexmethyldisilane (O[Si(CH3)3]2) Hyperbranched  polyglycerol Mono thiol functionalized HPGs High MW mono thiol functionalized HPGs Low MW mono thiol functionalized HPGs Human Serum Albumin Hertz (1/s) Thickness of the layer (nm) Hour(s) Current (ampere) Anti-mouse goat Immunoglobulin G Joule (Nm = m2 kg/s2) Langmuir Equilibrium constant Frumkin Equilibrium constant L →K orbital transition light Potassium bromine Diphenyl methylpotassium Dipotassium hydrogen phosphate Boltzman constant Rate constant of adsorption Kilohertz Kilovolt Liquid Chromatography Laser Doppler Vibrometer Liter Pounds Mass per unit length of the membrane Matrix-Assisted Laser Desorption/Ionization-Time Of Flight Micro-Electro-Mechanical Systems Molecular Weight (g/mol)  xx MHz Mg m mPEG mM mg min ml mm NA N NaCl NaH2PO4.H2O nm O P PBS PDMS PEB PECVD PEG PMMA PVDF PZT p pN pH poise RΩ R RCA1 r Megahertz Magnesium Meter Methoxy polyethylene glycol thiol (CH3O-(CH2CH2O)n-(CH2)4-SH) Milimolar Miligram Minute(s) Mililiter Milimeter Avogadro’s number Newton (m·kg/s2) Sodium chloride Sodium phosphate monobasic monohydrate Nanometer Oxygen Particles Phosphate Buffered Saline Polydimethylsiloxane ((H3C)3SiO[Si(CH3)2O]nSi(CH3)3) Post Exposure Bake Plasma Enhanced Chemical Vapor Deposition Polyethylene glycol ((HO)CH2[CH2CH2O]nCH2(OH)) Polymethyl methylacrylate ((C5O2H8)n) Polyvinylidene fluoride ((CH2CF2)n) Lead Zirconate Titanate (PbZrTiO3) p value for significant statistical differences PicoNewton Measure of the acidity or alkalinity of a solution Viscosity unit (1 mPa·s = 10-3 Pa·s = 1 cP = 10-2 poise) Electrolyte resistance between the working and reference electrodes Sphere radius of protein (nm) Standard clean-1, clean with mixture of H2O2:NH4OH:H2O = 1:1:5 Radial distance of the rotating disk (cm)  xxi rpm S SB SEM SOI Si Si3N4 SiO2 Stokes s (unit) s, p, and f T (unit) T TMAH TRITC Ti UV-Vis spectroscopy u VEOF Vpp V VASE VDW W XPS Z’ Z’’ z  Rotate per minute Sulfur Soft Bake Scanning Electron Microscope Silicon-On-Insulator Silicon Silicon nitride Silicon dioxide 1 cSt = 10-2 stokes = 1 mm²/s Second(s) Electron orbits Kelvin temperature (K) In-plane tension of membrane (N/m2) Tetra methyl ammonium hydroxide ((CH3)4NOH) Tetramethyl rhodamine isothiocyanate Titanium Ultraviolet-Visible spectroscopy Flow velocity (m/s) EOF velocity (m/s) Peak-to-peak voltage (V) Volt   (m2·kg/s3·I) Variable Angle Spectroscopic Ellipsometer Van der Waal Watt (J/s = m2 kg/s3) X-ray Photoelectron Spectroscopy Real impedance (Ω) Imagery impedance (Ω) Out of plane distance (nm)   xxii ACKNOWLEDGEMENTS First of all, it is my pleasure to thank Dr. Mu Chiao, my research advisor, for creating a stimulating, friendly and open environment in which I pursued my doctoral work. His continuous support and unabated interest with all aspects related to my present study, and his thoughtful guidance and encouragement to inspire new ideas, are invaluable. I am very thankful for his mentorship and friendship. I am also very grateful to Dr. Jayachandran N Kizhakkedathu, who introduced me to the basic polymer knowledge and techniques and provided continuous technical support throughout my study. It is a really nice and comfortable experience to cooperate and discuss with him. My sincere thanks go to Dr. Kainthan Rajesh Kumar, Dr. Yuquan Zou, Dr. Nicholas A. A. Rossi, and Dr. Johan Janzen for their help on polymer synthesis and surface analysis equipments. Their kindly supports help me to be familiar with polymer and surface science. I also thank the members of my supervisory committee, Prof: Mohamed S. Galada (Mech. Eng.), Prof: Rizhi Wang (Material Eng.), and Prof: John D. Madden (Elec. Eng.) for keeping in touch with my work and offering valuable suggestions. I also learned a lot from students in the laboratories: Mr. Billy Siu, Ms. Nazly Pirmoradi, Mr. Farid khan, Mr. Reza Rashidi, Mr. Hadi Mansoor, Mr. Ki- Young Song, Mr. Tung Siu, and Ms. Jiamei Bai. Many thanks go to your help and friendship. The financial support from Natural Science and Engineering Research Council of Canada is greatly appreciated. I would like to thank my beloved wife, Cathy. I am fortunate to share my life with her these years. She created a home for us and had being a lovely presence to me. Many thanks also go to my parents, my parents in law. It is your love and encouragement that allow me to overcome many challenges. A special thank go to many dear saints in the church of Vancouver and Seattle, I am blessed to share belief with you and have you pray with me in many situations. Finally, I would like to dedicate this thesis to my new born daughter, Jessy, who is one of the most precious gifts that I ever have!!     xxiii CO-AUTHORSHIP STATEMENT I hereby declare that this thesis incorporates materials that are results of joint research. The hyperbranched polyglycerol (Chapter 5) and its derivative (Chapter 6) were synthesized and characterized by Dr. Kainthan Rajesh Kumar and Dr. Nicholas A. A. Rossi, respectively. The AFM measurements and analysis in Chapter 5 of the thesis were collaborated with Dr. Yuquan Zou under the supervision of professor Jayachandran N. Kizhakkedathu. The plasma proteins in Chapter 3 and 4 were labeled with Alexa FluorTM 488 by Ms. Yevgeniya Le. My main contributions are on following areas: • Key ideas • Experimental designs • Fabrication and simulation • Data analysis and interpretation • Manuscripts preparation   1 CHAPTER 1 INTRODUCTION 1.1 Overview The emergence of microelectro-mechanical systems (MEMS), with the integration of microelectronic circuitry into micromachined mechanical structures, to produce completely integrated systems, has the advantages of small size, low cost, and reliability. Recently, BioMEMS, which applies MEMS into biological environments in vitro or in vivo, has attracted much attention because it has led to the emergence of a new class of therapeutic medical devices. Among the BioMEMS are biosensors, drug delivery systems, microfabricated devices for electrophoretic separation, biological analysis systems, and so on. These kinds of devices, made through micromechanics or which incorporate microengineered components are envisioned to command a significant market share long into the future. For the success of these implantable medical devices, the view of biocompatibility of materials has gone from ‘the host tolerates the device’ to ‘the device tolerates the host’ and vice versa [1,2]. For example, a subcutaneous glucose sensor is comprised of immobilized enzymes, electrodes, and a semi-permeable membrane. The membrane acts as a protective layer that keeps the enzymes and electrodes from being attacked by the immune system while allowing glucose from the interstitial space to diffuse through. To be successful, the semi-permeable membrane must not induce pathogenic reactions in the host. On the other hand, the glucose permeability of the membrane must be maintained at all times. Materials that are widely used as biomaterials for BioMEMS include silicon, metals, ceramics, carbons, glasses, modified natural biomolecules, synthetic polymers, and composites of various material types. Exposing materials to a biological environment by implanting a  2 biomedical material/device can elicit a number of reactions such as protein adsorption and cell adhesion to the surface of the device. Therefore, predicting the results of possible interactions between biological systems and materials is an important in the design of medical devices. The goal of this thesis is to study the interactions between surfaces and biological molecules, such as albumin, immunoglobulin, and plasma proteins for attenuating the nonspecific adsorption of proteins to the surface (“biofouling”). Physical methods, such as surface charging and vibrations of surface, and chemical methods, such as affinity coating and covalent grafting of linear or branched polymers, are applied to macro and microscale devices. Polymers developed for attenuating nonspecific protein adsorption are further modified to functionally capture protein (specific adsorption) for developing biomedical liquid chromatography devices.  1.2 Literature Review Two particularly important issues in the biocompatibility of surfaces are: 1) thrombogenicity, which involves blood coagulation and the adhesion of blood platelets to biomaterial surfaces, and 2) the fibrous-tissue encapsulation of biomaterials that are implanted in tissues. The initial response of a biological system to material surfaces is determined by the proteins and cellular interactions that can set the stage for inflammatory or thrombotic responses. Chronic inflammatory responses, resulting in thick fibrous capsules, are undesirable since they can cause damage of surrounding tissue and the failure of the device. More generally, all proteins are known to have an inherent tendency to adsorb very rapidly and tightly onto surfaces, which strongly influences the subsequent interactions of various types of cells with the surfaces. Thus, to improve the biocompatibility of microdevices, the focus is on reducing protein adsorption and cell adherence.  3 Proteins are biological macromolecules constructed for specific and unique functions. They are high molecular weight polyamides comprised of up to 20 different amino acids. When an implanted material or device is in contact with the blood, a layer of proteins surrounds the device and is adsorbed on the solid surface within seconds. Blood consists of plasma (mostly water and proteins), red and white cells, and platelets. The plasma proteins are mainly albumin (50~60%), immunoglobulins (15~35%), fibrinogen (~3%), and other regulatory proteins [3,4]. Albumin has several important functions and is the most abundant protein in the plasma. Its molecular weight is 66,000 Da, and at neutral pH, the number of charged groups is roughly 200. Fibrinogen plays a special role with respect to blood-material interactions, such as in blood coagulation and platelet aggregation. It is categorized as a cell-adhesive protein and has a strong tendency to adsorb onto various surfaces. Immunoglobulin, on the other hand, has three different forms: alpha, beta, and gamma, with the gamma globulins (called IgG) serving as antibodies to help protect the body against foreign objects. Antibodies are classified in five types: IgG, IgM, IgA, IgD, and IgE. IgG is the most common antibody, and is present mostly in the blood and tissue fluids. In this thesis, we evaluate the adsorption behavior of the most common and key proteins: albumin, immunoglobulin, and plasma proteins. Reducing their adsorption is a key factor contributing to the success of microdevices. Different methods for reducing protein adsorption are reviewed in Chapters 2 to 6, and in this chapter, the protein-surface interaction is discussed, and research directions for reducing protein adsorption are reviewed. The fundamental classes of nonspecific interactions, which are important in protein adsorption, involve four major types [5,6]. First, the van der Waals (VDW) interaction includes momentary attractions between molecules, diatomic free elements, and individual atoms and others, besides those due to covalent bonds or the electrostatic interaction of ions, that occur with  4 one another or with neutral molecules. The VDW interaction is the result of the temporary fluctuating dipoles of molecules and is correlated as the molecules come closer to originate an attractive force. A VDW force always exists between two molecules or two surfaces, and can be described according to a power-law dependence on the separation between protein and surface. Second, electrostatic interaction is due to the attraction or repulsion of two or more groups carrying net charges. In addition to the VDW interaction, the electrostatic double layer (EDL) force is the other major force existing between two molecules in aqueous solution. VDW and EDL forces are actually described in the DLVO theory, developed by Derjaguin & Landau (1941), and Verwey & Overbeek (1948) for colloid stability. The electrostatic force between two similarly charged molecules is repulsive and decays exponentially with the separation. Third, hydrophobic interaction is a large entropically-driven process attributed primarily to the water structure effect occurring adjacent to hydrophobic interfaces. The hydrophobic force is an attractive interaction in water between inert, non-polar molecules or surfaces, and is much larger than the VDW force occurring between non-polar molecules or surfaces. The magnitude of the hydrophobic attraction depends on the hydrophobicity of surfaces, which is related to the interface energy. Finally, steric interactions occur at many biological surfaces where flexible polymer-like hydrophilic groups (i.e., polysaccharides, ligand groups at the ends of tethers, and lipids with large headgroups) are exposed to the aqueous phase. Steric forces arise between the surfaces containing these flexible polymer coils when they approach each other in an aqueous solution, the main property explaining a polymer’s antifouling performance. Among these nonspecific interactions, hydrophobic and electrostatic interactions are said to be the most important; the former is significant when a device surface is more hydrophobic and the later dominates when highly charged groups are present at the contact surfaces. The  5 magnitude of these interactions can be experimentally measured by atomic force microscopy or a surface force apparatus [7]. The magnitude of these interactions is in the order of several pN [8,9]. Using human serum albumin as an example, the total repulsive force and the attractive hydrophobic force between albumin molecule and the surface is 7.6*10-11 and 9*10-11 (N), respectively, measured by the surface force apparatus [7]. On the other hand, interactions such as “lock-and-key” and recognition interactions are defined as specific interactions. They are not catalogued as reasons for “biofouling” but are often used as the components for biosensor and affinity chromatography. To achieve the goal of anti-biofouling, two forces need to be cooperating. One is a vertical force to overcome the total attractive interactions mentioned previously, that occur between protein and the surface, and the other is a horizontal force to carry away any desorbed protein so that it will not adsorb onto the surface again. The mechanism of ultrasound cleaning micro-sized particles was discussed by Qi [8], and then was applied to MEMS-based filtering, as discussed by Caton [10]. The mechanism of attenuating protein adsorption will be discussed in detail in Chapter 2. The polymer coating used for attenuating protein adsorption is reviewed in Chapters 4 and 5. In this introduction, instead of describing the properties necessary for resisting protein adsorption, the functionality of polymers are discussed with regards to their specific binding. Surfaces can be coated with polymers to produce a relatively simple environment; for example, a charged surface can be created by simply changing the terminal functional groups of the polymer, such as -CH3, - COOH, -NH2, or -OH [11,12]. Thus, polymers with terminal -CH3, -OH, -COOH, and -NH2 groups can provide hydrophobic, hydrophilic, negative charge, and positive charge surfaces, respectively. Applications that are derived from this method include protein or cell patterning  6 [13], biosensors [14], chromatography [15], and so on. For example, Falconnet applied a lift-off method to covalently bind proteins with polymer (with an RDG sequence so that cells can bind to the proteins), and then the cells can bind to the proteins to produce the cellular pattern [16]. The possibility of modifying the terminal groups of the polymer has led to more interest in different types of branched polymers, as discussed in Chapter 5. Because many terminal groups can exist on a given polymer molecule, with multiple functionalities, further modifications are possible. For example, in liquid chromatography, polymers with various terminal groups are used as analytical tools (mentioned in Chapter 6). Instead of silicon-based materials, polymer- based materials may be useful for device fabrication. The polymer-based materials have many advantages over the silicon-based materials, such as in rapid development to prototype, short time to result, lower cost, higher sensitivity, and easier mass production [17-19]. Moreover, polymer-based devices have already shown their usefulness as bioanalytical tools because of their ability to form complex valves, separation columns, pumps, mixers, etc. [20-22]. The concepts presented in Chapter 6 may further be applied to the development of tools for proteomics, an interesting new discipline, especially in the post-genome era, because of the close relationship between disease and proteins. In the literature, the main idea is that proteins, not genes, are at the business end of biology [23]. Proteome profiling, which separates and identifies the varieties of proteins in a given biological sample, is one of the fundamental, yet critical, tasks for the success of proteomics [24]. Furthermore, enhanced performance chromatography could be used as a proteome profiling tool.  1.3 Organization of the Thesis This thesis is organized with a general introduction and literature review in Chapter 1.  7 In Chapter 2, the mechanism of attenuating protein adsorption is described on a theoretical basis, and one route for reducing protein adsorption experimentally, is presented. In Chapter 3, the concept in Chapter 2 is applied to an MEMS-based membrane to evaluate its antifouling performance. These two chapters examine the physical method for attenuating protein adsorption. In Chapter 4, a physical method is combined with the use of a chemical polymer coating, and the improved performance for resisting protein adsorption is compared to either the physical or the chemical method alone. The possible mechanism for antifouling, including the polymer and protein patterns, is being developed and will likely be the subject of future work. In Chapter 5, two different polymer structures (linear and branched) are described for their ability to resist protein adsorption. Branched polymer structures are of great interest in current research, and in this chapter, the synthesis and coating of branched polymers on surfaces, using affinity methods, are described in detail. Branched polymers show an even better antifouling performance compared to linear polymers of similar molecular weight. In Chapter 6, the possibility of grafting branched polymers in a PDMS-based microfluidic device is examined for a liquid chromatography application. The modified polymer is able to separate proteins of interest, and other applications may develop in the future, based on this concept. Finally, in Chapter 7, the conclusions for the work presented in this thesis are drawn, in light of current research in related fields and from possible future work.   8 1.4 Bibliography [1] M. Schlosser and M. Ziegler, in D. M. Fraser (Eds.), Biosensors in the Body: Continuous in vivo Monitoring, 1st ed., John Wiley & Sons, Chichester, 1997, Chapter 5. [2] D. M. Fraser, in D. M. Fraser (Eds.), Biosensors in the Body: Continuous in vivo Monitoring, 1st ed., John Wiley & Sons, Chichester, 1997, Chapter 1. [3] B. Blombäck and L. A. Hanson, Plasma Proteins, Pitman Press, Bath, 1975, p. 17-21. [4] J. L. Brash, in J. L. Brash and T. A. Horbett (Eds.), Proteins at interfaces: Physicochemical and Biochemical Studies, American Chemical Society Press, Washington D.C., 1987, Chapter 1. [5] D. Leckband and J. Q. Israelachvili, Rev. Biophys., 34 (2001) 105-267. [6] J. Israelachvili, Intermolecular & Protein-surface Forces, 2nd ed., Academic Press, London, 1991, p. 176-212. [7] D. Leckband, Annu. Rev. Biophys. Biomol. Struct., 29 (2000) 1-26. [8] Q. Qi and G. Brereton, IEEE Transactions on Ultrasonics, Ferroelectrics and Frequency Control, 42 (1995) 619-629. [9] E. Blomberg, P. M. Claesson, and Johan C Floberg, Biomaterials, 19 (1998) 371-386. [10] P. Caton, Microfiltration and Flexural Plate Wave Devices, PhD thesis, University of California at Berkeley, 2001. [11] S. Chen, L. Liu, J. Zhou, and S. Jiang, Langmuir, 19 (2003) 2859-2864. [12] L. Liu, S. Chen, C. M. Giachelli, B. D. Ratner, and S. Jiang, J. Biomedical Materials Research, 74A (2005) 23-31. [13] R. Singhvi, A. Kumar, G. P. Lopez, G. N. Stephanopoulos, D. I. Wang, G. M. Whitesides, and D. E. Ingber, Science, 264 (1994) 696-698. [14] G. D. Meyer, J. M. Morán-Mirabal, D. W. Branch and H.G. Craighead, IEEE Sensors Journal, 6 (2006) 254-261. [15] N. Malmstadt, A. S. Hoffman, and P. S. Stayton, Lab Chip, 4 (2004) 412-415. [16] D. Falconnet, Molecular Assembly Patterning by Lift-off at the Micro- and Nanoscale for Applications in the Biosensors, PhD thesis, Swiss Federal Institute of Technology, 2005. [17] D. C. Duffy, J. C. McDonald, O. J. A. Schueller, and G. M. Whitesides, Anal. Chem., 70 (1998) 4974-4984. [18] T. T. Huang, N. S. Mosier, and M. R. Ladisch, J. Sep Sci, 29 (2006) 1733-1742.  9 [19] N. Lion, F. Reymond, H. H. Girault, and J. S. Rossier, Curr. Opin. Biotechnology, 15 (2004) 31-37. [20] S. K. Sia, and G. M. Whitesides, Electrophoresis, 24 (2003) 3563-3576. [21] A. Dodge, E. Brunet, S. Chen, J. Goulpeau, V. Labas, J. Vinh, and P. Tabeling, Analyst, 131 (2006) 1122-1128. [22] T. Thorsen, S. J. Maerkl and S. R. Quake, Science, 298 (2002) 18. [23] Proteomics’s New Order, Nature, 437, 169-170, 2005. [24] S. L. S. Freire and A. R. Wheeler, Lab Chip, 6 (2006) 1415-1423.        10 Chapter 2 Electric Field and Vibration-Assiated Nanomolecule Desorption and Anti- fouling for Biosensor Applications 12  2.1 Introduction Biofouling is considered as one of the greatest challenges in the field of in vivo biosensing [1]. As soon as a device is implanted into the body, a healing response is initiated through a series of complex events that include acute inflammation, formation of granulation tissue, and eventual scar formation [2]. The immediate response is to flood the injured area with blood followed by adsorption of blood proteins onto the device surface, which is followed by activation of the molecular and cellular defense systems [2,3]. The coagulation, complement, immune, and inflammatory pathways can all be initiated by this process as can platelet and white cell activation. The extent to which these responses occur depends on the nature of the protein adsorption, and the specific binding or rejection reactions that occur at the material-blood interface. Usually, an adsorbed protein layer (thickness from 0.5 to 9 μm) accumulates over time. Approximately 21 days after implantation, an avascular fibrous capsule of up to 100 μm will form around the device [4], which eventually causes the biosensor to be “blind,” as the mass transfer between biosensor and the analyte of interest is considerably reduced [4-6]. Passive surface treatment methods using chemical compounds such as PEG have been used to minimize non-specific binding of proteins that produces biofouling. Although these modifications are highly suitable for single proteins, they often fail in the case of highly complex  1 A version of this chapter had been published. Yeh, J. P. Y., Kizhakkedathu, J. N., Madden, J. D. and Chiao, M. (2007) Electric Field and Vibration-assisted Nanomolecule Desorption and Anti-biofouling for Biosensor Applications on Colloid Surface B 59:67-73.   11 mixtures of proteins, for example, blood plasma [7,8], or under in vivo conditions. In addition, for different applications, the density and chain length of PEG must be optimized to minimize non-specific biological interactions. Furthermore, the anti-biofouling ability of PEG also depends on other parameters; for example, more proteins were found to adsorb to a PEG grafted surface at higher temperature [7]. As an alternative to polymer modification, researchers have used nanostructured surface coatings (e.g., nanowires) to reduce protein adsorption [4]. The effects of an active surface coating on biofouling and non-specific protein binding have also been researched. Gluing or depositing a piezoelectric material on the surface of a sensor has been shown to minimize biofouling [9,10]. On an oceanographic sensor, an anti-biofouling process using polyvinylidene fluoride (PVDF) actuator mounted on a transparent glass plate was explained by the electromechanical behavior of PVDF, and this anti-biofouling process was proven experimentally. The application of a mechanical wave, induced by PVDF, was shown to increase the transmittance of a glass plate, as an indication of reduced biofouling [9]. Furthermore, non-specific binding of proteins was reduced by shear waves from high-frequency macro scale quartz resonators with 3.5 W output. At this power output, the specific binding of proteins (signal) remains, while the non-specific binding (noise) is reduced, thus increasing the signal/noise ratio of the biosensor [10]. It was also observed that a magnetic nanowire-assembled surface adsorbs less protein if the nanowires vibrate due to an external oscillating magnetic field [4]. In all of the above-mentioned reports; however, the mechanism underlying the attenuation of non-specific protein adsorption is not fully understood. External fluid flow through the surface has also been shown to attenuate the adsorption of colloids on the surface of biomaterials [5,6,11,12]. Clinically, a glucose sensor was tested in vivo with a continuous saline flow over the sensor surface [5,6]. The glucose sensor was shown to  12 have a longer life than that of regular glucose sensors with no saline flow on the surface, and in this case, the sensor reading drifted less over time. Nevertheless, the difficulty in integrating a flow generator with a biosensor in vivo limits its application. Ultrasound-induced acoustic streaming has also been used to remove μm-sized particles in microfiltration applications [11,12]. The centrifugal (lift) force generated by the vibration was determined to be the main source for overcoming the adhesion force between the particles and the surface. Horizontal forces then carry the desorbed particles away from the surface by the flow. Nevertheless, such forces generated by vibration are not sufficient at the nanoscale, where van der Waals, electrostatic, and hydrophobic interaction forces are more dominant than the centrifugal force [13]. In this chapter, we report a novel anti-biofouling mechanism initiated by a vibrating piezoelectric plate for applications in MEMS-based devices. We propose to assemble a micromachined active piezoelectric membrane with implantable sensors to reduce biofouling. Figure 2.1 illustrates the conceptual application of a micromachined piezoelectric material designed for integration with a biosensor (not shown). The piezoelectric membrane is activated by voltage and induces the flow (acoustic streaming) while being vibrated. By an independent rotating plate experiment, where no electric field is present, we also found that shear stress is not sufficient to remove the adsorbed proteins on a PZT plate. An ideal combination of electric field, coupled with acoustic streaming close to the surface, was found to be effective in removing the adsorbed proteins.     13    Figure 2.1 Schematic diagram of the proposed piezoelectric membrane as an implantable sensor coating. The membrane is composed of two driving electrodes with a piezoelectric material in between. Adsorbed protein can be desorbed by an electric field and carried away by acoustic streaming generated by the vibration.  14 2.2 Materials and Methods 2.2.1 Materials Bovine serum albumin-fluorescein isothiocyanate (BSA-FITC, A9771) and anti-mouse goat immunoglobulin G-FITC (IgG-FITC, F5265) were purchased from Sigma-Aldrich and used without further purification. The lead zirconate titanate plates (PZT, SM10-2525-00) were purchased from Sensor Technology Ltd. and diced into 0.5 cm x 1 cm x 0.5 mm plates for use in the vibration experiments, and into 0.5 cm x 0.5 cm x 0.5 mm plates for use in the experiments without vibration. Silicon substrate with thickness 400μm was purchased from Helitek Company LTD. The resonance frequency of the PZT plates was 16 kHz in a phosphate-buffered saline (PBS). The PZT plates had three kinds of surface coatings: (1) silver (original coating as electrodes of PZT); (2) titanium (0.25 μm, 99.995% purity, Kurt J. Lesker Company, EVMTI45); and (3) gold (50 nm, 99.999% purity, Kurt J. Lesker company, EVMAUXX50G), deposited by an E-beam Evaporator. A rotor was taken from a hard drive and activated by applying either DC or AC voltage. The rotating speed was increased linearly with the applied voltage. The rotating plates were copper sheets purchased from Small Parts, Inc. (SMC-021-B), and the titanium coating (0.3 μm, the same material as used for coating the PZT surface) was deposited by E-beam evaporation.  2.2.2 Methods The diced PZT plates and silicon substrates were wired for connection to a function generator. The rotating plate system was composed of a rotor and plates, and was connected to the function generator. A function/arbitrary waveform generator (Agilent 33220A 20MHz) was used to apply the sinusoid or DC voltage on both sides of the PZT plate and rotor. A Laser Doppler Vibrometer  15 (LDV, Polytec DFV-5000) was used to measure the vibration velocities of the PZT plates and the speeds of the plates in the rotating plate experiments. The PZT vibrating velocities measured by the LDV can be converted to vibration amplitudes by dividing it by 2πf, where f is the frequency. We swept the frequency from 1 kHz to 100 kHz using the function generator and the LDV was used to measure the vibration response. At each frequency, the peak velocity of the sinusoidal velocity signal was represented as the maximum vibration response at that frequency. The resonance frequency was determined by from the peak vibration amplitude-frequency diagram. For the rotating plate system, the LDV measured the rotating speed in frequency (1/s). An oscilloscope (Tektronix TDS420) was used to read the waveforms measured by the LDV. All protein adsorption experiments were done in PBS with 0.15 M NaCl (pH 7.4). Protein adsorption was achieved by incubating PZT plates with protein solutions at different concentrations in 12-well polystyrene plates at 37 oC for 30 min. or 1 hr. Following the adsorption, the PZT surface was rinsed 10 x with PBS to remove loosely bound proteins. The rotation plate was incubated in the protein solution (2 mg/ml) for 1 hr and rinsed with the PBS solution. Vibration experiments were operated in PBS solution, while rotating-plate experiments were operated in PBS and in water at room temperature. Images of the surfaces were taken on a fluorescence microscope (Nikon eclipse TE 2000-U with X-Cite 120 fluorescence illumination system, FITC filter and DS-U1 suite digital camera). Fluorescence intensity of the images can be represented by the grey value of photos after conversion to a grey scale, and is linear with the amount of adsorbed proteins on the surface [14]. The intensity analysis was done using Adobe PhotoshopTM 6.0. The more commonly used quantitative protein assay methods, such as spectroscopy, were not adopted, due to the detection limit requirements in this work [15].  16 The influence of surface tension on protein desorption, which results from the passage through a liquid-air interface was also considered. In this experiment, we took the amount of adsorbed protein after 3 rinse cycles as 100%, and measured the amount of protein remained on the surface after 10, 20, and 40 rinse cycles, with respect to this value. For each case, 104%, 97%, and 89% of the protein remained on the surface, respectively. In addition, three PZT plates with similar backgrounds were chosen to verify the influence of drying in each step before the fluorescence images were taken. The first plate was dipped in PBS and then dried. The second plate was incubated in BSA for 1 hr and then dried. The third plate was incubated in BSA for 1 hr, vibrated in PBS for 5 min, and then dried. All three plates underwent only one drying step and fluorescence images were taken. From the analysis of the fluorescence intensity, no significant influence of surface tension was seen on protein desorption as a result of drying or passing through the air-liquid interface, and therefore, it was not considered further. Electrochemical impedance measurements were taken on a Solartron 1260 frequency response analyzer and 1287 potentiostat. The PZT plate surface was used as the working electrode. A platinum plate was used as the counter electrode with an area 10 x larger than that of the working electrode. Using a large-area electrode is particularly important for measuring impedance, since the impedance of a counter electrode must be negligible compared with the impedance of the working electrode. A standard Ag/AgCl electrode functioned as the reference electrode, and PBS with 0.15 M NaCl was the electrolyte. The measured impedance, represented by the Nyquist plot, was fitted by an equivalent electrical circuit model. The measurement and fitting procedure were done with Z-View 2.6 software from Scribner Associates, Inc.    17 2.3 Results and Discussions Figure 2.2(a) shows that the amount of BSA adsorption on a PZT plate coated with silver was a function of time and BSA concentration. The experimental data is fitted to the Langmuir isotherm (R2>0.9). The BSA concentration used in experiments was set to 2.0 mg/ml since, at this concentration, the adsorbed amount reaches a stable plateau of fluorescence intensity, as shown by the fitted Langmuir isotherm (Fig. 2.2(a)). The typical incubation time was set to 1 hr, to allow sufficient BSA adsorption. Figure 2.2(b) shows fluorescence images of four PZT plates (natural frequency = 16 kHz) with silver surfaces under four vibrating conditions (the applied signal is sinusoid) for 5 min: (1) no vibration; (2) 10 Vpp at 16 kHz; (3) 10 Vpp at 1 kHz; and (4) 3 Vpp at 16 kHz. The quantified fluorescence intensity given in Figure 2.2(c) shows that, under the vibrating condition (2), maximum attenuation of protein desorption occurs. Under this condition, 58 ± 5.5% desorption in BSA occurred in the sample, compared to the sample without vibration (condition 1). Furthermore, for PZT plates that vibrate off the natural frequency (condition 3) or with a smaller applied voltage (condition 4), protein desorptions of 16.3 ± 7.3%, and 15.7 ± 4.7% were observed, respectively. The protein desorption observed here may be attributed to two possible reasons. First, when a voltage is applied to PZT, an accumulation of charge will take place on the surface due to the capacitance character of PZT. When the charge polarity of surface is the same as that of the adsorbed proteins, the electrostatic repelling force between the surface and proteins reduces the adhesive forces at a certain voltage and repels proteins from the surface. Second, when a PZT plate is vibrating, the induced acoustic streaming will create shear stress on the proteins. The acoustic streaming velocity is ~ (vibration amplitude)2, and the shear stress is linear to the streaming velocity [16], which correlates well with our observation that the PZT plates with  18 larger vibration amplitudes have greater BSA desorption (Fig. 2.2(c)). In our case, we believe that a combined effect of both electric-field and vibration is responsible for the protein desorption. We have also looked at the individual effects of electric field and vibration, and discuss this point in a later section. We have also studied protein desorption behavior from other metal surfaces (e.g., titanium) and with other proteins (e.g., IgG). BSA desorption on titanium-coated PZT plates and anti- mouse IgG desorption on silver-coated PZT plates are shown in Figure 2.2(d). With a vibration of 10 Vpp at 16 kHz, 39 ± 5.2% of BSA and 43 ± 9.7% of IgG were removed from the titanium and silver surfaces, respectively. These results show that vibration-induced desorption works for at least two surfaces and for more than one protein.               19                  Figure 2.2 (a) BSA adsorption isotherms on a PZT surface incubated at 37 oC for 1 hr and 30 min. The average fluorescence intensity was analyzed by Adobe Photoshop 5.0, and is proportional to the adsorbed BSA on the surface. The relationship between adsorbed BSA and the concentration of BSA in solution fits the Langmuir isotherm well. (b) Fluorescence microscope images of PZT plate surfaces with 4 vibrating conditions: (1) No vibration, (2) 10 Vpp at 16 kHz, (3) 10 Vpp at 1 kHz, and (4) 3 Vpp at 16 kHz for 5 min. (c) BSA adsorption measurement based on Fig. 2(b). (d) Protein adsorption characterization of BSA onto PZT with titanium surfaces, and anti-mouse IgG onto silver surfaces.  (c) (1) (3) (4) (2) (b) (d) (a)  20 2.3.1 Adsorption Isotherm and Fluorescence Quenching In Fig. 2.2(a), the BSA adsorption behavior was fitted by a Langmuir isotherm. The Langmuir isotherm theory of particle adsorption onto a flat surface, based on a kinetic viewpoint, was first proposed by Langmuir (1918) [17,18]. The Langmuir Adsorption Model is derived from the equilibrium between empty surface sites (Es), particles (P), and filled particle sites (EsP). The particles are assumed to bind at a series of distinct sites on the surface, and the adsorption process was treated as a reaction wherever a particle P reacts with an empty site, Es, to yield an adsorbed complex (filled surface sites) EsP. s sP E E P+ U                                                                                                (2.1) The Equilibrium constant K of this reaction is given as: [ ] [ ],  [ ][ ][ ] [ ] s s s s E P E PK EP E K P= =                                                                             (2.2) [Es] = [Eso] - [EsP], [ ] [ ] [ ][ ] s s so E P E P EK P + =                                                      (2.3) where [Eso] is the concentration of total sites of the surface. If we set a dimensionless number θl as [EsP]/[Eso], then [ ]1,  [ ] 1 [ ] l l l K P K P K P θ θ θ+ = = +                                                                           (2.4) For a general expression, this equation can be described as Langmuir adsorption isotherm: max 1 l l KC KC ΓΓ = +                                                                                                 (2.5) where Γl = amount adsorbed, Γlmax = maximum amount adsorbed, C = aqueous concentration (or gaseous partial pressure), and K = Langmuir equilibrium constant, which signifies the affinity between the adsorabate and the surface. K increases with binding energy of adsorption and  21 decreases with temperature. K varies with different molecules and surfaces. The discussion of K on different adsorption systems mentioned in this thesis is summarized in Appendix A.4. Inherent within the Langmuir model, the following assumptions are valid for the simplest case: the adsorption of an adsorbate to a series of equivalent sites on a surface. 1. The surface containing the adsorbing sites is a perfectly flat plane with no corrugations. 2. The adsorbing particle adsorbs to an immobile state. 3. All adsorbing sites are equivalent and no interactions occur between the adsorbed particles on adjacent sites. 4. Each site can hold only one particle. Langmuir adsorption model may deviate significantly in many cases, primarily because it fails to account for the roughness of the substrate. Rough inhomogeneous surfaces have multiple sites available for adsorption; the heat of adsorption varies from site to site. The model also ignores the interaction between the adsorbed molecules, which is experimentally evidenced. In Fig. 2.2(b), the protein quantity is characterized by the fluorescence intensity. Nevertheless, the fluorescence intensity of dye has been reported to possibly be quenched near a noble metal surface [19-22]. Quenching refers to a process that decreases the fluorescence intensity of a given dye. A variety of processes can result in quenching, including energy transfer from dye to a metal surface. The quantum efficiency of emission from a dye represents how much fluorescence (energy) a dye can emit for a given energy absorption and is equal to the radiative decay rate (Rrad)/(radiative decay rate + nonradiative decay rate (Rnonrad)) [22]. The nonradiative rate has been reported to be distance (between dye and metal) dependent [22,23]. Quantum yield increases with the distance because of the reduction of the nonradiative rates. The reduction  22 often occurs within 10 nm, and reaches a maximum at the very proximity of a few nm. For example, the quantum efficiency of Cy5 dye decreases by a factor of 40 when the distance decreases from 16 to 2 nm [22]. The FITC-BSA used in this thesis is labeled with an average of seven FITC molecules on each BSA, and the BSA dimensions are 4 nm x 4 nm x 14 nm. We measure the BSA layer thickness (2 mg/ml, incubation for 1 hr) on a gold-coated silicon substrate by ellipsometry, and the thickness is 6.02 ± 0.41 nm, which is very close to our radius approximation ((14*4*4)1/3) for BSA. Hence, the distance between an FITC fluorescence dye and the surface is estimated to be around 6 nm. For a Cy5 dye, the fluorescence intensity quenches by about 50% within a 6 nm distance; though relevant data for FITC cannot be found. Then, we assume that the area occupied by a single BSA is (6.07)2 = 36.845 nm2. Hence, the adsorption quantity of a BSA monolayer is 68000 Da*(1/(36.845*10-18 m2))/6.02*1023 = 3.0366 mg/m2, which is close to what has been reported (4 mg/m2) [24]. To further investigate the fluorescent emission from the BSA monolayer, we incubated different substrates with BSA at very low concentrations (2*10-6, 6.6*10-5, 2*10-5, and 2*10-4 mg/ml) for 1 hr. The area of the surface immersed in the BSA solution was about 0.3 cm2. If BSA forms a monolayer on the surface, then we can calculate: ~ 3.0366*0.3*10-4 ~ 9.11*10-5 mg of BSA must be adsorbed from the solution. The volume of BSA solution is 0.5 ml, the concentration of BSA solution is ~ 1.82*10-4 mg/ml ~ 2*10-4 mg/ml. Hence, if we can observe fluorescence at concentrations lower than 2*10-4 mg/ml incubation concentration, even there is fluorescence quenching, we can still use fluorescence intensity to monitor BSA on the surface. Also, in real situations, all of the BSA in the solution will not adsorb onto the surface and more than 2*10-4 mg/ml concentration of BSA would be needed to form a monomer layer.  23 The fluorescence images of control Ti, Au surfaces, and those incubated with 2*10-4 mg/ml BSA solution for 1 hr are shown in Figure 2.3(a)~(d). Fluorescence was still observed for surfaces incubated with low concentration BSA solutions (Fig. 2.3(b) and (d)). The fluorescence measurement (shown in Fig 2.4) suggests that even though fluorescence quenching may occur, the fluorescence intensities are still strong enough to be observed and measured.      Figure 2.3 Fluorescence images of (a) control Ti (250 nm)-PZT plate, (b) low concentration (2*10-4 mg/ml) BSA incubated 1 hr on Ti (250 nm)-PZT plate, (c) control Au (50 nm)-PZT plate, and (d) low concentration (2*10-4 mg/ml) BSA incubated 1 hr on Au (50 nm)-PZT plate. Stronger fluorescence intensities are observed in (b) and (d) compare to (a) and (c), respectively.  (a) (b) (d) (c)  24 We then incubated the substrates (Ti (250 nm)-PZT plate and Au (50 nm)-PZT plate) with BSA at different concentrations (2*10-6, 6.6*10-5, 2*10-5, 2*10-4, 2*10-3, 2*10-2, 0.2, and 2 mg/ml) for 1 hr. The adsorption behavior is shown as fluorescence intensities (Figure 2.4) with different BSA concentrations. 0.0 0.5 1.0 1.5 2.0 0 20 40 60 80 100 120   BSA Cncentration (mg/ml) M od ifi ed  F lu or es ce nc e In te ns ity Equation y = (a*b*x^(1-c))/ (1 + b*x^(1-c)) Adj. R-Sq 0.9125 Value a 105 b 7.94234 c 0 0.0 0.5 1.0 1.5 2.0 0 20 40 60 80 100 120  M od ifi ed  F lu or es ce nc e In te ns ity BSA Cncentration (mg/ml) Equation y = (a*b*x^(1-c))/( 1 + b*x^(1-c)) Adj. R-Squa 0.61883 Value a 105 b 10.6868 c 0   Figure 2.4 The dependence of modified fluorescence intensities (modified with fluorescence intensity of surface incubated in 2 mg/ml BSA for 1 hr) of (a) Ti (250 nm)-PZT plate and (b) Au (50 nm)-PZT plate. The data is fitted with the Langmuir adsorption isotherm.  For comparisons of BSA adsorption on different metal-coated surfaces, the Langmuir isotherm was fitted with only two parameters: maximum adsorption (proportional to the fluorescence intensity) and the equilibrium constant K, which are shown as “a” and “b” in the inset tables in Fig. 2.4, respectively. The equilibrium constants K in Fig. 2.4(a) and (b) are 10.69 and 7.94, respectively. The equilibrium constants K are close, which means that the hydrophilicity of surfaces in Fig. 2.4(a) and (b) are similar. The contact angles of Ti- and Au- coated PZT plates were measured as 75.12° and 77.68°, respectively. The fitting of Fig. 2.4(b) with a Langmuir isotherm is inadequate and the fluorescence intensities from surfaces incubated (a) (b)  25 with 0.2 mg/ml BSA are lower than expectation from a Langmuir isotherm in both Fig. 2.4(a) and (b). The poor fitting of Fig. 2.4(b) with a Langmuir isotherm (R2 = 0.619) may be explained by: 1) an improper use of adsorption isotherm [25], and 2) the existence of a bimodal adsorption isotherm [26]. The Langmuir isotherm (Equ. (2.5)) can be rewritten as: max max 1 1 1 l l lKC = +Γ Γ Γ                                                                                    (2.6) From Equation (2.6), the plot of 1 lΓ with 1/C should be a straight line, with a constant slope of max1 lΓ . The 1 lΓ - 1/C plot of Fig. 2.4(b) is shown in Figure 2.5, but the slope is not a constant, and changes with 1/C. Hence, another isotherm, the Frumkin isotherm, should be considered for a better fitting [25]. The Frumkin isotherm, which considers the lateral interactions of adsorbates and represents a more accurate depiction of adsorption compared to the Langmuir isotherm, is written as [25]: max max , l l a l F l l NK C e a kT φΓ− ΓΓ= = −Γ −Γ                                                             (2.7) Where KF is the Frumkin equilibrium constant, k is the Boltzmann constant, T is the temperature in Kelvin, Nφ  is the interaction energy of one molecule with its N nearest neighbors on the completely covered surface. A negative sign in the parameter a means the repulsion between absorbates, while a positive sign means the attraction between absorbates. When the parameter a is equal to zero (no lateral interaction), this isotherm is reduced to the Langmuir isotherm. In this isotherm, the equilibrium constant K (K = KF*eaθl ) is changing with surface coverage.    26     Figure 2.5 The plot of 1/Fluorescence intensities of Au (50 nm)-PZT plates with 1/C (data from Fig. 2.4(b)).      The poor fitting of Fig. 2.4(b) with the Langmuir isotherm can also be explained by the possible bimodal adsorption of BSA on the surface. From the data of Roscoe et al., a bimodal isotherm (two plateaus) of BSA was investigated on the titanium surfaces at all temperatures [26]. The authors argued that following the BSA adsorption, some conformational changes in the BSA took place during the adsorption process. When the energies needed to break intramolecular interactions are greater than the energies released in forming protein-metal interactions, the adsorption process is endothermic, which suggests the presence of conformational changes. Indeed, the authors found entropic changes during the adsorption of BSA onto the titanium surfaces [26]. The bimodal isotherm can be explained by first, at low concentrations (0.12 mg/ml), the adsorption process proceeds by a random uncorrelated mode (involving conformational changes), that results in a completely disorganized structure (with a less dense random protein layer). Secondly, at high concentrations (0.23 mg/ml), the protein layer favors a 0 1x105 2x105 3x105 4x105 5x105 0.00 0.02 0.04 0.06 0.08 0.10   1/ Fl uo re sc en ce  In te ns ity 1/C  27 more dense structure. Between the two concentrations, the protein layer structure is transitional. The degree of order of the protein layer depends on the relaxation time of the protein configuration, and the collision frequency of protein molecules with the surface. When the collision frequency is sufficiently large to overcome the kinetic limitation of the configuration change, a close-packed protein layer is formed. Thus, the appearance of two plateaus in a bimodal isotherm indicates the occurrence of a protein monolayer with different conformations. These two conformations of BSA on surfaces are shown in Figure 2.6, where the heights of the BSA were determined to be 4.94 and 6.66 nm (1.87 and 2.52 mg/m2) for the two conformations at the first and the second plateaus, respectively [26].     Figure 2.6 Illustration of BSA conformations on surfaces in a bimodal isotherm.  In our research, the quantities of adsorbed BSA were determined by fluorescence intensities of BSA-FITC. Assuming that the FITC dyes have a normal distribution in BSA, the fluorescence intensities of the BSA-dye for the two conformations shown in Fig. 2.6 are different because the fluorescence quenching ratio depends on the distance between the dye and the surface. The energy transfer follows a 1/d4 dependence [27], and the quantum efficiency of energy transfer in the surface energy transfer can be written as [27]: 40 1 1 ( ) dF d F d φ ∞ = = +                                                                                         (2.8)           1st plateau                                           2nd plateau 4.94 nm 6.66 nm  28 Where d0 is the distance at which the quantum efficiency of the dye is 50%; F and F∞ are the fluorescence intensities of the dye at distance d, and infinity, respectively. From Equation (2.8), the quantum efficiency of the dye at the first conformation is lower than that at the second conformation of BSA. Hence, the actual fluorescence intensities measured must be adjusted according to the fluorescence quenching. Equation (2.7) can be re-arranged as:  max max max 1 1 1( ) l l a l l F l e K C Γ Γ− =Γ Γ Γ max max max max max max 1 1 1ln( ) ln ln 1 1 1ln( ) ln ln l F l l l l d F d a KC Fa K FF F F C ∞∞ ∞ Γ− + = − Γ ⇒Γ Γ Γ − + = −                                              (2.9) After compensating for the fluorescence quenching, Equation 2.9 can be written as: 40 max 40 max max (1 ( ) )1 1 1ln( ) ln ln (1 ( ) ) d F d dF da K Fd F F CF d ∞ ∞ ∞ +− + = − +             (2.10) For finding the parameter of 0d , maxF∞ , and a, we plot the data from Fig. 2.5 in the plot of fluorescence intensity vs. ln(1/C), fitted with Equation (2.10) (Figure 2.7). Since the elliptic BSA conformation appears at a lower surface concentration, and d is around 4.94 nm, we only use data for the lower concentration (<= 0.02 mg/ml) for the fitting with the plot, to determine d0.       29   Figure 2.7 Dependence of fluorescence intensities of Au (50 nm)-PZT plates with ln(1/C) (data from Fig. 2.4(b)).     The R2 of the fitted curve is 0.933, and the parameter of 0d , maxF∞ , and a are 6.24 nm, 191, and -17.19, respectively. The fluorescence intensities of the surface, incubated with the 2 mg/ml BSA solution, are adjusted by the fitted 0d , and thus modified to be from 100 to 177.2. The fluorescence intensities of the surface, incubated with 0.2 mg/ml BSA are adjusted by the average values of low (<2*10-2 mg/ml) and high (2 mg/ml) concentrations, and thus modified to be from 58.2 to 137.5. Then we plotted max max 1 1ln( ) d d FaF F F∞ ∞ − +  with 1ln C  (Figure 2.8).   Figure 2.8 Dependence of max max 1 1ln( ) FaF F F− + on 1ln C  (data from Fig. 2.4(b)).    4 6 8 10 12 14 10 15 20 25 30 35    Experimental data  Fitting data Fl uo re sc en ce  In te ns ity ln(1/C) -2 0 2 4 6 8 10 12 14 -25 -20 -15 -10 -5 0    Frumkin-without FI compensation  Frumkin-with FI compensation ln (1 /F d- 1/ F ∞ m ax )+ aF d/F ∞m ax ln(1/C)  30 From Fig. 2.8, the R2 values of the Frumkin isotherm fittings of the fluorescence intensities, without and with adjustments, are 0.788 and 0.971, indicating that the fitting of the Frumkin isotherm is improved by adjusting for the fluorescence intensities. In addition, the intercept is - 21.63, which is axln F mK F∞− , hence, the KF is 1.28*107 ml/mg. The equilibrium constant in Frumkin isotherm is changing with the surface coverage as K = KF*eaθl. Since the parameter a is negative, the equilibrium constant decreases exponentially with the increase of the surface coverage, which means more difficult for a BSA to adsorb unto the high BSA covered surface. Furthermore, the Langmuir equilibrium constant, K, can be related to KF*eaθl. The average surface coverage, 0.5, is used for a rough calculation of K from KF. The K is hence 1.28*107*e- 17.19*0.5, which is 2379.1. The Langmuir equilibrium constant of BSA on gold nanoparticles is reported as 2.58*103 [28] ~ 1.38*104 [27], and these values are close to our result (2379.1). Compared to the Langmuir equilibrium constant (5.1 and 7.3) from Fig. A.9(a) and (b), this value (2379.1) is much larger. Since the fitting in Fig. A.9(a) and (b) only considers high surface coverage (0.64~0.93), the K is evaluated as 1.28*107*e-17.19*(0.64+0.93)/2, which is 17.9 and close to 5.1 or 7.3. Based on the data from Fig. 2.3~2.8, and the thickness (~6.02 ± 0.4 nm) of the BSA, measured by ellipsometry (detailed measurement shown in Chapter 5), the BSA monolayer (rather than multilayer) formed on the surface after incubation in 2 mg/ml BSA. Hence, we can make several valid assumptions: 1) the conformation of BSA is close to that of a sphere at a high surface concentration; 2) the mechanism described in this chapter deals with a BSA monolayer; 3) fluorescence quenching happens, but since the BSA surface concentration used in this chapter is on the same plateau as the adsorption isotherm, the quenching ratio is the same, and our assumption is valid that the quantities of BSA is linear with the fluorescence intensity; and 4) the  31 deviation from the Langmuir isotherm exists on surfaces with lower surface coverage, but it fits data fairly and conveniently for high coverage surfaces, and the Langmuir isotherm is used to fit the adsorption data in this thesis.  2.3.2 Electrochemical Impedance and Electrical Equivalent Circuits Fitting In addition to fluorescence imaging, we used electrochemical impedance as a complementary tool to measure the protein desorption. Figure 2.9 shows the Nyquist plot measured by an impedance analyzer for three PZT plates as working electrodes immersed in PBS under three different conditions: (a) a bare PZT plate, (b) a PZT plate incubated with BSA without vibration, and (c) a PZT plate incubated with BSA and vibrated. Z’ stands for the real part of impedance, while Z” signifies the imagery part of impedance.    Figure 2.9 Nyquist plot of three plates as working electrodes: bare PZT, BSA adsorbed, and BSA adsorbed after vibration surfaces in PBS. The curves can be modeled with electrical equivalent circuits.     32 The measured Nyquist plot can be modeled by electrical equivalent circuits (EEC) [26,29-32]. Three EECs were used to model the electrochemical impedance data (shown in Figure 2.10). These three circuits are EEC 1 (two parallel RCPE [29]), EEC 2 (one parallel RC and one parallel RCPE in serials [30]), and EEC 3 (one parallel RCPE [26,31,32]).     Figure 2.10 Three fitting equivalent electrical circuits, (a) one solution resistance and two parallel RCPE, (b) one solution resistance, one parallel RC and one parallel RCPE in serials, and (c) one parallel RCPE.       R, C, and CPE represent resistors, capacitors, and constant phase elements, respectively. All three circuits have the resistor, RΩ, representing the resistance of the solution between working and reference electrodes. Figure 2.10(c) is the modification of a Randles Cell, where the parallel RC represents the surface double layer capacitor and a charge transfer or polarization resistance (from solution to electrode). (a) (b) (c)  33 The constant phase element (CPE, c(iω)-n) is considered as a non-ideal capacitor, where c is the capacitance, n is an exponent, which equals 1 for an ideal capacitor. But when n equals 0.5, the solution-solid interface is dominated by a pure diffusion process (Warburg impedance). Generally speaking, n has a value between 0.5 to 1, which means that most liquid-solid interfaces have a mixed charge transfer and mass transport processes. The CPE is necessary for the heterogeneous electrode (imperfect capacitor) [29,30]. For EEC1 (Figure 2.10(a)), the two RCPE elements represent the interfaces of silver-solution and silver-BSA film, respectively [29]. For EEC2 (Figure 2.10(b)), RC and RCPE elements represent the silver-BSA film and the silver-solution interface, respectively [30]. For EEC3 (Fig. 2.10(c)), the RCPE elements represent the interface of solution-BSA film together with silver surface [26,31]. The fitting results from EEC1, EEC2 and EEC3 are listed in Table 2.1. The bare surface specimen is fitted by EEC3, since the surface has no BSA adsorption. As seen in Table 2.1, the charge transfer resistance (R1) in all three fittings has the most significant changes compared to those in bare surfaces. With more BSA adsorbed on the surface, R1 is reduced significantly, which suggests that the surface charge density, resulting from protein adsorption, is directly proportional to the amount of adsorbed protein on the surface. This implies that the charge transfer is accompanied by the BSA adsorption [30]. In fact, BSA is negatively charged at pH 7.2 in PBS, and the resistance comes from the adsorbed BSA itself (R2), which is small compared to R1. It follows that more BSA adsorption increases R2, which indicates that the resistance of the BSA layer (R2) is increasing with the adsorbed BSA; while the adsorption of BSA plays a much more important role in the transfer of charges from electrolyte to electrode (decreasing R1). The value of n is between 0.5~1 indicating that the rate of the dissolution of  34 silver due to BSA adsorption was controlled by both surface diffusion and by charge transfer processes.               Table 2.1 Summary of fitting parameters of electric equivalent circuits  EEC RΩ (Ωcm2) R1 (Ωcm2) R2 (Ωcm2) c1 (mF cm−2 sn−1) n1 c2 (mF cm−2 sn−1) n2 c (mF cm−2) Bare 3 54.8 9103  0.294 0.68 1 24.46 918.29 6.86 0.365 0.79 0.0981 0.8 2 25.47 891.56 5.67 0.354 0.80   0.0246 BSA 3 27.09 979.3  0.407 0.75 1 23.48 4950 2.99 0.327 0.65 0.0761 0.67 2 27.39 4978 3.14 0.349 0.64   0.024 BSA- Vibration 3 27.95 5171  0.343 0.64  Two Bode plots are used to interpret the impedance data, except for the Nyquist plot (shown in Fig. 2.9). Two Bode plots of BSA-adsorbed surface ((a) Z’ vs. log(frequency) and (b) phase angle vs. log(frequency) are shown in Figure 2.11.  Figure 2.11 Bode plots of BSA adsorbed surfaces: (a) Z’ v.s. log(frequency) and (b) phase angle v.s. log(frequency).  (a) (b)  35 The impedance spectra show three different frequency regions. In the high-frequency region (>100 Hz), the impedance (Z’) is independent of frequency (Fig. 2.11(a)) and the phase angle is near 0° (Fig. 2.11(b)). This behavior probably comes from the resistance, RΩ, of the phosphate buffer solution between the working and the reference electrode. A linear relationship is observed between the Z’ and log(frequency) in the medium-frequency region (3 Hz to 10 Hz). This may be due to the capacitive behavior of the electrode/electrolyte interface. In the low frequency region (3 Hz to 150 mHz), the phase angle reaches a maximum and then starts to decrease. This is attributed to the Faradaic processes, which is associated with the charge transfer at the electrode/PBS interface.  For the Ag electrode in PBS, the following half reduction reactions are possible [33]. Ag ÆAg+ + e-                                       Eo = -0.8 V                                     (2.11) O2 + 2H2O + 4e- Æ 4OH-                     Eo = 0.4 V                                      (2.12) 2H2O + 2e- Æ H2 + 2OH-                     Eo = -0.8277 V                               (2.13)  Equation (2.11) is the dissolution of Ag; Equation (2.12) is the oxidative reaction of water with the presence of O2; and Equation (2.13) is the oxidative reaction of water in the absence of O2. Since we did not degas the PBS before the measurements, this reaction is less likely to happen, but for comparison, we discuss this case here. The potential shown in Equ. (2.11)~(2.13) is the standard potential (Eo) of the half reduction reaction (298 K, 1 atm). For reactions under non-standard conditions, the half reduction potentials are modified by the Nernst equation.    36 Hence, the possible redox reactions could be: (2.11) + (2.13) with the presence of O2 4Ag + O2 + 2H2O Æ Ag+ + 4OH-        E = -0.4V                                       (2.14) (2.12) + (2.13) in the absence of O2 2Ag + 2H2O Æ Ag+ + H2 + 4OH-        E = -1.6277V                                 (2.15)  In the presence of O2 (PBS without degassing), the redox reaction is favored to happen (smaller negative potential). The open circuit potentials (OCP, also measured in the impedance measurement) for the three conditions are -0.191, -0.2467, and -0.3863 for bare silver, after BSA adsorption and vibration, and after BSA adsorption, respectively. The data indicate that the adsorption of BSA produces a change in the mechanism and kinetics of the reactions such that, the dissolution of the silver electrode (corrosion) is enhanced. It has been shown that the release of silver ion can be enhanced by the adsorption of proteins in vitro [34,35]. The rate of cell death was also correlated with the amount of silver released [341]. In addition, the role of proteins in corrosion of a variety of clinically available dental casting alloys was also studied [35]. It was found that more elemental release occurs into the saline-BSA solution compared to saline alone for Ag, Cu, Pd, and Zn, observations which are consistent with the impedance and OCP measurements in this study.  2.3.3 Lateral Forces For flexure plate wave (FPW)-induced acoustic forces, two important types of forces need to be considered: 1) Stokes drag from acoustic streaming, and 2) the acoustic radiation force (which pushes particles to accumulate in pressure nodes) from standing waves. The acoustic radiation force induces a movement on particles relative to the medium, while the acoustic streaming  37 moves the entire fluid. The acoustic streaming force scales with the radius (R) of the particle while the acoustic radiation force scales with the volume (R3). Hence, for particles of smaller size (<1 μm), acoustic streaming dominates over the acoustic radiation force [12,36,37]. In the present case, the particles described here are proteins with diameters in the nanometer range; we neglect acoustic radiation force and only consider acoustic streaming force. The flow pattern and strength is governed by the acoustic pressure field, which is greatly influenced by the flow velocity. For converting pressure to velocity, the acoustic impedance (Z), relates pressure to velocity exerting on a particle in solution (PBS in our case) [38]: Z = pressure/velocity                                                                                   (2.16) Since the acoustic impedance in PBS is constant, the increase in pressure linearly increases the velocity. Thus for a vibration driven by a harmonically oscillating piezoelectric plate, the time-dependence of pressure can be described as p*cos(ωt). The streaming vortex goes from the pressure node to the point with highest pressure [39,40].  The relationship between vibration amplitude and flow velocity follows [41]: 2 3 25 ( ) ( ) ( )(1 ) 4 2 S S P S fAmp du κ κα αν α κ β= + −                                                     (2.17) Where u is the maximum velocity of the acoustic stream; f is the frequency of vibration; Amp is the z-directed component of membrane vibration amplitude; Pυ  is the phase velocity of the plate wave; κs is the propagation constant, equal to 2π/λ; λ is the wavelength of FPW; α = κs(1-( Pυ / Fυ )2)1/2; Fυ  is the sound velocity in PBS; β2 = ρf/2η, which relates to the viscosity effect of PBS; ρ is the density of PBS; η is the dynamic viscosity of PBS; and d is the membrane thickness. From Equation (2.17), the flow velocity is proportional to the square of the vibration  38 amplitude, which is linearly related to the driving voltage. This relationship between vibration amplitude and flow velocity has also been experimentally observed (Figure 2.12) [16,41,42].   Figure 2.12 Fluorescence image of PZT plate after vibration 5 min under 10 Vpp at resonance. The inset is the simulation of a PZT plate vibrating at resonance frequency. The highest vibration amplitude is located at the center of PZT plate.   The BSA desorption is position-dependent on a PZT plate and matched to the vibration amplitude of the PZT plate (Fig. 2.12). Modified flow velocity (the center point of PZT is taken as reference) vs. BSA desorption percentage is shown in Figure 2.13.      Figure 2.13 The dependence of BSA desorption and modified flow velocity.    30 40 50 60 70 0.0 0.1 0.6 0.7 0.8 0.9 1.0 1.1   M od ifi ed  F lo w  V el oc ity BSA Desorption (%) Equation y = a + b*x Adj. R-Square 0.99029 Value Standard Error Intercept 0 -- Slope 0.01725 6.96449E-4  39 In addition, two important scenarios are needed to investigate the role of acoustic streaming in BSA desorption mechanism: 1) the surface charge without flow application, and 2) the flow without surface charge application. For the first scenario, an experiment with a driving AC voltage on a silicon membrane (no vibration induced flow) and in the second scenario, a rotation plate experiment is performed to prove the mechanism. For scenario 1, a silicon substrate (1 cm x 0.5 cm x 0.05 cm) deposited with 250 nm thick titanium is prepared by e-beam evaporator. A sinusoid AC driving voltage (16 kHz, 10 Vpp) is applied on the silicon. Since silicon is not a piezoelectric material, it will not vibrate in PBS solution and no flow will be generated. The titanium-coated silicon substrate is incubated with fluorescently labeled BSA at 2 mg/ml concentration for 1 hr. The fluorescence intensities of the surface before and after AC application are shown in Figure 2.14. From Fig. 2.14, no significant difference is seen in fluorescence intensities between the scenarios, to suggest that the AC application without acoustic streaming cannot remove the adsorbed BSA from the surface.  Before AC After AC 0 20 40 60 80 100   M od ifi ed  F lu or es ce nc e In te ns ity  Figure 2.14 Modified fluorescence intensities of surfaces before and after AC application. The error bar is the standard deviation of 6 measured data.   40 For scenario 2, a rotating plate system was used to separate the effects of electric field and acoustic streaming on protein adsorption. In a rotating plate, no charge accumulation takes place on the metal surface. The rotating plate system setup consists of a rotor connected to a sample Cu plate coated with Ti (Figure 2.15). Different sinusoid AC driving voltages (0.7, 1, and 3 V) are applied on the motor for different rotating speeds. The titanium-coated copper plate is incubated with BSA at 2 mg/ml concentration for 1 hr. The fluorescence image, after rotating the plate at the highest voltage for 5 min, is shown in Figure 2.16. Unlike the fluorescence image on a PZT plate (Fig. 2.12), the distribution of fluorescence on a copper plate is more uniform and the fluorescence intensity is comparable to that seen before rotation.    Figure 2.15 Schematic diagram of rotating plate system.       Figure 2.16 Fluorescence image of a copper plate after rotating for 5 min. The dark region in the center is connecting to a motor for rotation.     41 The comparison between shear stresses induced by a rotating plate and a PZT plate is discussed in the following section. The fluid wall shear stress ( ωτ , dyn/cm2) on the surface of a rotating plate is defined as [43]: 3 1/ 20.8 ( / )rωτ η ω υ=                                                                                     (2.18) where η (cPoise) is the dynamic viscosity of the medium; r (cm) is the radial distance from the center of the disk; ω (rpm) is the radial velocity of the rotating disk (which is measured by LDV); and υ (Stokes) is the kinematic viscosity of the medium. The medium in this experiment is water at room temperature, so η and υ  are 0.0089 (Poise, g/cm.s) and 0.008583 (Stokes, cm2/s), respectively. The average shear stress over a rotating plate is expressed as: 2 3 1/ 2 2 2 3 1/ 2 , 11 0.8 ( / ) 2 | 0.267 0.8 ( / ) r r w ave rr r rdr rτ η ω υ π π η ω υ= = ×∫ i (dyne/cm2)  (2.19) where η and υ  are 0.0089 (Poise, g/cm.s) and 0.008583 (Stokes, cm2/s), respectively. The radius, r, is from 0.1 to 0.38 cm. The average wall shear stresses induced by different rotating speeds are obtained by incorporating these numbers and the rotating speed into Equation (2.19), which are listed in Table 2.2. For example, the average shear stress induced by a 40.79 Hz rotation is 0.267*0.8*8.9*10-3*((40.79)3/8.583*10-3)0.5 = 5.35 dyne/cm2.               Table 2.2 Shear stress applied on BSA molecules and related parameters Applied Voltage (Vpp) Rotating Speed (Hz) Shear Stress (dyne/cm2) 0.7  1.49 3.73 x 10-2 1 6.66 3.53 x 10-1 Rotating Plate 3 40.79 5.35 x 100 Applied Voltage (Vpp) Max Vibration Amplitude (nm) Shear Rate (1/s) Shear Stress (dyne/cm2) Vibrating PZT 10 500 2.71 x 10-3 2.41 x 10-5   42 For acoustic streaming, the shear stress is defined as u zτ μ= ∂ ∂ ; where η is the dynamic viscosity of PBS (0.0089 Poise); u is the flow velocity; z is the distance from the surface; and u z∂ ∂ is shear rate. The velocity of acoustic streaming is calculated from Equ. (2.17), which is 8.84*10-6 m/s (10 Vpp) and is the maximum velocity parallel to the plate surface. The flow on the wall is assumed to be no-slip, hence the velocity at the wall is zero. The velocity of flow reaches a maximum at the evanescent decay length, which is the reciprocal of α in Equ. (2.17). For our experimental setup, 1/α equals 3.26*10-3 m. Assuming that the dependence of velocity with distance from wall is linear, the shear rate of acoustic streaming is 8.84*10-6/3.26*10-3 = 2.71*10-3 1/s. The shear stress, u zτ μ= ∂ ∂ , is 8.9*10-4 (kg/m.s) * 2.71*10-3 (1/s) = 2.41*10-6 N/m2 = 2.41*10-5 dyne/cm2. Parameters for calculating shear stress from acoustic streaming are also listed in Table 2.2. The fluorescence intensities for plates rotating at different speeds for 5 min were compared with plates without rotation (results shown in Figure 2.17) and no significant difference occurs between different speeds. Since shear stresses induced by a rotating plate are much larger than those of acoustic streaming, we conclude that the acoustic streaming shear stress itself cannot desorb BSA from a PZT plate surface.        3.73*10 -2    0.353     5.35 (dyne/cm2)  Figure 2.17 Fluorescence intensity of Ti coated plate surfaces at different applied shear stresses on one BSA molecule.   43 In conclusion, the BSA desorption is related to both the vibration amplitude and surface charge. Two experiments were performed: 1) driving AC voltage on a silicon substrate, and 2) a rotation plate experiment. Neither experiment showed attenuation of BSA adsorption.  2.3.4 Vertical Forces We explored the influence of electric-field on protein desorption by applying a DC voltage across the PZT plate in a BSA solution (pH 7.4). The results are given in Table 2.3. The BSA adsorption on a silver surface on PZT charged with -1 V is much less than that of a surface charged with + 1V. Since BSA is negatively charged at pH 7.4 (the isoelectric point of BSA is 4.9), we believe that the electrostatic repulsion on the metal-coated PZT surface with a negative potential contributed to less protein adsorption [44]. When an AC voltage is used, since the voltage signal is sinusoidal, negative and positive charging occurs on the PZT plate alternatively. The half period of the AC signal operated in resonance frequency in our experiment was (1/16kHz)/2 = 3.125*10-2 ms. The time constant required for the total system (function generator, wires, PZT plates) to charge the plate surfaces is equal to the system resistance (mainly from the connection of wires and the PZT) multiplied by the capacitance of PZT. The resistance between the wires and the metal electrode in PBS is 2 to 5 Ω and the capacitance of a PZT plate vibrating at resonance frequency is 1.5*10-5 F. Therefore, the time constant is calculated from 3 to 7.5*10- 2 ms. The negative average voltage of a 10 Vpp AC signal, measured after immersing the PZT plate into PBS, is -2*5/π = -3.18 V [45]. The charging voltage during this period is -3.18 * (1- exp(-3.125/7.5)) to -3.18 * (1-exp(-3.125/3)) = -1.08 to -2.06 V [45]. Thus, the voltage is large enough to desorb the proteins from surfaces, as supported by our experiments with DC voltage.  44 We believe that the surface charge is the major reason for the desorption of BSA from the metal- coated PZT surface.  Table 2.3 Protein adsorption on positively and negatively charged surfaces      To further substantiate our experimental results, we calculated the surface interaction of BSA theoretically, under the conditions studied. At the molecular level, three protein-surface interactions dominate; i.e., electrostatic, hydrophobic, and van der Waal (VDW) interactions [13]. These interactions were modeled using BSA and calculated as functions of the distance from a metal-coated PZT surface. All of the variables used in the following equations are given in Table 2.3. The electrostatic double-layer interaction force Fe between a sphere (radius R) and a flat surface can be modeled as [13]:                                                                                                                                                    (2.20) where Z is a constant derived from the Gouy-Chapman electrostatic double-layer model. The Debye length 1/κ can be modeled as [13]:                                                                                                                                                    (2.21) For BSA in a PBS solution with 0.15 M NaCl at 37 oC, 1/κ is equal to 0.8 nm. The constant Z at 37 oC can be defined as [13]:                                                                                                                                                 (2.22) Voltage(V) Fluorescence Intensity Normalized -1  58.92 ± 4.7 n=3 35.2 ± 2.8% +1 165.38 ± 10.34 n=3 100 ± 6.3% ( )DeF RZe N κκ −= [ ]0.5 1 0.3 (nm) NaClκ = 2 2 11 264 ( ) tanh ( / 4 ) 9.38*10 tanh ( /107) ( / )o o oZ kT e ze kT J mπεε ψ ψ−= =  45 The relationship between surface potential ψo and surface charge density σ given by the Graham equation for 1:1 electrolytes such as NaCl at 37 oC is [13,46]:                                                                                                                                                 (2.23) Experimentally, σ can be estimated by measuring the voltage V across the PZT plate in the BSA solution, and can be modeled as [47]:                                                                                                                                                 (2.24) where Co is the PZT capacitance in air at resonance and A is the area of the PZT plate. Z can be calculated from Equations (2.22), (2.23), and (2.24). Furthermore, the physical dimension of a BSA molecule is 14 nm x 4 nm x 4 nm [48], and the geometric mean radius R can be evaluated as 3.04 nm (~1/2*(14*4*4)1/3). The electrostatic interaction energy Ee of a BSA molecule is [13]:                                                                                                                                                (2.25) The VDW interaction force FV between a sphere (radius R) and a flat surface can be derived as [13]:                                                                                                                                                 (2.26) where H is the Hamaker constant [49]. The VDW interaction energy EV of a BSA molecule can be modeled as [13]:                                                                                                                                                (2.27) The hydrophobic interaction energy Eh between two spherical groups with radius R is [47]:                                                                                                                                                    (2.28)  where k is Boltzman constant and is ~1.38*10-23; T is temperature (K); R is the radius of protein (Å); and D is the distance from surface (Å). Also the hydrophobicity of surface is needed to be considered since it will affect hydrophobic interaction energy. The hydrophobic interaction 2/ 6  ( )VF HR D N= − [ ]0.5 20.116 sinh( / 53.4) ( / )o NaCl C mσ ψ= 19 1.25/  4.89*10 ( )De eE F e Jκ − −= = 20* 1.01*10  ( )VE F D D J −= − = − /141.903 DhE Re kT −= − 2/ * / ( / )oQ A C V A C mσ = =  46 energy is related to surface tension of liquid-surface and can be evaluated by measuring water contact angle on PZT. The following equation depicts that surface tension is a function of contact angle [13].                                                                                                                                                    (2.29) where γS = 25-30 mJ, γL = 72 mJ. The contact angle, θ, of hydrophobic polymer mentioned in reference [50] is 95º, and is 77.8º for gold coated PZT surface. γi-Au is about 0.36*γi-polymer. So the hydrophobic interaction energy of gold coated PZT surface is 0.36 of hydrophobic interaction energy calculated by Equation (2.28). Equations (2.25), (2.27), (2.28), as well as the summation of electrostatic, VDW, and hydrophobic energy, are plotted in Figure 2.18. The VDW and hydrophobic interactions are attractive, while the electrostatic interaction is repelling. The attractive interactions have the same magnitude order as that of the repelling interaction (Table 2.4). Therefore, we hypothesize that the electrostatic repelling interaction can overcome the adhesive interactions. From the theoretical calculations and experimental observations on protein surface interaction forces and shear stress, we conclude that two steps are involved in the surface protein desorption process in the present case. First, the electrostatic repulsion overcomes the VDW and hydrophobic adhesion interactions. Then, acoustic streaming near the surface carries the desorbed protein away and prevents the proteins from being adsorbed back again. In addition, the removal of desorbed proteins is shear stress-modulated.  cosi SL S Lγ γ γ γ θ= = −  47 0 1 2 3 4 5 -1.0x10-19 -5.0x10-20 0.0 5.0x10-20 1.0x10-19 1.5x10-19 2.0x10-19 E ne rg y (J ) Distance (nm)  VDW  Hydrophobic  Electrostatic  Total  Figure 2.18 Calculated protein-surface interaction energy between a BSA molecule and a negatively charged PZT surface (pH = 7.4). In the proximity of the surface, the total attractive energy of VDW and hydrophobic interaction can be compensated by the repelling electrostatic interaction.           48 Table 2.4 Variables used in Equ. (2.20)-(2.28) a,b Electrostatic interaction Hydrophobic interaction VDW interaction Total Equ. (2.20) Equ. (2.22) Equ. (2.23) 1/k(nm) 0.8 ψo(mV) -175 R(nm) 3.04 R(nm) 3.04 σ (C/m2) 0.5962 Z(J/m) 8.06*10-11 D(nm) 0.3 D(nm) 0.3 C(farad) 1.6*10-5 R(nm) 3.04 k 1.38*10-23 H(J)  10-20 V(volt) -1 D(nm) 0.3 T(K) 310  Ee = 1.67*10-19 Eh = -6.82*10-20 EV = -1.69*10-20 ET = 8.18*10-20 a D is the distance from surface when the BSA is adsorbed on the surface. We used a cut-off distance of 0.3~0.4 [46]. b The calculated magnitude of attracted protein-protein-surface forces is compared with the report in the literature [51]. Blomberg et al. measured a force-distance profile of human serum albumin (HSA) and a mica surface by a protein-surface force apparatus. The average energy of a HSA molecule and a mica surface is around 3.6*10-19 J, which is close to the 6.82*10-20 J calculated in Table 2.4.           49 2.4 Conclusions We have demonstrated a proof-of-concept nanomolecule desorption mechanism on a surface. A combination of electric field and mechanical vibration were used to remove proteins (BSA and IgG) from metal-coated PZT surfaces. Our theoretical calculations on protein-surface interactions also support this mechanism. Protein desorption was evident from both fluorescence microscope studies and electrochemical impedance spectroscopy experiments. This active protein desorption or removal mechanism can potentially be used in implantable biosensors without, or in combination with, passive biocompatible polymer surface treatment. Even though only partial desorption of protein is observed here by vibration, a potential future study could combine other anti-fouling methods, with the methods used here, to produce better results.                       50 2.5 Bibliography [1] D. M. Fraser, in D. M. Fraser (Eds.), Biosensors in the Body: Continuous in vivo Monitoring, 1st ed., John Wiley & Sons, Chichester, 1997, Chapter 1. [2] B. D. Ratner, S. J. Bryant, Annu. Rev. Biomed. Eng., 6 (2004) 41. [3] J. M. Anderson, Cardiovasc. Pathol., 2 (1993) S33. [4] K. M. Ainslie, G. Sharma, M. A. Dyer, C. A. Grimes and M. V. Pishko, Nano Lett., 5 (2005) 1852. [5] G. P. Rigby, P. W. Crump, P. Vadgama, Med. Biol. Eng. Comput., 33 (1995) 231. [6] G. P. Rigby, P. W. Crump, P. Vadgama, Analyst, 121 (1996) 871. [7] K. L. Prime and G. M. Whitesides, J. Am. Chem. Soc., 115 (1993) 10714. [8] R. Gref, M. Lück, P. Quellec, M. Marchand, E. Dellacherie, S. Harnisch, T. Blunk and R. H. Müller, Colloids and Surfaces B: Biointerfaces, 18 (2000) 301. [9] M. Rahmoune and M. Latour, Smart Matter. Struct, 4 (1995) 195. [10] G. D. Meyer, J. M. Morán-Mirabal, D. W. Branch and H.G. Craighead, IEEE Sensors Journal, 6 (2006) 254. [11] Q. Qi, G. J. Brereton, IEEE T. Ultrason. Ferr., 42 (1995) 619. [12] P. F. Caton, Microfiltration and Flexural Plate Wave Devices. Ph.D. Dissertation, University of California at Berkeley, Berkeley, CA, 2001. [13] D. Leckband and J. Q. Israelachvili, Rev. Biophys., 34 (2001) 105. [14] T. T. Huang, J. Sturgis, R. Gomez, T. Geng, R. Bashir, A. K. Bhunia, J. P. Robinson and M. R. Ladisch, Biotechnol. Bioeng., 81 (2003) 618. [15] A monolayer of BSA adsorbed on a surface is about 4 mg/m2 [24]. The surface area of the PZT plate is 0.5 cm2; when the protein attaches to the PZT as a monolayer, the amount of protein adsorbed is 4*10-4 mg. When the PZT plate is immersed in a 1 ml protein solution with 0.1 mg/ml (the minimum concentration in Fig. 2.2(a)) and 2 mg/ml, the incubation concentration, the adsorption amount is only 0.4% and 0.02% of the total proteins in solution, respectively. These amounts are below the detection limit of conventional quantitative protein assay methods, such as UV/vis spectroscopy (Perkin Elmer UV/vis spectrometer user’s manual). [16] N. T. Nguyen, A. H. Meng, J. Black and R. M. White, Sensor. Actuat. A-Phys., 79 (2000) 115.  51 [17] R. I. Masel, Principles of Adsorption and Reaction on Solid Surfaces, John Wiley & Sons, Inc., New York, 1996, pp.235-246. [18] D. D. Do, Adsorption Analysis: Equilibria and Kinetics, Imperial College Press, London, 1998, pp.11-48. [19] J. L. Shepherd, A. Kell, E. Chung, C. W. Sinclar, M. S. Workentin, and D. Bizzotto, J. Am. Chem. Soc., 126 (2004) 8329. [20] D. Bizzotto, B. Pettinger, Langmuir, 15 (1999) 8309. [21] E. Dulkeith, A. C. Morteani, T. Niedereichholz, T. A. Klar, J. Feldmann, S. A. Levi, F. C. J. M. van Veggel, D. N. Reinhoudt, M. Mӧller, and D. I. Gittins, Phys. Rev. Lett.,  89 (2002) 203002. [22] E. Dulkeith, M. Ringler, T. A. Kla, J. Feldmann, A. M. Javier, and W. J. Parak, Nano Lett., 5 (2005) 585. [23] J. I. Gersten, in J. R. Lakowicz and C. D. Geddes (Eds.), Topics in Fluorescence Spectroscopy, Volume 8, Radiative Decay Engineering, Springer, New York, 2005, Chapter 6. [24] W. Norde, and C. E. Giacomelli, Journal of Biotechnology, 79 (2000) 259. [25] J. O’M Bockris, A. K. N. Reddy, and M. Gamboa-Aldeco, Modern Electrochemistry 2A, 2nd ed., Plenum Press, New York, 1998, pp. 936-943. [26] D. R. Jackson, S. Omanovic and S. G. Roscoe, Langmuir, 16 (2000) 5449. [27] T. Sen, K. K. Haldar and A. P. J. Phys. Chem. C, 112 (2008) 17945. [28] B. Pan, D. Cui, P. Xu, Q. Li, T. Huang, R. He and F. Gao, Colloids and Surfaces A: Physicochem. Eng. Aspects, 295 (2007) 217. [29] E. M. Pinto, D. M. Soares and C. M. A. Brett, J. Appl. Electrochem., 37 (2007) 1367. [30] E. M. Pinto, D. M. Soares and C. M. A. Brett, Electrochimica Acta, 53 (2008) 7460. [31] S. Omanovic and S. G. Roscoe, Langmuir, 15 (1999) 8315. [32] S. E. Moulton, J. N. Barisci, A. Bath, R. Stella and G. G. Wallace, Langmuir, 21 (2005) 316. [33] UNC-Chapel Hill Chemistry Fundamentals Program, http://www.shodor.org/unchem/index.html [34] M. C. Cortizo, M. Fernandez, L. D. Mele and A. M. Cortizo, Biological Trace Element Research, 100 (2004) 151. [35] J. C. Wataha, S. K. Nelson and P. E. Lockwood, Dental Materials, 17 (2001) 409.  52 [36] S. M. Hagsater, T. G. Jensen, H. Bruus and J. P. Kutter, Lab Chip, 7 (2007) 1336. [37] S. M. Hagsater, A. Lenshof, P. Skafte-Pedersen, J. P. Kutter, T. Laurell and H. Bruus, Lab Chip, 8 (2008) 1178. [38] D. A. Christensen, Ultrasonic Bioinstrumentation, John Wiley & Sons, New York, 1988, Chapter 3. [39] Q. Wan, T. Wu, J. Chastan, W. L. Roberts, A. V. Kuznetsov and P. I. RO, Flow Turbul. Combust., 74 (2005) 195. [40] S. Hyun, D. R. Lee and B. G. Loh, Int. J. Heat Mass Tran., 48 (2005) 703. [41] R. M. Moroney, R. M. White and R. T. Howe, Appl. Phys. Lett., 59 (1991) 774. [42] N. T. Nguyen and R. M. White, Sensor Actuat. A-Phys., 77 (1999) 229–236. [43] W. Kao, Biomaterials, 21 (2000) 2295. [44] Although a DC voltage could potentially be used to reduce protein adsorption, a disadvantage could be interfacial electrochemistry that occurs between the electrode and the solution. By using a high-frequency AC voltage, no electrochemical reaction can occur in the electrode-electrolyte interface. Typically, the voltage scan rate used in an impedance analyzer is 100 V/s, so the time required to conduct a chemical reaction is much longer than the charging time of a PZT by AC voltage. [45] J. R. Cogdell, Foundations of Electrical Engineering, Prentice Hall, Englewood Cliffs, 1990. [46] J. Israelachvili, Intermolecular & Protein-surface Forces, 2nd ed., Academic Press, London, 1991, pp 176-212. [47] J. Wang, Analytical Electrochemistry, 2nd ed., John Wiley & Sons, New York, 2000; pp 18- 25. [48] M. Zhang, T. Desai and M. Ferrari, Biomaterials, 19 (1998) 953. [49] J. J. Marra, Colloid Interf. Sci., 107 (1985) 446. [50] S. I. Jeon and J. D. Anderade, J. Colloid and Interface Sci., 142 (1991) 159. [51] E. Blomberg, P. M. Claesson and J. C. Froberg, Biomaterials, 19 (1998) 371.      53 CHAPTER 3 AN INVESTIGATION ON VIBRATION-INDUCED PROTEIN DESORPTION MECHANISM USING MICROMACHINED MEMBRANE AND PZT PLATE 2134  3.1 Introduction Biofouling is a process that starts immediately after a foreign surface comes to contact with biological fluids and is considered to be one of the greatest challenges in the field of in vivo biosensing [1-3]. Proteins, cells, and other biological components adhere to the sensor surface after the implantation to form an avascular fibrous capsule to encapsulate the device and will eventually impair the function of the bio-sensor [4]. Adsorbed proteins on the surface is believed to be responsible for the modulation of this healing response through a series of complex events that include immune and inflammatory response, formation of granulation tissue, and scar formation [5]. Numerous studies have indicated that surface chemistry and microstructure of an interacting foreign surface may modulate the protein adsorption [2,3,6,7] which eventually determines host response. Thus, prevention or manipulating the protein adsorption to the surfaces is very critical to increase the life of implantable devices.  Many different approaches have been used in the past to prevent protein adsorption on the surface. Examples include polymer based coatings [5,8-16], surface functionalization using gas plasma [17], surface charges [18], detergent coating [19], nanostructured surfaces [6,20,21], and  21A version of this chapter had been published. Yeh, P. Y., Le, Y., Kizhakkedathu, J. N. and Chiao, M. (2008) An Iinvestigation of Vibration-induced Protein Desorption Mechanism Using a Micromachined Membrane and PZT Plate on Biomed. Microdevices 10:701-708.  54 mechanical vibration [6,22-24]. Among these methods, passive methods (without surface treatment) are more attractive for certain applications [25-27]. The use of mechanical means to reduce non-specific protein binding had been investigated by researchers [6,22,23]. It was observed that the magnetic nanowires, when oscillated by an external oscillating magnetic field, adsorb less proteins compared to static nanowires [6]. By using a piezoelectric material in the construction of the sensor is another approach to combat biofouling [22,23]. On an oceanographic sensor, the application of a mechanical wave induced by polyvinylidene fluoride (PVDF) was shown to reduce biofouling [23]. High shear waves from high-frequency (~MHz) macro scale quartz resonators (thickness-shear mode) with a 3.5 W output can attenuate non-specific proteins binding [22]. An external fluid flowing through a surface had also been shown to attenuate the adsorption of colloids and biomolecules on the surface [20,28]. Clinically, a glucose sensor had demonstrated to have longer life and less drifting sensor reading in vivo with a continuous saline flow over the sensor surface than regular glucose sensors without saline flow [20]. In addition, ultrasound-induced acoustic streaming (flow) has also been shown to be able to remove μm-sized particles from the surface [28]. One of the main advantages of these methods is that there is no surface treatment is involved, so the concept can potentially be adapted to various surfaces. Previously, using a conventional macro scale piezoelectric plate, our research group has shown the combined effect of acoustic flow initiated from mechanical vibration and surface electrostatic repelling can attenuate adsorption of BSA and IgG [24]. In the current research, we demonstrate the attenuation of adsorption of single proteins (BSA and IgG) and proteins from blood plasma by the vibration of micromachined membrane by a miniature PZT plate. The    55 vibration of membrane is also studied and simulated by ANSYS. By scaling down the device size and investigation in blood plasma makes study more closer to ‘in vivo’ situation. Figure 3.1 shows the schematic diagram of a micromachined piezoelectric membrane that can be integrated with a biosensor (not shown). The vibrating membrane couples with the fluid to generate shear flow near the membrane. The fluid access via in the membrane allows fluids to pass through to the sensor.   Figure 3.1 Schematic diagram of the proposed piezoelectric membrane as an implantable sensor coating. Proteins can be desorbed by surface charge and acoustic streaming force generated by vibration.    56 3.2 Materials and Methods 3.2.1 Materials Bovine serum albumin-fluorescein isothiocyanate (power, BSA-FITC, A9771) and anti- mouse goat immunoglobulin G-FITC (solution, IgG-FITC, F5265) were purchased from Sigma- Aldrich and used without further purification. The lead zirconate titanate plates (PZT, SM10-2525- 00, thickness 0.5mm) were purchased from Sensor Technology Ltd (Ontario, Canada). and were diced into 3mm x 1mm x 0.5mm plates. Silicon-on-insulator (SOI) wafers were purchased from Ultrasil Inc. with 2/1/400 μm silicon/SiO2/silicon layers. Alexa FluorTM 488 labeled blood plasma was prepared and purified by following a protocol from Invitrogen. The flurophore density of labeled plasma was measured and was equal to 1.57 mole of dye per mole of protein. The plasma was diluted in 0.1M NaHCO3 and phosphate buffer saline (PBS, 0.01M potassium phosphate, 0.15M NaCl, 0.2mM NaN3, pH 7.2), and the final concentration is 2.8 mg/ml.  3.2.2 Device Fabrication and Characterization Tetra methyl ammonium hydroxide (TMAH) was used to anisotropically etch a SOI wafer to form a SiO2/silicon membrane (2000μm x 500μm x 3μm). The PZT plate (3000μm x 1000μm x 500μm) was then attached to the silicon substrate by conducting silver epoxy (R.P. Electronic Components Ltd. BC Canada). Similar approaches have been adopted in fabricating PZT-based micropumps [29,30]. The silver epoxy is applied only around the silicon membrane, hence, there is no bonding between PZT and the silicon membrane but the electrical signal can be applied on bottom electrode of PZT through the silver epoxy. Another electrical signal was connection to the top electrode by a bond wire. The membrane vibration is generated by vibration of PZT plate and the vibration response of the membrane was measured by a laser doppler vibrometer (LDV,  57 Polytec DFV-5000). The LDV measured membrane vibrating velocities at corresponding frequencies. The velocity measurement has higher signal-to-noise ratio than displacement measurement. The displacement of the membrane is then calculated by dividing velocity by 2πf, where f is the corresponding frequency. The frequency is swept from 120 kHz ~ 900 kHz at 10 Vpp by a function generator (Agilent 33220A) and the frequency response from the LDV is monitored by an oscilloscope (Tektronix TDS420). The vibration of piezoelectric membrane is activated by an AC signal and the membrane operates in a FPW (flexural plate wave) mode since the membrane thickness is much smaller than the wavelength of FPW. The advantage of FPW mode is that phase velocity is lower than the wave velocity in fluid. Hence, the energy can be restricted within an evanescent decay length and can be transferred to generate acoustic streaming near the surfaces [31]. To generate FPW motion using a bulk PZT plate be found in Guo and Lal, 2001 [32]. Other ways to achieve FPW actuation without using bulk PZT plates include deposition of an interdigitated electrode on a piezoelectric film [33,34]. Simulation was performed by a finite element software (ANSYS 10.0TM). A solid model representing the Si/SiO2 membrane (2000μm x 500μm x 3μm)/PZT plate (3000μm x 1000μm x 500μm), was built.A 8-nodes, 3D coupled field (piezoelectricity capable) solid element, “solid 5” and a 8-nodes, 3D structural solid element, “solid 45” were used for PZT plate and Si/SiO2 membrane, respectively. The simulation was done in vacuum, however, for an actual device, one side of membrane is contacting with fluid. Therefore, the simulation needs to be modified for approximating the actual fluid loading conditions. Theoretically, if d (the thickness of a membrane vibrating in FPW mode) is << λ (the wavelength of FPW) and vibrates asymptotically, the frequency (f) can be approximated as [31,35]:  58      where T is the in-plane tension. Β and M are the bending stiffness and mass per unit length of the membrane, respectively. ρ is the density of the surrounding fluid, and δ is evanescent decay length, which can be written as [31,35]:                                                                                                                                                      (3.2) where Pν  is the phase velocity of sound in the membrane, and Fν  is the sound velocity in the fluid. For a membrane vibrating in a fluid, δ is close to λ/2π [31], since phase velocity of sound in the membrane is lower than the sound velocity in fluids. From Equation (3.1), the frequency spectrum simulated from ANSYS was used to evaluate T and B from the resonant frequency at corresponding wavelength (λ), M is calculated from knowing the thickness and density of composite of the membrane. Then the modified f for membrane vibrating in water can be calculated through Equ. (3.1).  3.2.3 Protein Adsorption and Vibration The protein adsorption experiment was done in PBS buffer (pH 7.4). The following protein solutions were used: BSA (2 mg/ml, IgG (0.1 mg/ml) and human plasma (0.3 mg/ml protein concentration) in PBS. Protein solutions were filled into the cavity in the back side of the membrane and were incubated at room temperature for a specific period of time to allow proteins to adsorp. The cavity was then washed by PBS solution (5 times) to remove any unadsorbed or loosely bound proteins from the chamber surface. To desorb proteins, the PZT plate was vibrated for a period of time in PBS solution, and then was washed by PBS 5 times. 1  T B f Mλ ρ δ += + 2   1 ( ) 2 p F v v λδ π= − (3.1)  59 Fluorescence detection was used for determining the amount of adsorbed protein [36,37]. Fluorescence images of the membrane surfaces before and after protein adsorption, and after further vibration were taken using a Nikon eclipse TE 2000-U fluorescence microscope with X- Cite 120 fluorescence illumination system (FITC filter and DS-U1 suit digital camera). The detailed procedure of analysis of fluorescence intensity to relative protein adsorption can be found in our previous work [24].  3.2.4 Scanning Electron Microscope (SEM) Surfaces were imaged using a Hitachi S-3000N scanning electron microscope with a tungsten electron source. The accelerating voltage can be varied over the range of 0.5~30 kV. The device image is taken without coating and tilt.  3.3 Results and Discussions The proof of concept setup of PZT on a micromachined membrane is shown in Figure 3.2(a), and the SEM micrograph of front view and back view of the device are shown in Figure 3.2(b) and (c), respectively. The cavity in Fig. 3.2(c) was filled with protein solution or PBS, such that the backside of the membrane  will contact the solution in experiments. In a FPW mode (the thickness of membrane is 3 μm, and the wavelength ~4000 μm), the membrane was constrained by four peripheral ends and moves like a standing wave.      60                Figure 3.2 (a) The experimental setup (b) Front side view of a micromachined PZT plate (3000μm x 1000μm x 500μm) on membrane with an electrical wire bonded onto the plate surface (c) The backside view of the Si membrane. A thin SiO2 (1μm) layer remained on the surface. In the following experiments, the proteins are adsorbed onto the backside of the Si membrane (with a SiO2 surface).      (a) Plastic cover PBS Silicon Membrane (SiO2/Si) Silver Epoxy PZT To function generator  (b) (c) 500μm 2000μm V ~ ~ ~~ ~ ~ ~ ~ ~ ~ ~ ~ PZT Silicon wire  61 Figure 3.3 shows the vibration spectrum at the center of a Si/SiO2 membrane/PZT plate, when contacting with a PBS solution from the backside. The PZT was driven by a 10 Vpp sinusoid waveform from 120~900 kHz. The measured resonant frequency was 308 kHz (Fig. 3.3) and the measured out-of-plane vibration amplitude was 105 nm. The simulated frequency spectrum is shown in Fig. 3.3. The simlated resonance frequency is 276 kHz, which is very close to the experimental value of 308 kHz with a 10.4% error. The error between simulation and experiment may be coming from the simplification of evanescent decay length (δ), real dimension of membrane, parameters of simulation, and not perfectly flat membrane. If a phase velocity (354 m/s, product of the resonance frequency and the corresponding wavelength of vibrating membrane) and sound velocity in water (1948 m/s) were put into Equation (3.2), a more accurate δ can be obtained. The simulated frequency then shifted from 276 kHz to 280 kHz with the error 9.1% compared to the measured value.   Figure 3.3 The simulation and experimental vibration spectrum of the PZT plate/silicon membrane. The vibration at 308, 320, 500, and 575 kHz correspond to 1.5, 2.5, 3.5 wavelength bending mode, and longitudinal mode, respectively.    0 200 400 600 800 0 20 40 60 80 100   R el at iv e Vi br at io n Am pl itu de Frequency (kHz)  Experiment  Simulation  62 In addition, the displacement at different locations on the membrane at the resonance frequency was measured by moving the laser spot of LDV from center to left periphery of the membrane along the center line compared with simulated membrane displacement at 276 kHz. As shown in Figure 3.4, the displacement is plotted versus the distance measured from the center point of the membrane.  0 200 400 600 800 1000 0 20 40 60 80 100 120   Experiment  Simulation R el at iv e Vi br at io n Am pl itu de Distance (μm)  Figure 3.4 The simulation and experimental vibration amplitude distribution across the Si/SiO2 membrane along the membrane center line. X axis is given as the distance from the center of membrane to the left periphery. The half length of the membrane is 1000μm.  The kinetic adsorption isotherms of BSA and IgG were constructed from the intensity of fluorescent images taken at various incubation times and is shown in Figure 3.5. The initial adsorption rate is high and reached saturation. The saturation time value is consistent with other  63 reports in the literature [38]. Based on the isotherms, we have chosen the 1 hr incubation time for subsequent experiments as protein adsorption reaches a plateau value.                  Figure 3.5 Kinetic adsorption isotherms of BSA and IgG on a SiO2 surface. The protein solution concentrations were 2 mg/ml for BSA and 0.1 mg/ml for IgG. After each different protein incubation time, the surface was washed with PBS 5 times before the fluorescence intensity measurements. Equation (A.1) was used to fit the data.  Figure 3.6 shows the results of relative protein desorption due to vibration. About 57 ± 10% of BSA and 47 ± 13% of IgG were removed from the membrane surface after the vibration is applied. The values are close to our previous work [24] using macroscale PZT plates.  0 10 20 30 40 50 60 0 50 100 150 200 R2=0.96 R2=0.99  Fl uo re sc en ce  In te ns ity Incubation Time (mins)  IgG  BSA  64 Control 1 Vib 1 Control 2 Vib 2 0 20 40 60 80 100 120   R el at iv e Fl uo re sc en ce  In te ns ity  Figure 3.6 Comparison of relative fluorescence intensities, which linearly depend on protein adsorption quantity, with and without PZT vibration. The vibration was initiated by a 10 Vpp AC signal at 308 kHz for 5 min. The proteins were incubated in 2 mg/ml BSA and 0.1 mg/ml IgG for 1 hr. Control 1 and 2 are the surfaces incubated in BSA and IgG solution, respectively. For Vib 1 and Vib 2, BSA and IgG adsorbed surfaces respectively were used. The fluorescence intensity at every condition was measured after washing the surface after that condition by PBS for 5 times.   In addition, we have tested two other vibration conditions with smaller membrane vibrating amplitude for protein desoprtion: (1) lower AC voltage input (1 Vpp) at 308 kHz (resonant frequency) and (2) 10 Vpp at 400 kHz (off resonant frequency) and the results are given in Figure 3.7. Compared to the vibration at 10 Vpp and 308 kHz, these two cases had less amount of proteins desorbed from the surface. Protein desorption from surface needs two interweaving directional forces, out-of-plane and horizontal (in-plane) forces [24]. Shear stress from the fluid  65 flow contributes to the horizontal force, and is decreasing with decreased vibrating amplitude [39]. Surface charge contrbutes to the out-of-plane force to repel proteins away from surface [18]. The surface charge comes from the applied voltage on a dielectric material. The charge accumulation was observed by measuring the surface voltage of SiO2 side of membrane in PBS when applied DC or AC voltage to PZT plate. The measured voltage is 65% of the real voltage applied directly on PZT. In DC application, 3.33 V was measured on the SiO2 surface when 5 V is applied on the PZT plate. In AC application, 6.5 Vpp was measured on the SiO2 surface when 10 Vpp is applied on the PZT plate.  Figure 3.7 Relative fluorescence intensities, which linearly depend on protein adsorption quantity, show a strong dependence on the vibration amplitude of the membrane. Protein desorption cannot be effective if the surface vibrated at off-resonance frequency or under low voltage AC signals application. In all the experiments, samples were incubated in BSA (2 mg/ml) for 1 hr and vibration conditions were applied for 5 min. Conditions for Vib 1, Vib 2, and Vib 3 were 308 kHz 10 Vpp, 308 kHz 1 Vpp, and 400 kHz 10 Vpp, respectively. The fluorescence intensity at every condition was measured after washing the surface after that condition by PBS for 5 times.    Control Vib 1 Vib 2 Vib 3 0 20 40 60 80 100 120   R el at iv e Fl uo re sc en ce  In te ns ity  66 A DC voltage was applied on the PZT plate after BSA adsorption on the SiO2 surface. The relative protein desorption at different DC voltage is listed in Table 3.1. The BSA adsorption was normalized with respect to the sample without vibration. The relative protein adsorption after applying a -2 V DC voltage for 1 min and 3 min were 82.94 ± 9.3 % and 84.28 ± 9.4 %, respectively. The BSA is negatively charged at pH 7.2 in PBS, and the surface is also negatively charged because of the -2V DC application. Furthermore, less BSA adsorption (69.04 ± 8.4%) was observed by combining DC application and three additional washing steps during DC application.  Table 3.1 Protein adsorption on surfaces with different conditions          Significant statistic differences  (p<0.05) between 1&2, 1&3, and 1&4  If an AC sinusoidal voltage is used, there is negative and positive voltage alternatively. The half period of the AC signal (only negative portion) operated in resonance frequency is (1/308 kHz)/2 = 1.623*10-3 ms. The time constant for charging the surface is equal to the system resistance (mainly from connection of wires and PZT) multiplied by the capacitance of the membrane (SiO2 (1μm)/Si (2μm)). The resistance between the wires and the metal electrode in PBS is 2 to 5 Ω and the capacitance of SiO2/Si membrane is 2.3*10-11 F. Hence, the time Conditions Normalized Fluorescence Intensity 1 BSA 100 ± 4.2 n = 4 2 DC (2V) 1 min 82.94 ± 9.3 n = 4 p < 0.05 3 DC (2V) 3 min 84.28 ± 9.4 n = 4 p < 0.05 4 DC (2V) 1 min, 3 washes 69.04 ± 8.4 n = 4 p < 0.05  67 constant for charging is 4.6~11.5*10-8 ms [40]. The much longer half period of 308 kHz AC signal than time constant means that surface can be fully charged by AC signal with operating frequency 308 kHz. The negative average voltage of a 6.5 Vpp (voltage we measured if 10 Vpp is applied) AC signal is -2*3.25/π = -2.1 V [40], and this is large enough to desorb BSA from SiO2 surface and is supported by the experiments with DC voltage ( Table 3.1). The vibration effect on attenuation of protein adsorption from blood plasma was also investigated. Plasma consists of mainly albumin (50~60%), immunoglobulins (15~35%), fibrinogen (~3%), and other regulatory proteins [41]. In our experiment, 0.3 mg/ml totoal protein concentration, corresponds to 0.15~0.18 mg/ml, 0.045~105 mg/ml, and 0.009 mg/ml  albumin, immunoglobulin, and fibrinogen, respectively. About 55.3 ± 5% of plasma proteins were desorbed from surface due to vibration (Figure 3.8(a)) compared to that without vibration. The vibrated surface (protein desorbed) is further incubated with plasma for another 20 min and vibrated again to remove the adsorbed plasma proteins. The second cycle of adsorption and vibraton resulted in the removal of 56.4 ± 12% of adsorbed proteins shows the continuous protein desorption can be achieved by appropriate vibration strategies. We have also carried out another experiment where the surface was exposed to plasma proteins for 4 hr (long term) and vibrated at different conditions (Figure 3.8(b)). Results show that ~ 59.2 to 72.1 % of the adsorbed proteins can be removed by the vibration at various conditions. A comparision of relative flurescence intensity of the control samples at 1 hr incubation (Fig. 3.8(a)) and 4 hr incubation (Fig. 3.8(b)) shows that the 4 hr incubated sample has ~69.5 % higher protein adsorption than 1 hr incubated sample. Neverthless, vibration removed most of the adsorbed protein even for extended incubation periods.   68 Control Vib 1 Adsorption 1 Vib 2 0 20 40 60 80 100 120 140 160  R el at iv e Fl uo re sc en ce  In te ns ity  Control Vib 1 Vib 2 Vib 3 Vib4 0 20 40 60 80 100 120  R el at iv e Fl uo re sc en ce  In te ns ity  Figure 3.8 Comparisons of relative fluorescence intensities, which linearly depend on plasma adsorption quantity at (a) different times without and with PZT vibration, (b) 4 hr without and with PZT vibration in different media and time intervals. The conditions in (a) are Control- incubation in plasma protein solution for 1 hr; Vib 1-vibration application in PBS for 5 min after incubation; Adsorption 1-second plasma protein incubation for 20 min after Vib 1 on the sample sample; Vib 2- vibration application in PBS for 5 min after Adsorption 1 on the sample sample. The conditions in (b) are Control- Incubation in plasma protein solution for 4 hr; Vib 1- vibration application in plasma prtein solution for 5 min; Vib 2-vibration application in PBS for 5 min after Vib 1; Vib 3- vibration application in PBS for 10 min after Vib 2; Vib 4- vibration application in PBS for 20 min after Vib 3. All vibrations are applied by a 10 Vpp AC signal at 308 kHz. The plasma protein concentration was 0.3 mg/ml. The fluorescence intensity at every condition was measured after washing the surface after that condition by PBS for 5 times.  3.4 Conclusions In summary, we have demonstrated a proof-of-concept study on protein desorption by a vibrating Si based membrane. The membrane vibrates in a FPW mode standing wave with 4000/3 μm longitudinal wavelength constrained at two ends at 308 kHz. A simulation by ANSYS shows good match on resonant frequency (276 kHz) of maximum vibration amplitude (a) (b)  69 and membrane vibrating amplitude distribution at this frequency. In vitro protein desorption experiments show that ~57 ± 10% of the adsorbed bovine serum albumin (BSA) ,47 ± 13%  of   the immunoglobulin G (IgG), and 55.3 ± 5% of the plasma proteins can be achieved by vibration. With DC application, the surface charge effect on protein desorption has also been verified. The present method offers the possibility to attenuate protein adsorption without or in combination with other surface treatment methods. Other piezoelectric materials such as ZnO can also be used in bulk or film form in future for MEMS-based membrane for implantable biosensor applications. A micromachined vibrating membrane is used to remove adsorbed proteins on a surface. A lead zirconate titanate (PZT) composite (3 mm x 1 mm x 0.5 mm) attached on a silicon membrane (2000 μm x 500 μm x 3 μm) and vibrates in a flexural plate wave (FPW) mode with wavelength of 4000/3 μm at the resonant frequency of 308 kHz. The surface charge on the membrane and fluid shear stress contribute in minimizing the protein adsorption on the SiO2 surface. In vitro characterization shows that 57±10% of the adsorbed bovine serum albumin (BSA), 47 ± 13% of the immunoglobulin G (IgG), and 55.3~59.2 ± 8% of the proteins from blood plasma are effectively removed from the vibrating surface. A simulation study of vibration-frequency spectrum and vibrating amplitude distribution matches well with the experimental data. Potentially, a MEMS-based vibrating membrane could be the tool to minimize biofouling in vivo.     70 3.5 Bibliography [1] D. M. Fraser, in D. M. Fraser (Eds.), Biosensors in the Body: Continuous in vivo Monitoring, 1st ed., John Wiley & Sons, Chichester, 1997, Chapter 1. [2] N. Wisniewski, F. Moussy, and W. M. Reichert, Fresenius J. Anal. Chem., 366 (2000) 611. [3] N. Wisniewski and M. Reichert, Colloids Surf. B Biointerfaces, 18 (2000) 197. [4] B. D. Ratner and S. J. Bryant, Annu. Rev. Biomed. Eng., 6 (2004) 41. [5] A. Whelan and F. Regan, J. Environ. Monit., 8 (2006) 880. [6] K. M. Ainslie, G. Sharma, M. A. Dyer, C. A. Grimes, and M. V. Pishko, Nano Lett., 5 (2005) 1852. [7] P. Roach, D. Farrar, and C. C. Perry, J. Am. Chem. Soc., 128 (2006) 3939. [8] R. Gref, M. Lück, P. Quellec, M. Marchand, E. Dellacherie, S. Harnisch, T. Blunk, and R. H. Müller, Colloids Surf. B Biointerfaces, 18 (2000) 301. [9] N. P. Huang, R. Michel, J. Vörös, M. Textor, R. Hofer, A. Rossi, D. L. Elbert, J. A. Hubbell, and N. D. Spencer, Langmuir, 17 (2001) 489. [10] C. M. Nolan, C. D. Reyes, J. D. Debord, A. J. Garcia, and L. A. Lyon, Biomacromolecules, 6 (2005) 2032. [11] R. G. Nuzzo, Nat. Mater., 2 (2003) 207. [12] K. L. Prime and G. M. Whitesides, J. Am. Chem. Soc., 115 (1993) 10714. [13] C. Allen, N. D. Santos, R. Gallagher, G. N. C. Chiu, Y. Shu, W. M. Li, S. A. Johnstone, A. S. Janoff, L. D. Mayer, M. S. Webb, and M. B. Bally, Biosci Rep., 22 (2002) 225. [14] E. A. S. Doherty, R. J. Meagher, M. N. Albarghouthi, and A. E. Barron, Electrophoresis, 24, (2003) 34. [15] H. Jacobs, D. Grainger, T. Okano, and S. W. Kim, Artif. Organs., 12 (1988) 506. [16] H. Makamba, J. H. Kim, K. Lim, N. Park, and J. H. Hahn, Electrophoresis, 24 (2003) 3607. [17] G. B. Sigal, M. Mrksich, and G. M. Whitesides, J. Am. Chem. Soc., 120 (1998) 3464. [18] S. Pasche, J. Vörös, H. J. Griesser, N. D. Spencer, and M. Textor, J. Phys. Chem. B, 109 (2005) 17545. [19] M. Zhang and M. Ferrari, Biotech. Bioeng., 56 (1997) 618. [20] G. P. Rigby, P. W. Crump, and P. Vadgama, Analyst, 121 (1996) 871.  71 [21] G. P. Rigby, P. W. Crump, and P. Vadgama, Med. Biol. Eng. Comput., 33 (1995) 231. [22] G. D. Meyer, J. M. Morán-Mirabal, D. W. Branch, and H.G. Craighead, IEEE Sens. J., 6 (2006) 254. [23] M. Rahmoune and M. Latour, Smart Mater. Struct., 4 (1995) 195. [24] P. Y. Yeh, J. N. Kizhakkedathu, J. D. Madden, and M. Chiao, Colloids Surf. B Biointerfaces, 59 (2007) 67. [25] A. L. Klibanov, V. P. Torchilin, S Zalipsky, in V. P. Torchilin and V. Weissig (Eds.), Liposomes, Oxford University Press, Oxford, 2003, p. 231-265. [26] S. Sharma, R. W. Johnson, and T. A. Desai, Biosens. Bioelectron., 20 (2004) 227. [27] H. Zhu and M. Snyder, Curr. Opin. Chem. Biol., 7 (2003) 55. [28] Q. Qi, and G. J. Brereton, IEEE T. Ultrason. Ferr., 42 (1995) 619. [29] S. C. Chen, C. H. Cheng, and Y. C. Lin, Sens. Actuators A Phys., 135 (2007) 1. [30] J. S. Yoon, J. W. Choi, H. Lee, and M. S. Kim, Sens. Actuators A Phys., 135 (2007) 833. [31] D. S. Ballantine, R. M. White, S. I. Martin, A. I. Ricco, E. T. Zellers, G. C. Frye, and H. Wohltjen, Acoustic Wave Sensors: Theory, Design, and Physico-Chemical Applications, Academic Press, San Diego, 1997, Chapter 3. [32] H. Guo and A. Lal, in IEEE Ultrasonics Symposium (IEEE, Atlanta, 2001), p. 799. [33] A. W. Wang, R. Kiwan, R. M. White, and R. L. Ceriani, Sens. Actuators B Chem., 49 (1998) 13. [34] N. T. Nguyen and R. M. White, Sens. Actuators A Phys., 77 (1999) 229. [35] M. S. Weinberg, C. E. Dube, A. Petrovich, and A. M. Zapata, J. Microelectromech. Syst., 12 (2003) 567. [36] T. T. Huang, J. Sturgis, R. Gomez, T. Geng, R. Bashir, A.K. Bhunia, J.P. Robinson, and M.R. Ladisch, Biotechnol. Bioeng., 81 (2003) 618. [37] R. Bashir, Adv. Drug Deliv. Rev., 56 (2004) 1565. [38] M. Malmsten, and Bo Lassen, in T. A. Horbett, and J. L. Brash (Eds.), Protein at Interfaces II, American Chemistry Society, Washington D.C., 1995, Chapter 16. [39] N. T. Nguyen, A. H. Meng, J. Black, and R. M. White, Sens. Actuators A Phys., 79 (2000) 115. [40] J. R. Cogdell, Foundations of Electrical Engineering, Prentice Hall, Englewood Cliffs, 1990. [41] B. Blombäck, and L. A. Hanson, Plasma Proteins, Pitman Press, Bath, 1975, p. 17-21.  72 CHAPTER 4 ATTENUATION OF PROTEIN ADSORPTION BY USING SELF- ASSEMBLING MONOTHIOL-TERMINATED POLYETHYLENE GLYCOL AND VIBRATION 315   4.1 Introduction  Protein adsorption on biomedical devices initiates a biofouling response, which involving a cascade of host reactions, including platelet activation, blood coagulation, and complement activation [1,2]. The implanted medical devices, such as biosensors [3] and drug delivery systems [4] typically fail due to biofouling within one month of implantation. Many approaches have been used to prevent these non-specific biological interactions [5-8]. For example, hydrophilic polymer based coatings [9-11] have been used as an anti-biofouling agent for biosensors and drug delivery systems, which demonstrated reduction of biofouling and improved biocompatibility [12-14]. In addition, such coatings have shown to extend the life-span of biomedical devices [15,16]. There are many methods of applying hydrophilic polymer coatings. The most common method uses the affinity between sulphur of thiol PEG and gold, silver, and copper atoms [17-19]. Covalently polymer bonding has also been investigated by researchers [20,21]. Generally, grafting density, chain length and chain morphology are the important parameters in the attenuation of protein adsorption by a polymer coating [17,22]. Sofia et al. found that  31A version of this chapter will be submitted for publication. Yeh, P. Y., Chiao, M. and Kizhakkedathu, J. N. Attenuation of Protein Adsorption Using a Combination of Polymer Grafting and Application of Mechanical Vibration.   73 differing protein adsorption behavior occurs in linear and branched polymer chains [22]. Unsworth et al. investigated the effects of chain density and chain length on the adsorption of proteins [17]. Other parameters, temperature, for example, also change the ability of polymer to attenuate protein adsorption [12]. Application of mechanical vibration was reported to reduce non-specific protein adsorption. Meyer et al. used high-frequency quartz resonators to attenuate non-specific protein binding and increase the sensitivity of a biosensor [23]. Ainsile et al. activated the vibration of a magnetic nanowire assembled surface by an external oscillating magnetic field to reduce protein adsorption [24]. Recently, our group investigated the combined electrostatic and vibration effect initiated by PZT, a piezoelectric material, to attenuate protein adsorption [25]. One of the main advantages of applying vibration is that no surface treatment is involved, so the concept can potentially be adapted to various surfaces. In this chapter, we combine a linear hydrophilic polymer surface coating (PEG) with mechanical vibration to achieve a high level of attenuation of adsorption of BSA, IgG, and plasma proteins. We show that the combined effect is synergistic and highly effective for BSA and IgG attenuation at very low graft densities of PEG chain on the surface; at such low graft densities without vibration, the surface adsorbs large amount amounts of BSA and IgG. With the help of vibration, there is further attenuation of protein adsorption on surfaces at almost every grafting density of PEG. This is a highly critical result in the search for effective methods to prevent biofouling in both in vitro and in vivo applications.     74 4.2 Materials and Methods 4.2.1 Proteins and Polymers Bovine serum albumin-fluorescein isothiocyanate (BSA-FITC, A9771) and anti-mouse goat immunoglobulin G-FITC (IgG-FITC, F5265) were purchased from Sigma-Aldrich and used without further purification. The lead zirconate titanate plates (PZT, SM10-2525-00, thickness 0.5mm) were purchased from Sensor Technology Ltd. and were diced into 0.5cm x 1cm plates. The methoxy poly (ethylene glycol) thiol powder (mPEG-SH, CH3O-(CH2CH2O)127- CH2CH2CH2CH2-SH, average molecule weight 5703 Da, density 1.08 g/cm3 at 20 ºC) was purchased from Nektar Therapeutics. PEG is used in this chapter to represent mPEG-SH. Absolute ethanol was purchased from the University of British Columbia’s chemical store and was used to dissolve PEG powder. Alexa FluorTM 488 labeled blood plasma was prepared and purified by following a protocol from Invitrogen. The plasma was diluted in 0.1 M NaHCO3 and phosphate buffer saline (PBS, 0.01 M potassium phosphate, 0.15 M NaCl, 0.2 mM NaN3, pH 7.2), with a final concentration of 2.8 mg/ml. The plasma proteins were dyed with Alexa Fluor (ratio of mole dye/mole of protein was 1.57).  4.2.2 Substrate Preparation The PZT plates were coated both sides with 300 nm thick titanium layers (titanium pellets, 300 nm, 99.995% purity, Kurt J. Lesker company, EVMTI45) as electrodes for voltage application. Evaporated SiO2 from bulk SiO2 was then applied as a 50 nm thick insulation layer on one side of the PZT.. A 5 nm thick chromium layer (chromium pellets, 99.95 % purity, Kurt J. Lesker company, EVMCR35D) was evaporated as an adhesion layer. A 50 nm thick gold layer  75 (gold pellets, 99.999% purity, Kurt J. Lesker company, EVMAUXX50G, diameter:1/8”, length:1/8”) was evaporated as a functional layer for PEG modification. All the evaporated materials were deposited by an evaporator with an electron beam source. Illustration 4.1 shows a cross-section of the surface coatings (Ti/PZT/Ti/SiO2/Cr/Au) and preparation of the patterned PEG grafted surface. The patterned PEG grafted surface was done by photolithography using a Canon PLA-501F mask aligner. The AZ 4110 photoresist used for lithography was purchased from Hoechst Celanese Corporation. The photoresist was further removed by acetone after PEG grafting.        Illustration 4.1 The depiction of a cross-section of Au/Cr/SiO2/Ti/PZT/Ti plates, and the scheme of grafting PEG on plain or patterned plates.  4.2.3 Polymer Modification Prior to the polymer coating, the substrates were cleaned using a freshly prepared piranha solution (H2SO4: H2O2 = 7:3) to remove organic contamination. Note that piranha solution shoul be handled with extreme care; it is a strong oxidant and reacts violently with many organic materials. It also presents an explosion danger. All work should be performed under a fume hood, Au 50 nm Cr 5 nm SiO2 50 nm Ti 300 nm PZT 500 μm Ti 300 nm Photolithography + PEG incubation PEG incubation Remove Photoresist Photoresist  76 and proper protective equipment should be worn. The substrates were rinsed with DI water thoroughly and dried with a nitrogen gun. The PEG solution was prepared by dissolving powdered PEG in 100% ethanol at 40 oC, and was kept at room temperature during polymer coating. Due to strong affinity between sulphur and gold atoms, the PEG can be chemisorbed on PZT plates. The plates were immersed in PEG solution at different initial concentrations (from 2x10-9 to 6 mg/ml) for 16 hr. We have investigated the kinetics of polymer adsorption by measuring the IR spectrum of the PEG-coated PZT plate surfaces at different PEG incubation time (9, 16, 24, and 48 hr) in 6 mg/ml PEG solution (data not shown). There were no significant changes of intensity in the IR peaks (C-O-C and –CH2 stretching) when the incubation was longer than 16 hr. Therefore, a 16-hr incubation time was chosen for subsequent experiments. The surfaces following polymer coating were washed thoroughly with 100 % ethanol to remove unbound polymer, and then the surfaces were kept in deionized water before use.  4.2.4 Protein Adsorption and Vibration Experiments The substrates were initially equilibrated with a phosphate buffered saline (PBS) buffer for 3 hr.  The protein adsorption experiments were conducted by incubating substrates with or without PEG modification in a 2 mg/ml BSA, 0.1 mg/ml IgG, and 0.3 mg/ml plasma solution in 48-well polystyrene plates at room temperature for 1 hr. Following incubation, the substrates were rinsed thoroughly with PBS (>10 times) to remove loosely adsorbed proteins. In vibration experiments, the diced PZT plates were connected to a function/arbitrary waveform generator (Agilent 33220A 20MHz) to receive a 10 Vpp sinusoidal signal. The resonance frequency of the PZT plates was 16 kHz in PBS, measured by a Laser Doppler  77 Vibrometer (LDV, Polytec DFV-5000); detailed resonant frequency findings have been documented in our previous work [25].  4.2.5 Surface Characterization 4.2.5.1 X-ray Photoelectron Spectroscopy (XPS) Analysis Surface chemical analysis was performed through X-ray photoelectron spectroscopy (XPS, Leybold MAX200 X-ray Photoelectron Spectrometer XPS/ESCA Dual anode (Mg Kα and Al Kα) achromatic X-ray source), using an Al Kα X-ray source (1486.6 eV). The samples (0.5cm x 1 cm) were mounted on the XPS stage. The XPS spectra of all studied elements, such as C 1s, O 1s, S 2p, and Au 4f were measured using a constant analyzer with 48 eV pass energy. All binding energies (BEs) were referenced to the Au 4f peak (BE, 84 eV).  4.2.5.2 ATR-FTIR Spectroscopy Surface ATR-FTIR spectra were collected on a Nexus 670 FT-IR ESP (Nicolet Instrument Corp., Waltham, MA) with a MCT/A liquid nitrogen cooled detector, KBr beamsplitter, and a MkII Golden Gate Single Reflection attenuated total reflectance (ATR) accessory (Specac Inc., Woodstock, GA). The sample stage contains a diamond window and a sapphire anvil on a torque limiting screw set to deliver 80 lbs of force. The samples were dried in an oven before analysis. IR spectra of the samples were recorded in the range of wave number 800–4000 cm−1 at room temperature.     78 4.2.5.3 Fluorescence Intensity Measurement Images of the protein-adsorbed surfaces were taken by a fluorescence microscope (Nikon eclipse TE 2000-U with X-Cite 120 fluorescence illumination system, FITC filter (since dyes are excited by blue light and emit green light) and DS-U1 suit digital camera). The green fluorescence intensity of the images from the conjugated dyes was converted to grayscale images by Adobe PhotoshopTM 6.0. The measured intensity is proposional to the relative quantity of adsorbed proteins on the surface [26]. The calibration scale was set by two points: the grayscale intensity of the bare surface was set at 0%, and the protein-adsorbed bare surface was set at 100%. The relative adsorption of proteins on various surfaces was normalized based on the calibration scale. The results were analyzed statistically by using Student’s t-test for two-group comparisons at 95% confidence levels (two tails); p values ≤ 0.05 are considered significant, which corresponds to t > 2.045 for 29 degrees of freedom (30 statistical samples) to 2.131 for 15 degrees of freedom (16 statistical samples).  4.3 Results and Discussions 4.3.1 PEG Coating and Surface Characterization The composition and atomic bonding of the surface materials was investigated by XPS scan. The PEG grafted surfaces were incubated in 2x10-7, 2x10-5, 2x10-3, 0.1 and 6 mg/ml initial PEG concentration solutions. Figure 4.1(a) shows the XPS spectra (0-600eV) of surfaces with or without PEG. From the composition analysis, Au, C, and O are the three most abundant elements. The element Au (50nm) comes from the deposited functional layer for PEG coating. Cr (5nm) and SiO2 (50nm) layers are beneath Au and can not be detected. The presence of C and O can be attributed to two sources: (1) contamination when samples were exposed to air; and (2) grafting  79 of PEG. Investigating the binding energy and C/O ratio will differentiate these two sources. The three main peaks that represent Au4f1/2, C1s, and O1s can indicate the composition change of surfaces with the changes of grafted PEG. With an increase in initial PEG concentration, the intensity of the Au4f1/2 peak decreased considerably, while the intensity of C1s and O1s peaks increased (Fig. 4.1(a)) indicating an increase in surface PEG grafting concentration. In addition, with an increase in initial concentration of PEG, the ratio of the atomic percentage of Au, C, O showed a dramatic change, as depicted in Figure 4.1(b). Fig. 4.1(b) also shows the C/O ratio of the surface incubated in 6.0 mg/ml initial PEG concentration to be 2.027. The theoretical C/O ratio for PEG (CH3O-(CH2CH2O)127-CH2CH2CH2CH2-SH,) is 2.023. This result suggests that the surface is almost fully covered by PEG chains.           Figure 4.1 (a) XPS spectra (binding energy from 0 to 600 eV) for (1) control (bare surface) and surfaces incubated in (2) 2x10-7 mg/ml, (3) 2x10-5 mg/ml, (4) 2x10-3 mg/ml, (5) 1, (6) 6 mg/ml PEG initial concentration solutions. (b) Variant atomic percentage of elements C, O and Au on surfaces incubated in different PEG initial concentration solutions. Increases of elements C and O and decrease of element Au indicate increase of PEG moieties on the surface. (a) 600 400 200 0 0.0 8.0x105 1.6x106 (5) (6) A u 4d (3 /2 ) A u 4d (5 /2 )  A u 4f (1 /2 ) O  1 s C  1 sC P S Binding Energy (eV)  (1) (2) (3) (4) (b) Control2E-7 2E-5 0.002 0.1 6 0 10 20 30 40 50 60 70 PEG initial conc. (g/l)  A to m ic  %  C  O  Au 2x10-7 2x10-5  80 High-resolution C1s scans provided more detailed information about the amount of PEG grafting as a function of PEG initial concentration. As shown in Figure 4.2, a distinct increase in C-O (presence of PEG) and decrease in C-C peaks (contamination) occurs with the PEG initial concentration. Bare Au surfaces showed small C–O and C–C peaks due to contamination. The fractional [C–O] composition in C1s peaks is considered to be the quantitative measure of PEG grafting density [3,27]. Theoretically, the amount of grafted PEG is proportional to the increased ratio of C-O/C-C [3]. However, for more precise and convenient quantification of PEG on a surface, a high-resolution scan of the Au4f1/2 peak is studied in this chapter because of the linear relationship between the natural logarithm of Au4f1/2 peak integral intensity and the thickness of surface PEG film [27]. The integral intensities of Au4f1/2 peaks of surfaces incubated in different initial PEG concentration solutions are shown in Figure 4.3(a). Theoretically, the intensity of Au4f1/2 from the gold substrate decreases exponentially with increased PEG film thickness due to the attenuation of photoelectrons in the PEG film [19,27]. This method is suitable for homogeneous film on the surface and has commonly been accepted by researchers [19,20,22,27,28], the self- assembly PEG film on the surface is the good candidate for such analysis. The thickness of PEG film can be estimated by the following equation [29]:   where h is the thickness of grafted PEG on the surface, λ is the photoelectron attenuation length of Au through grafted PEG film (λ is a constant to the thickness of coating polymer), θ is the take-off angle (angle between sample surface and the analyzer) during XPS measurement, AuO is the integral Au4f1/2 peak intensity of a bare Au surface, Au is the integral Au4f1/2 peak intensity ln ( ) sinO Au h Au λ θ= − (4.1)  81 of a PEG grafted surface. The latter is smaller because that the amount of photoelectron of Au is attenuated due to passing a PEG film. Substitute AuO, Au, and θ (90o in our XPS measurement) into Equation 4.1, the value of d/λ versus the initial PEG concentration (with its corresponding Au) can be plotted, and is shown in Figure 4.3(b). The thickness of PEG (since λ is a constant) increased dramatically at higher initial concentrations (> 0.1 mg/ml). When the PEG grafting space is larger than two times the Flory radius (the radius of the occupation of PEG), the PEG exists as random coil. While the grafting space decreases due to the increased grafting density, the steric repulsion pulls the tethered chains away from the surface, forming a brush-type layer. Hence, the thickness of the film increases dramatically [30]. The value of λ is 15~50 Å for conventional organic films [28]. Hence, the thickness of grafted PEG from 6 mg/ml initial PEG concentration solution is estimated as 6 to 20 nm, which is close to the reported value [20] for PEG with ~5000 Da molecular weight. The inset figure of Fig. 4.3(b) is the same plot of Fig. 4.3(b) but with linear X axis, which shows the PEG adsorption behavior in different initial PEG concentration solutions. The curve of PEG thickness-initial concentration can be fitted well by the typical Langmuir isotherm model, and the standard deviation (R2) is 0.9387. Other peaks from the XPS spectrum, such as C1s peaks, can also be used for PEG thickness evaluation. These two peaks increase in intensity with increased grafted PEG thickness because of the hydrocarbon chains of the monolayers [19,27]. However, unlike peak Au4f1/2, a plot of C1s intensity against PEG thickness is curved [27]. Hence, the peak Au4f1/2 is used for more precise calculation. .    82                Figure 4.2 High resolution C 1s peak scan of (1) control (bare surface) and surfaces incubated in (2) 2x10-7 mg/ml, (3) 2x10-5 mg/ml, (4) 2x10-3 mg/ml, (5) 1, (6) 6 mg/ml PEG initial concentration solutions. The increase of C-O and decrease of C-C peak intensity is shown with the increase of PEG initial concentration.          290 285 280 C PS  (6) (5) (4) (3) (2) (1) Binding Energy (eV) C-O C-C  83                     Figure 4.3 (a) The log-log plot of integral intensity of peak Au4f1/2 with PEG initial concentration, (b) The linear-log plot of the corresponding PEG grafted thickness and PEG concentration. The inset in Figure 4.3(b) is the linear-linear plot of PEG thickness and PEG concentration, which shows the PEG adsorption behavior and can be fitted by Langmuir isotherm with R2=0.9387. 1E-7 1E-6 1E-5 1E-4 1E-3 0.01 0.1 1 10 0 1 2 3 4   Th ic kn es s/ λ ( nm ) PEG initial conc. (g/l) (a) 1E-7 1E-6 1E-5 1E-4 1E-3 0.01 0.1 1 10 105 106   C PS PEG initial concentration (g/l)  10-    10-      10-5  10-     10   10-2   0.1     1      10 Th ic kn es s/ λ ( nm )    PEG Conc. (g/l) R2=0.9387 (b) 10-7 0-6    10-5 10-4  10-3 1 -2 .  84 The chemical characteristics of grafted PEG were further analyzed and confirmed by ATR- FTIR spectroscopy. Figure 4.4 shows subtractive spectra of PEG grafted substrates from bare gold surfaces. The presence of absorption bands around 1100 cm-1 (C-O-C stretching) and at 2960 and 2869 cm-1 (-CH2 stretching) on PEG grafted surfaces are the characteristic of PEG, which verifies the presence of PEG grafted on gold surfaces [21,29,31]. The increased intensity of absorption bands (both C-O-C and –CH2 stretching) with increased initial PEG concentrations designates the higher graft density of PEG on gold surface in higher PEG initial concentration solutions. Our XPS data given earlier also support this result.             Figure 4.4 The infrared spectra of PEG grafted surfaces (with gold-coated PZT substrate), incubated in (1) 6 mg/ml, (2) 0.1 mg/ml, and (3) 2x10-2 mg/ml PEG initial concentration solutions, measured by ATR-FTIR. The C-O-C and –CH2 peaks are characteristic peaks of PEG. The intensity of these peaks increases with the surface incubated in increased PEG initial concentration. 4000 3500 3000 2500 2000 1500 1000 500 0.00 0.01 0.02 0.03 0.04 0.05 0.06 -CH2 C-O-C  A bs or ba nc e (a .u .) Wavenumber (cm-1) (1) (2) (3)  85 4.3.2 Protein Adsorption Studies The ability of polymer coatings to resist protein adsorption with or without mechanical vibration was investigated. Before the protein adsorption experiment, electrochemical impedance measurements were used to determine if the PEG was desorbing during the vibration. The electrochemical impedance measurements performed on the gold and PEG-grafted gold surfaces are shown in the Appendix C. The resistance between electrolyte and surface is listed in Table C.1, and the Nyquist plot of PZT plates before and after vibration is shown in Figure C.1. There was no significant surface-grafted PEG loss after application of vibration. Figures 4.5 and 4.6 show the relative fluorescence intensities of different substrates incubated in different concentrations (from 2x10-9 to 6 mg/ml) of BSA and IgG solutions. The background images were taken before the substrates were incubated in protein solutions. Following protein incubation, the substrates were washed by dipping in PBS 10 times to remove loosely adsorbed proteins. Then the substrates were connected to a function generator to apply vibration. Following vibration, the substrates were washed again by dipping in PBS 10 times. All the substrates were gently dried in air before the fluorescence images were taken. The relative intensity of each initial PEG concentration was measured for up to 30 images (2-4 samples) in Fig. 4.5 and up to 16 images (2 samples) in Fig. 4.6. Statistical analysis was performed and with significance shown by asterisk in Fig. 4.5 and 4.6. Having the thickness calculated from Equation 4.1, the chain density can be evaluated by the following equation [20,22]:                                                                                                                                                    (4.2) where Σ is the grafted chain density (chains/nm2), h (nm) is the XPS calculated PEG thickness, ρ is the density of PEG (1.08 g/cm3, provided by the supplier), and MW is the molecular weight of 602.3 /h MWρΣ =  86 PEG (5703Da, provided by the supplier). For example, the surface incubated in 6 mg/ml has a thickness of 6-20nm, and the corresponding grafting density is 0.68~2.28 chains/nm2. The chain density of PEG is proportional to the thickness of PEG. More precise experiments on well- defined surfaces such as silicon substrate are needed to determine a more precise PEG thickness for grafting density evaluation, but are beyond the scope of this chapter. However, it is important to estimate the chain density of PEG since it is a key and clear variable in the inhibition of protein adsorption.              Figure 4.5 Relative fluorescence intensities, which linearly depend on BSA adsorption quantity, of surfaces incubated in different PEG initial concentration solutions with or without application of vibration. * indicates significant statistic differences (p<0.05).  R el at iv e Fl uo re sc en ce  In te ns ity  0 20 40 60 80 100 120 140 PEG initial concentration (g/l) * * * * * Bare  2x10-9  2x10-8  2x10-7  2x10-6   2x10-5   2x10-4  2x10-3  2x10-2    0.1     0.5       1        2         6 Without vibration Vibration   87 As shown in Fig. 4.5, BSA adsorption decreases with the increase in initial PEG concentration (since the grafting thickness and density of PEG are increased). When the grafting density is increased, the PEG covers more adsorbed sites for BSA adsorption, and the steric repulsion is increased, such that less BSA is able to penetrate the PEG layer and adsorb on the space between PEG chains. Surfaces incubated in low initial concentration (2x10-9 to 2x10-6 mg/ml) PEG solutions showed some BSA adsorption attenuation. For surfaces incubated in higher PEG initial concentration solutions, the BSA adsorption could be reduced to almost 20%, compared with bare gold substrates (control experiment). However, the attenuation of BSA adsorption leveled off when the surfaces were incubated in an initial PEG concentration solution higher than 2x10-4 mg/ml. It has been suggested that when the PEG grafting density is too high, BSA adsorption increases due to the entrapment of BSA molecules within the PEG chains [32], and the loss of flexibility of PEG chains [32]. Fig. 4.5 shows a slight increase in BSA adsorption when the surfaces were incubated in higher PEG concentration solution (>2x10-3 mg/ml). When vibration was applied for 5 min, the adsorbed BSA was partially removed from the surfaces. The control bare gold surface saw a 40% BSA attenuation after vibration. For surfaces with lower initial PEG concentrations (2x10-9, 2x10-8 and 2x10-7 mg/ml) solutions, vibration works more efficiently to remove adsorbed BSA from the surface then surfaces incubated within higher initial concentrations (higher than 2x10-7 mg/ml). Surfaces with the lowest initial PEG concentration solution (2x10-9 mg/ml) saw a 70% further reduction in the adsorbed BSA following vibration. Possible reasons could be: 1) the increased chain flexibility at lower grafting density [32] because of vibration, which Increases the sweeping volume of the PEG chain, and 2) the weak adhesion between protein and surface. Because of the hydrated water molecular bond around the PEG chain, the protein will be repelled from the sweeping volume of the PEG chain.  88 In addition, vibration induces flow that generates shear stress to help removing the desorbed BSA. However, the effect of vibration on protein attenuation is not significant at surfaces with higher PEG grafting density. This may due to the reduction of chain flexibility, which comes from strong interaction between PEG chains, and the entrapment of BSA molecules within the PEG chains [32]. The improved attenuation of BSA adsorption after application of vibration can be clearly observed on the PEG patterned surface. As shown in Figure 4.6, the brighter area is bare gold surface without grafted PEG while the darker area is gold surface with grafted PEG incubated in 6 mg/ml initial PEG concentration solution for 16 hr. The patterned surface was incubated in 2 mg/ml BSA solution for 2 hr at room temperature. Figure 4.6(a) shows the fluorescence image of the surface before vibration was applied, and Figure 4.6(b) shows the surface image after vibration. The fluorescence intensity ratios of brighter/darker without and with vibration (10 Vpp, 5 mimutes) are 2.85 (Fig. 4.6(a)) and 3.6 (Fig. 4.6(b)), respectively.        Figure 4.6 Fluorescent photos of patterned surfaces (a) before vibration and (b) after vibration. The brighter area is gold surface without grafted PEG and the darker area is the surface with grafted PEG incubated in 6 mg/ml PEG solution for 16 hr. (a) (b) Au, no PEG Au, with PEG  89 Figure 4.7 shows the attenuation of IgG adsorption. Initial PEG concentration did not affect the IgG adsorption attenuation compared to BSA. The protein repulsion property of a surface depends on both the nature of the surface (chemistry, physical structure etc) and properties of the protein [2,33,34]. Thus, it is not surprising to see the difference between the two proteins (BSA and IgG). However, as shown in Fig. 4.7, vibration is an effective way to further desorb IgG from the surfaces; there was 45.2% attenuation of IgG adsorption from the bare surface after vibration. Vibration further attenuated 68.9 and 64.8% of IgG adsorption from PEG grafted surfaces incubated in 2x10-9 and 2x10-8 mg/ml PEG initial concentration solutions, respectively. Vibration desorbs IgG from surfaces incubated in higher PEG initial concentration solutions compared to BSA. This may be because of the different dimensions of the proteins. The dimensions of BSA and IgG are 4x4x14 nm and 4.5x4.5x23.5 nm, respectively [35]. Small proteins can diffuse through the polymer brush by overcoming the kinetic energy barrier. However, large proteins can only approach the surface by compressing the brush. This compressive mechanism favors secondary adsorption at the edge of the polymer brush [32,36]. Hence, the adsorption of smaller proteins to the surface is stronger than that of larger proteins. The PEG-grafted surfaces were shown to resist plasma protein adsorption as well. The plasma proteins were incubated for 1 hr. Plasma consists of mainly albumin (50~60%), immunoglobulins (15~35%), fibrinogen (~3%), and other regulatory proteins [37]. In this work, plasma proteins wre 0.3 mg/ml, corresponding to 0.15~0.18 mg/ml of albumin, 0.045~105 mg/ml of immunoglobulin, and 0.009 mg/ml of fibrinogen. Figure 4.8 shows plasma proteins relative adsorption on PEG grafted surfaces incubated in 2x10-9, 2x10-7, 2x10-5, 1, and 6 mg/ml initial PEG concentration solutions. Unlike BSA and IgG, the plasma proteins adsorbed more (106.5 ± 13.6%) on surfaces incubated in very low PEG initial concentration (2x10-9 mg/ml)  90 solution. But resistance to plasma protein adsorption (50.2~54.5% reduced) was shown on surfaces incubated in other higher PEG initial concentration (2x10-7, 2x10-5, 1, and 6 mg/ml) solutions. The application of vibration further reduced the plasma protein adsorption for bare gold and PEG grafted surfaces, from 13 to 38%. About 71.9 ± 3.3% reduction of plasma protein adsorption from PEG was achieved by vibrating the surface incubated in 6 mg/ml PEG initial concentration solution.              Figure 4.7 Relative fluorescence intensities, which linearly depend on IgG adsorption quantity, of surfaces incubated in different PEG initial concentration solutions with or without application of vibration. * indicates significant statistic differences (p<0.05).   R el at iv e Fl uo re sc en ce  In te ns ity  0 20 40 60 80 100 120 140 PEG initial concentration (g/l) Bare    2x10-9   2x10-8  2x10-7    2x10-6  2x10-5 2x10-3    0.02       0.1       0.5         2 * * * Without vibration Vibration   91            Figure 4.8 Relative fluorescence intensities, which linearly depend on plasma protein adsorption quantity, of surfaces incubated in different PEG initial concentration solutions with or without application of vibration. * indicates significant statistic differences (p<0.05).  4.4 Conclusions We have investigated an anti-fouling treatment by combining surface-grafted PEG and application of vibration. PEG grafted surfaces and vibration application alone were shown to attenuate the adsorption of BSA, IgG, and plasma proteins. The resistance of surfaces to BSA and plasma proteins adsorption depends on the surface grafting density of PEG, while there is no significant dependency on IgG adsorption and surface grafting density of PEG. However, vibration enhances the resistance to adsorption of BSA, IgG, and plasma proteins. In vitro results showed that attenuation of 72 ± 18% of BSA and 81.8 ± 6.2% of IgG adsorption can be achieved by vibrating the surface with very low grafting density of PEG (incubated in 2x10-9 mg/ml initial PEG concentration solution). 71.9 ± 3.3% reduction of plasma proteins adsorption was achieved R el at iv e Fl uo re sc en ce  In te ns ity  0 20 40 60 80 100 120 140 control 2.00E-09 2.00E-07 2.00E-05 1.00E-01 1.00E+00PEG initial concentration (g/l) ** * * * * * Bare           2x10-9         2x10-7               2x10-5                  1                  6 Without vibration Vibration   92 by vibrating the surface incubated in 6 a mg/ml PEG initial concentration solution. Furthermore, the enhanced resistance to BSA adsorption by grafted PEG and application of vibration was observed on the PEG patterned surfaces. We have also investigated the characteristics of grafted PEG by XPS and ATR-FTIR. The grafting of PEG was observed by the change of C1s bind energy. In addition, the grafted PEG thickness can be evaluated from the integral intensity of peak Au4f1/2 in the high-resolution XPS spectrum. The spectrum of ATR-FTIR evidences the existence of grafted PEG on the surfaces.                  93 4.5 Bibliography [1] M. B. Gorbet and M. V. Sefton, Biomaterials, 25 (2004) 5681. [2] B. D. Ratner and S. J. Bryant, Annu. Rev. Biomed. Eng., 6 (2004) 41. [3] S. Sharma, R. W. Johnson, and T. A. Desai, Biosens. and Bioelectron., 20 (2004) 227. [4] A. L. Klibanov, V. P. Torchilin, S Zalipsky, in V. P. Torchilin and V. Weissig (Eds.), Liposomes, Oxford University Press, Oxford, 2003, p. 231-265. [5] K. J. Kitching, V. Pan, and B. D. Ratner, in H. Biederman (Eds.), Plasma Polymer Films, Imperial College Press, London, 2004, p 325-377. [6] N. Nath, J. Hyun, H. Ma, and A. Chilkoti, Surf. Sci., 570 (2004) 98. [7] K. F. Bohringer, J. Micromech. Microeng., 13 (2003) S1. [8] C. Mao, Y. Qiu, H. Sang, H. Mei, A. Zhu, J Shen, and S. Lin, Adv. Colloid Interfac. Sci., 110 (2004) 5. [9] H. Ma, J. Hyun, P Stiller, and A. Chilkoti, Adv. Mater., 16 (2004) 338. [10] Z. Zhang, S. Chen, Y. Chang, and S. Jiang, J. Phys. Chem. B, 110 (2006) 10799. [11] W. K. Cho, B. Kong, and I. S. Choi, Langmuir, 23 (2007) 5678. [12] K. L. Prime and G. M. Whiteside, J. Am. Chem. Soc., 115 (1993) 10714. [13] S. Nagaoka and A. Nakao, Biomaterials, 11 (1990) 119. [14] J. M. Harris, in J. M. Harris (Eds.), Poly(thylenen glycol) Chemistry : Biotechnical and Biomedical Applications, Plenum Press, New York, 1992, Chapter 1. [15] I. Tsujino, J. Ako, Y. Honda, and P. J. Fitzgerald, Expert Opin. Drug Deliv., 4 (2007) 287. [16] H. M. Burt and W. L. Hunter, Adv. Drug Deliver. Rev., 58 (2006) 350. [17] L. D. Unsworth, H. Sheardown, and J. L. Brash, Langmuir, 21 (2005) 1036. [18] C. Pale-Grosdemange, E. S. Simon, K. L. Prime, and G. M. Whitesides, J. Am Chen. Soc., 113 (1991) 12. [19] P. E. Lainbinis, C. D. Bain, and G. M. Whitesides, J. Phys. Chem., 95 (1991) 7017. [20] B. Zdyrko, S. K. Varshney, and I. Luzinov, Langmuir, 20 (2004) 6727. [21] N. A. Alcantar, E. S Aydil, and J. N. Israelachvili, J. Biomed. Mater. Res., 51 (2000) 345. [22] S. J. Sofia, V. Premnath, and E. W. Merrill, Macromolecules, 31 (1998) 5059. [23] G. D. Meyer, J. M. Morán-Mirabal, D. W. Branch, and H.G. Craighead, IEEE Sensors Journal, 6 (2006) 254.  94 [24] K. M. Ainslie, G. Sharma, M. A. Dyer, C. A. Grimes, and M. V. Pishko, Nano Lett., 5 (2005) 1852. [25] P. Y. Yeh, J. N. Kizhakkedathu, J. D. Madden, and M. Chiao, Colloids Surf. B Biointerfaces, 59 (2007) 67. [26] T. T. Huang, J. Sturgis, R. Gomez, T. Geng, R. Bashir, A. K. Bhunia, J. P. Robinson, and M. R. Ladisch, Biotechnol. Bioeng., 81 (2003) 618. [27] C. D. Bain and G. M. Whitesides, J. Phys. Chem., 93 (1989) 1670. [28] K. C. Popat, S. Sharma, and T. A. Desai, J. Phys. Chem. B, 108 (2004) 5185. [29] B. Zhu, T. Eurell, R. Gunawan, and D. Leckband, J. Biomed. Mater. Res., 56 (2001) 406. [30] L. D. Unsworth, Z. Tun, H. Sheardown, and J. L. Brash, J Colloid Interf. Sci, 281 (2005) 112. [31] P. Harder, M. Grunze, R. Dahint, G. M. Whiteside, and P. E. Laibinis, J. Phys. Chem. B , 102 (1998) 426-436. [32] J. Zheng, L. Li, S. Chen, and S. Jiang, Langmuir, 20 (2004) 8931. [33] J. D. Andrade, Surface and Interfacial Aspects of Biomedical Polymers, Plenum Press, New York, 1985, Volume 2. [34] J. L. Brash, Curr. Opin. Colloid & Interf. Sci., 1 (1996) 682. [35] M. Zhang, T. Desai, and M. Ferrari, Biomaterials, 19 (1998) 953. [36] L. Li, S. Chen, J. Zheng, B. D. Ratner, and S. Jiang., J. Phys. Chem. B 109 (2005) 2934. [37] B. Blombäck and L. A. Hanson, Plasma Proteins, Pitman Press, Bath, 1975, p. 17-21.       95 CHAPTER 5 SELF-ASSEMBLED MONOTHIOL-TERMINATED HYPERBRANCHED POLYGLYCEROLS ON GOLD SURFACE:  A COMPARATIVE STUDY ON THE STRUCTURE, MORPHOLOGY AND PROTEIN ADSORPTION CHARACTERISTICS WITH LINEAR PEG 416 5.1 Introduction Protein adsorption at the biomaterial-tissue interface is the first and critical event that initializes a cascade of host responses, including platelet activation, blood coagulation, and complement activation [1]. Many approaches have been used to prevent such non-specific biological interactions [2-5]. Hydrophilic polymer based coatings [6-14] have been used as an anti-biofouling agent for a number of applications including coatings to biomedical devices such as BioMEMs [15], biosensors [16] and drug delivery systems [17]. Such coatings frequently extend the life span of biomedical devices [18,19] and the circulation half life of drug delivery systems [20,21]. Several factors affect the protein-repelling properties of such polymer thin films on the surface include the similarity of interfacial free energies of the polymer with that of water, interaction of proteins with polymers through hydrophobic or charge interactions and environmental factors such as temperature and pH [22-24]. In the case of neutral hydrophilic polymer brushes, the steric barrier due to high conformational entropy of anchored chains is the  41A version of this chapter had been published. Yeh, J. P. Y, Kainthan, R. K., Zou, Y., Chiao, M. and Kizhakkedathu, J. N. (2008) Self-Assembled Mono Thiol Terminated Hyper-branched Polyglycerols on Gold Surface:  A Comparative Study on the Structure, Morphology and Protein Adsorption Characteristics with Linear PEG on Langmuir 24:4907-4916.  96 main contributing factor towards protein repulsion [25-31]. Other factors include, the structure of the polymer on the surface (linear vs. branched) and molecular weight of the grafted chains. One of the commonly used polymers for making such films is PEG because of its properties such as low toxicity, low immunogenicity, and the ability to prevent nonspecific protein adsorption and cell adhesion [32-34]. Covalent coupling and chemisorption are the two common methods used for the attachment of polyethylene glycol onto surfaces and such methods have been studied extensively on various substrates [35,36]. Though this is one of the widely used polymers for biomedical applications, one of the disadvantages is its susceptibility to oxidation and subsequent degradation [37]. Another disadvantage is the availability of only two functionalities per chain because of its linear nature which limits further modification. Also when attached at one end, the apparent surface density of functionalities decreases with increase in molecular weight owing to inherent low graft density of such assemblies on the surface [38]. This is a huge disadvantage when developing functional surfaces for the immobilization of biomolecules. Thus development of surface modification techniques which could not only make surface non-fouling but also functional will be of great importance. Such developments will potentially increase the sensitivity and detection limit of immunoassays and protein arrays due to the decreased non-specific protein interaction and increased coupling capacity [39]. The use of dendritic polymers with multiple surface functional groups are gaining increased attention in recent years for such applications and it has been demonstrated that such surfaces can be modified with broad range of functional groups and biomolecules such as peptides, antibodies and DNA for various biomedical applications [40-42]. We have recently shown that hyperbranched polyglycerols are highly biocompatible [43,44], and can potentially be functionalized to various degrees due to the presence of large amount of  97 reactive hydroxyl groups [45]. A recent report of Haag and coworkers demonstrated that surfaces coated with hyperbranched polyglycerols are resistant to protein adsorption [46]. The thiol functionalized polyglycerols used in their study was synthesized by the post-modification of hydroxyl groups which could potentially generate multiple attachment points on the surfaces due to the presence of more than one thiol group per molecule [46]. In this chapter, we report the synthesis and characterization of mono thiol functionalized hyperbranched polyglycerols and a detailed study on the self assembly of these molecules on gold surface. Being functionalized with a single thiol group within the polymer, these polymers are expected to form only one attachment point on the surface. We studied the properties of the thin films formed by these polymers with respect to its molecular weight, graft density, thickness, and morphology and protein adsorption characteristics from single protein solutions. Further, we compared the results with thin polymer films formed by linear PEG thiols of similar molecular weights under identical conditions.  5.2 Materials and Methods 5.2.1 Materials Bovine serum albumin-fluorescein isothiocyanate (BSA-FITC, A9771) and anti-mouse goat immunoglobulin G-FITC (IgG-FITC, F5265) were purchased from Sigma and used without further purification. Methoxy poly (ethylene glycol) thiol 5000 (mPEG-5000, CH3O- (CH2CH2O)127-CH2CH2CH2CH2-SH, average molecular weight 5703 Da) was purchased from Nektar Therapeutics. Methoxy poly (ethylene glycol) thiol 2000 (mPEG-2000) was purchased from Laysan Bio, Inc (lot # 103-56; average molecular weight 2000 Da). 2, 2’-dihydroxy ethane disulfide was purchased from Aldrich and used as such. Glycidol purchased from Aldrich was  98 purified by distillation. 5,5 ′ -dithiobis (2-nitrobenzoic acid) (DTNB) was obtained from Sigma. Silicon substrate with thickness 400μm was purchased from Helitek Company LTD. Phosphate buffered saline (PBS) (1x, 0.067M PO4 and 0.15M NaCl, PH 7.4) was purchased from HyClone.  5.2.2 Polymer Synthesis a) Disulfide containing hyperbranched polglycerol (HPG-S-S-HPG) Hyperbranched polyglycerols containing disulfide group were synthesized by anionic ring opening multibranching polymerization of glycidol using partially deprotonated 2, 2’-dihydroxy ethane disulfide as initiator as described earlier [47].  Briefly, 0.44 ml of the disulfide initiator was taken in a three neck round bottom flask equipped with a mechanical stirrer. Approximately 10 % of the hydroxyl groups were deprotonated using diphenyl methyl potassium (K+CH-Ph2) (0.25 M solution in THF). The mixture was stirred for 30 min and the reaction flask was immersed in an oil bath maintained at a temperature of 100 o C. Glycidol (10 ml, corresponding to a monomer to initiator ratio of 40) was added drop wise over a period of 12 hr using a syringe pump. The polymerization mixture was stirred for an additional 2 hr. Another polymer was synthesized using a higher glycidol to initiator ratio. The product was dissolved in methanol, neutralized by passing three times through a column containing cation-exchange resin (Amberlite IRC-150). Both HPG disulfide polymers were then precipitated thrice from acetone and dried.  b) Mono thiol functionalized HPG Mono thiol functionalized HPGs (HPG-SH) were synthesized by reduction of the disulfide bond present in HPG-S-S-HPG polymers [48].  The polymer (200 mg) was dissolved in 10 ml  99 ethanol and excess of dithioerythritol (100 mg) was added to this solution. The pH was adjusted to 10 by adding ammonium hydroxide solution. The solution was stirred for 1 hr and the resulting HPG-SH was precipitated thrice from acetone. The molecular weight of the polymers was characterized by MALDI-TOF mass spectrometry analysis (DHB matrix) and the presence of thiol in the polymers was tested by UV-Vis spectroscopy following its reaction with DTNB [49].  5.2.3 Gold Substrate Preparation Silicon substrate was treated with RCA1 clean (NH4OH:H2O2:H2O is 1:1:5) for 15 min to remove any organic contamination, rinsed thoroughly with DI water and substrate was dried under a stream of nitrogen. Then, chromium layer (99.95 % purity, Kurt J. Lesker company, EVMCR35D) of 5 nm thick was deposited as an adhesion layer for Au layer deposition. Au layer (99.999% purity, Kurt J. Lesker company, EVMAUXX50G, X(diameter):1/8”, X(long):1/8”) of 50 nm was following deposited as a functional layer for PEG coating. Chromium and Au were deposited by an evaporator with an electron beam source. The substrate was characterized by AFM topography analysis.  5.2.4 Preparation of Monolayer on Gold Substrates Before polymer coating, the substrates were further cleaned by freshly prepared piranha solution (H2SO4:H2O2, 7:3 (v/v)) to remove any organic contamination. The substrates were rinsed with DI water thoroughly and dried with a nitrogen gun. The mPEG thiol solutions were prepared by dissolving powdered polymer in ethanol at 40 °C. The HPG-SH solutions were prepared by dissolving it in ethanol at room temperature. The gold coated plates were immersed  100 in the mPEG thiol and HPG-SH solution of different concentrations (from 2 × 10-5  to 20 mg/ml) at room temperature for polymer coating. We have also investigated the adsorbed mPEG thickness at different incubation times (9, 16, 24, and 48 hr) in 6 mg/ml mPEG-5000 solution. The polymer coated surfaces were washed thoroughly with 100 % ethanol to remove unbound polymer and dried.  5.2.5 Surface Characterization 5.2.5.1 Ellipsometry The Variable Angle Spectroscopic Ellipsometer (VASE) spectra were collected on M-2000V spectroscopic ellipsometer (J.A. Woolham Co. Inc., Lincoln, NE) at 50°, 55°, 60°  and 65°, at wavelengths from 370 to 1000 nm with M-2000TM  50W quartz tungsten halogen light source. The VASE spectra were then fitted with the multi-layer model (ambient-film-substrate) based on the WVASE32 analysis software, using the optical properties of a generalized Cauchy layer to obtain the ‘dry’ ellipsometric thickness of the adsorbed polymer layer. This ‘dry’ thickness is also known as the dehydrated thickness, measured under ambient conditions. For evaluating thickness of polymer layer on the gold surface, refractive indices of mPEG and HPG were determined. Transparent and relatively thick layers (~50 nm) of polymer (mPEG thiol and HPG thiol) were prepared (to avoid surface inference) on glass slide with a coarse surface on the back side. Then optical constant (n) curve of polymer with multiple wavelengths (λ, 400-1000 nm) was measured at incident angles 65, 70, and 75° at different places on the surface and was repeated for at least two different polymer films and averaged the data. The standard homogeneous multi-layer model was used to interpret the experimental data. This model contains two unknowns: the thickness of polymer layer and refractive indices of the  101 adsorbed polymer. Since the polymer films used in this study were thin and transparent (i.e., κ=0), we used Cauchy layer [50-52],  in which n equals to A+Bλ-2  (A, B are constants), to model the polymer layer. At high λ (700-1000 nm), the optical constant is approximating to A, so we can fix B = 0 and fit experimental data by changing A and thickness of the polymer layer. Once the thickness is decided, we fitted the whole experimental data (370-1000 nm) for calculating A and B. For mPEG-5000, the calculated refractive indices: A is 1.474 and B is 0.0081, which is close to the reported values for crystalline PEG (1.46 [50,51]  to 1.465 [24]  for A and 0.01 [51]  for B). These values are good estimates and have been commonly used by other researchers [50,53].  The calculated refractive indices of HPG ware 1.473 for A and 0.0082 for B and was used for evaluating the polymer film thickness on gold surface. To further characterize the polymer films, several other parameters were evaluated. The polymer surface coverage (adsorbed amount), Γ(mg/m2), was calculated from the ellipsometry thickness of the layer, h (nm) [24,54]: Γ = hρ                                                                                            (5.1) where ρ is density of attached molecules. The density of mPEG-SH (1.08 g/cm3) was provided by the supplier and HPG-SH was assumed to be the same with PEG. The chain density, Σ (chains/nm2), i.e., the inverse of the average area per adsorbed chain, was determined by [24,54]: Σ = ΓNA(10-21)/MW = 602.3Γ/MW                                                       (5.2) where NA is the Avogadro’s number and MW (g/mol) is the molecular weight of the grafted polymer.    102 5.2.5.2 ATR-FTIR Spectroscopy ATR-FTIR absorption spectra were collected on a Nexus 670 FT-IR ESP (Nicolet Instrument Corp., Waltham, MA) with a MCT/A liquid nitrogen cooled detector, KBr beam splitter, and a MkII Auen Gate Single Reflection attenuated total reflectance (ATR) accessory (Specac Inc., Woodstock, GA). The sample stage contained a diamond window and a sapphire anvil on a torque limiting screw set to deliver 80 lbs pressure. IR spectra of the surfaces were recorded from 600–4000 cm-1 at room temperature.  5.2.5.3 X-ray Photoelectron Spectroscopy (XPS) Aanalysis The surface chemical analysis of reacted samples was made with X-ray photoelectron spectroscopy (XPS, Leybold MAX200 X-ray Photoelectron Spectrometer XPS/ESCA Dual anode (Mg Kα and Al Kα) achromatic X-ray source) method using Al Kα X-ray source (1486.6 eV). The samples (0.5 cm x 1 cm) were mounted on the XPS stage. The XPS spectra of all studied elements such as C 1s, O 1s, S 2p, and Au 4f were measured with a constant analyzer with 48 eV pass energy. All binding energies (BEs) were referenced to the Au 4f peak (BE, 84 eV).  5.2.5.4 Atomic Force Microscopy The atomic force microscope used for this experiment was a commercially available multimode system with an atomic head of 100×100 μm2 scan range, using the NanoScope IIIa controller (Digital Instruments, Santa Barbara, CA). AFM was performed in air by contact mode using a commercially manufactured V-shaped silicon nitride (Si3N4) cantilever with gold on the back for laser beam reflection (Veeco, NP-S20). Typical tip radius and spring constant of the  103 cantilever was 5-40 nm and 0.06 N/m, respectively. The raw data was processed with NanoScope IIIa software by first-order flattening. Force measurements were performed using the same AFM equipped with a Nanoscope IIIa controller and a fluid cell. The experiments were performed in force mode under 0.1 M NaCl. All measurements were taken at an approach speed of 1.0 Hz. The raw AFM force data (cantilever deflection vs displacement data) were converted into the reduced force vs separation following the principle of Ducker et al [55]. The onset of the region of constant compliance was used to determine the zero distance, and the region in which force was unchanged was used to determine the zero force. The hydrated thicknesses of the grafted polymer layers in 0.1 M NaCl were determined by two points, zero distance and zero force, from force-distance curves. The average thickness obtained from several force curves at various locations was taken as the hydrated thickness.  5.2.6 Protein Adsorption Studies The substrates were initially equilibrated with phosphate buffered saline (PBS buffer) for 3 hr and then the protein adsorption experiments were conducted by incubating substrates with or without PEG modification in 2 mg/ml BSA and 0.1 mg/ml IgG solution in 48-well polystyrene plates at room temperature for 1 hr. Following the incubation, the substrates were rinsed thoroughly with PBS (>10 times) to remove loosely adsorbed proteins. The images of the protein adsorbed surfaces were taken by a fluorescence microscope (Nikon eclipse TE 2000-U with X- Cite 120 fluorescence illumination system, FITC filter and DS-U1 suit digital camera). The green fluorescence intensity of the images was transferred to grayscale by Adobe PhotoshopTM 6.0, which is linear with the relative quantity of adsorbed proteins on the surface [56]. The  104 calibration scale was set by two points: grayscale intensity of bare surface was set as 0% and protein adsorbed bare surface was set as 100%. The relative adsorption of proteins on various surfaces is normalized based on the calibration scale.  5.3 Results and Discussions 5.3.1 Polymer Synthesis and Characterization Synthetic strategy for the mono thiol functionalized polyglycerols is given in Illustration 5.1. The polymerization of glycidol was initiated from 2, 2’-dihydroxy ethane disulfide. Two polymers were synthesized at different initiator to glycidol ratio to achieve high and low molecular weights. 1H NMR showed the presence of disulfide initiator moieties (peak at 2.7 ppm for –CH2-S-) in the purified polymer (Figure D.1, appendix D). Reduction of disulfide group with dithioerythritol produced a mono thiol functionalized HPG. The formation of thiol is evident from the formation of a yellow colored solution upon reaction with DTNB and absorbance at λ = 420 nm [49]  (Figure D.2, appendix D). The disulfide polymers did not produce any color after mixing with DTNB. The synthesis of the disulfide precursor polymer was very simple, robust and quantitative. The molecular weight of the polymer can easily be adjusted by employing a suitable glycidol to initiator ratio and this method could be attractive alternative to similar dendritic PEG based polymers. Also the presence of several reactive hydroxyl groups in the polymer makes it an ideal candidate for coupling biomolecules. The mono thiol functionalized HPG polymers designated as HPG-SH-L and HPG-SH-H have molecular weights of 1586 Da and 4261 Da (the molecular weight of the highest intensity peak in the MALDI-TOF spectra, Figure D.3) respectively.   105    Illustration 5.1 Synthetic route for mono thiol funtionalized hyperbranched polyglycidols.  106 5.3.2 Surface Modification with Polymers Self assembly of thiol functionalized polymers on metallic surface is a widely studied and versatile method for surface modification [57].  In this study we used four mono thiol functionalized polymers, HPG-SH-L (low molecular weight), HPG-SH-H (high molecular weight), linear mPEG-5000 thiol (high molecular weight) and mPEG-2000 thiol (low molecular weight). Since the linear and hyperbranched polymers have single thiol group within the polymer, we expect to get single gold-sulfur attachment point per polymer chain on the surface (Figure 5. 1). Thus these systems could be a good model to study and compare the adsorption characteristics and properties of linear and branched polymer on the surface. Although one could expect a more ordered monolayer of PEG with a higher carbon spacer (e.g. C11) as reported previously [57], our intention here was to compare the surface adsorption characteristics of HPGs with two carbon spacer and PEGs having similar chemical structure. So in the present case, we used mPEG-5000 with a four carbon spacer (commercially available), mPEG-2000 (two carbon spacer) for our studies. The gold surfaces were incubated with polymer solutions of different concentration. Unlike the thiol functionalized polymers, the non reduced disulphide polymers and a control HPG polymer (without thiol group) did not adsorb significantly to the surfaces (see Table D1, appendix D). The poor adsorption of disulfide containing HPG on gold surface might be due to the inaccessibility of the “deeply buried” disulfide groups to gold surface. The presence of HPG on both sides of the S-S bond may be acting as barrier to stable bond formation between gold and sulfur (see the structure, Illustration 5.1).    107           Figure 5.1 Representation of the structures of polymeric films on gold surface (a) linear mPEG monothiol linear and (b) HPG monothiol.  5.3.3 Surface Characterization: FTIR and XPS To investigate the chemical characteristics of the HPG and mPEG coated surfaces, ATR- FTIR spectroscopy was employed. As shown in the Figure 5.2, the presence of polymer on the surface is evident from the characteristic peaks for HPG and PEG. The C-O-C stretching around 1200 cm-1  and C-H stretching at 2875 cm-1  showed the presence of HPG and PEG on the gold surface. Additional broad peak around 3400 cm-1  in the case of HPG is from the hydroxyl stretching (Figure 5.2(b) and (c), see also Figure D4, appendix D) [58,59].  The polymer grafting on the surface increased with increase in polymer incubation concentration as evident from the increased intensity of IR bands (Figure 5.2(a) and (b)). The intensity of C-O-C peaks was higher and sharp in the case of linear PEG coated surface compared to that of HPGs. Even at low  S n O CH3 O (            ) (a) (b) OH O O O O O OH OH O HO OH O O OH OH O OH OH O OH O OH OH S  108 polymer incubation concentration (0.02 mg/ml), the C-O-C peaks were clearly visible in the ATR-FTIR spectra of polymer coated gold surface for both types of polymers. The presence of HPG on the surface was further demonstrated by measuring the XPS spectra of the gold substrates coated with the polymers (Figure D.5). Figure 5.3 shows the high resolution C1s spectra of HPG-SH-L coated surface. The presence of a peak at 286.4 eV in the C1s spectra is indicative of carbon attached to oxygen which is different from the adventitious carbon [16].  With the increase in HPG-SH-L solution concentration, the C1s peak shifted to higher energy. The atomic composition (ratio of C/O) and thickness of the HPG-SH-L films were calculated from the XPS spectra [54]  and are given in Figure D.6 (appendix D). Results show that the dry thickness measured by ellipsometry (see next section) and those calculated from XPS intensities (Au peak) were comparable. The C/O ratio of the grafted HPG-SH-L film decreased with the increase of polymer incubation concentration.             109                    Figure 5.2 ATR-FTIR spectra of gold surfaces coated with (a) monothiol-functionalized linear mPEG-5000 at (1) 6 mg/ml, (2) 2 x 10-2 mg/ml, (3) 2 x 10-3 mg/ml, and (4) 2 x 10-5 mg/ml polymer incubation concentration and (b) monothiol-functionalized HPG-SH-L at (1) 6 mg/ml, (2) 1 mg/ml, (3) 2 x 10-2 mg/ml, and (4) 2 x 10-3 mg/ml polymer concentration. (c) Comparison of linear mPEG-5000, HPG-SH-L, and HPG-SH-H at 6 mg/ml solution concentration. Polymer films were produced by incubating the gold surface in polymer solution for 16 h. (b) 4000 3500 3000 2500 2000 1500 1000 0.00 0.01 0.02 0.03 0.04 0.05 0.06   A bs or ba nc e Wavenumber (cm-1) (1) (2) (3) (4) C-O-C-CH2-OH 4000 3500 3000 2500 2000 1500 1000 500 0.00 0.02 0.04 0.06 0.08 0.10 0.12 A bs or ba nc e Wavenumber(cm-1)  mPEG  HPG-SH-L  HPG-SH-H C-O-C -CH2 -OH (c) 4000 3500 3000 2500 2000 1500 1000 500 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16  A bs or ba nc e Wavenumbers (cm-1) (1) (2) (3) (4) C-O-C -CH2 (a)   110 276 278 280 282 284 286 288 290 292 294 296 0 20 40 60 80 100  (1)  (2)  (3)  (4)  M od ifi ed  In te ns ity Binding Energy (eV)  Figure 5.3  High resolution C1s scan from XPS spectra of surface grafted HPG-SH-L at (1) 6 mg/ml, (2) 1 mg/ml, (3) 2x10-3 mg/ml polymer incubation concentration, and (4) bare gold- coated substrate.  5.3.4 Polymer Film Thickness and Graft Density The dry thickness of the polymer film on gold surface was calculated from the ellipsometric measurements. Figure 5.4 shows the dependence of adsorption of linear mPEG-5000 thiol on incubation time. The measured polymer film thicknesses of PEG films at 6 mg/ml polymer incubation concentration showed no significant change after 9 hr. So we decided to keep 16 hr for further experiments. The dry thickness of HPGs and linear PEG films obtained with different incubation concentration is given in Figure 5.5(a). Results show that polymer concentration in solution  111 significantly influences the film thickness. The dry thickness increased gradually with polymer incubation concentration in the case of HPGs compared to linear PEG and became almost constant above 6 mg/ml polymer concentration (Fig. 5.5(a). Linear mPEG-5000 gave maximum thickness (~7.3 nm) compared to HPG-SH-L & HPG-SH-H (~3.8 nm) or mPEG-2000 (~ 4.2 nm). In the case of mPEG-5000, there is a jump in the film thickness at 0.02 g/ l concentration. The film thickness reported for very low polymer solution concentrations (on or below 2 × 10-3 mg/ml) may have some uncertainty due to the surface roughness of gold films, which could possibly influence the ellipsometric analysis. Moreover, the thickness values are very small (< 1nm) in those cases. Also, we were not able to detect significant amount of polymer on the surface by FT-IR measurements (Fig. 5.2). The graft density of chains on the surfaces was calculated from the dry thickness measured by ellipsometry, density of the polymer and molecular weight from Equations (5.1) and (5.2). The results are shown in Figure 5.5(b). The graft density of chains increased with increase in polymer incubation concentration for linear and branched polymers. The low molecular weight HPG-SH-L gave maximum graft density (~1.6 chains/nm2) compared to linear mPEG-2000 (~1.38 chains/nm2) or mPEG-5000 (~0.83 chains/nm2) or HPG-SH-H (~0.56 chains/nm2) at 6 mg/ml incubation concentration. The difference in the surface packing of branched and linear chains may be the reason for this behavior. Theoretically, a branched polymer will occupy more area on the surface than the linear polymer of equal molecular weight. A comparison of HPG- SH-L and HPG-SH-H (branched polymers) shows that the low molecular weight HPG packed more closely than high molecular weight HPG suggesting the influence of polymer size on the surface adsorption. Higher the polymer size, lower the graft density and vice versa. Due to the uncertainty in the thickness measurements at very low polymer incubation concentration (below  112 2 × 10-3 mg/ml) as mentioned previously, the graft density values reported should not be considered absolute in those cases. The linear mPEG-5000 produced much higher film thickness compared to HPGs or mPEG- 2000. At high graft densities, linear polymers are supposed to assume brush confirmation and the chains will be extended away from the surface due to steric repulsion [36,38]. Under such conditions, the thickness of the film depends on the graft density of chains. But for branched polymers, the thickness may not increase significantly once they reach critical surface concentration (overlapping) on the surface [60,61]. The chain extension normal to surface is less favorable due to the adjacent branching points compared to linear polymers. Thus the thickness of the branched polymer film on the surface will mainly depend on the size of the polymer at similar graft densities. Since HPGs are not a perfectly branched structure unlike dendrimers, one could expect slight extension of chain normal to surface. In the presence case, the height of the HPG films on the surface is more than the molecular sizes which is an indication of the chain extension on the surface. Similar chain extension is reported previously for perfectly branched dendrimers when adsorbed to the surface [60,62]. The hydrated thickness of polymer film (polymer incubation concentration 6 mg/ml) under aqueous conditions was calculated from the AFM force-distance curves. The hydrated thickness of the film was higher than the dry thickness (Table 5.1, Fig. 5.5(a)) showing the swelling of the polymer film under these conditions. HPG- SH-H gave higher hydrated thickness than HPG-SH-L. The higher thickness of the high molecular weight HPG at low graft density (Table 5.1) compared to low molecular weight HPG is indicative of chain extension on the surface. A detailed analysis of the force curves obtained from HPG and mPEG coated surface will be presented in a future publication.   113  0 10 20 30 40 50 0 2 4 6 8 10 P E G  th ic kn es s (n m ) Incubation time (hr) Figure 5.4 Effect of incubation time on the dry thickness of linear mPEG-5000 film measured by ellipsometry on the gold surface at 6 mg/ml polymer concentration.           114            Figure 5.5 Effect of incubation concentration and type of polymers on the (a) thickness and (b) graft density of polymer films on gold surface. Polymer films were produced by incubating the gold surface in polymer solution for 16 hr.           (a) 10-5 10-4 10-3 10-2 10-1 100 101 0 1 2 3 4 5 6 7 8 9 10    mPEG-2000  mPEG-5000  HPG-SH-L  HPG-SH-H Th ic kn es s (n m ) Bulk Ploymer Concentration (g/l) (b) 10-5 10-4 10-3 10-2 10-1 100 101 0.0 0.5 1.0 1.5 2.0    mPEG-2000  mPEG-5000  HPG-SH-L  HPG-SH-H C ha in  D en si ty  (c ha in s/ nm 2 ) Bulk Polymer Concentration (g/l)  115 Table 5.1 Characteristics of polymer-grafted surfaces obtained from AFM analysis       Polymer Incubation Concentration (mg/ml)       2x10-5 2x10-2 6  Particle height (nm) rms roughness (nm) Particle height (nm) rms roughness (nm) Particle height (nm) rms roughness (nm) Particle height (nm) rms roughness (nm) Hydrated thickness a (nm) Gold 3.87 ± 0.23 1.05 ± 0.12 mPEG 5.05 ± 0.5 1.02 ± 0.15 10.23 ± 1.89 1.66 ± 0.16 12.98 ± 1.42 1.95 ± 0.55 14.79 ± 1.69 HPG-SH-L 5.23 ± 0.29 1.17 ± 0.11 5.58 ± 0.52 1.23 ± 0.10 7.74 ± 0.42 1.69 ± 0.25 7.57 ± 1.14 HPG-SH-H  4.75 ± 0.35 1.05 ± 0.31 5.08 ± 0.13 1.49 ± 0.23 7.92 ± 0.45 1.58 ± 0.12 11.8 ± 1.35  a Measured in 100 mM NaCl at 22 °C.  116 5.3.5 Polymer Film Morphology The topography of the self assembled polymer films obtained from HPGs and linear mPEG- 5000 was measured by the contact mode atomic force microscopy under identical conditions. Such measurements could potentially provide information about the surface ordering of linear and branched polymers. Figure 5.6 shows the 3D image and section plot of polymer adsorbed gold surface at different incubation concentrations. Characteristics of the film such as surface roughness, height of the surface structures are given in Table 5.1. All the three polymers produced uniform surface coverage at concentrations studied as evident from both 2D scan (data not shown) and 3D images. There are finite differences between the polymer coated surface and bare gold surface. The topography of the polymer films obtained from linear mPEG-5000 and hyperbranched polyglycerols was also different. At similar graft densities (0.51~0.6 chain/nm2 at 0.02 mg/ml polymer incubation concentration), the morphology was different for linear mPEG- 5000 and HPG-SH-L. HPG-SH-L produced more uniform structures compared to linear mPEG- 5000 on the surface. Both HPG-SH-L and HPG-SH-H produced uniform structures and showed almost similar topography except that large surface features were produced by HPG-SH-H (Figure 5.6(j)). The surface roughness of the polymer films formed by the HPG (both molecular weights) was consistently lower than that of linear mPEG-5000 at all polymer concentrations and was higher than the bare gold surface (Table 5.1). The surface roughness increased with increase in polymer concentration on the surface for all polymers. More extended structures were formed by the linear mPEG-5000 compared to HPGs. The difference in the flexibility of the chains may be the primary reason for this behavior. A closer examination of the section plots (Fig. 5.6) reveals that the polymers are evenly coated on to the surface except in the case of linear mPEG- 5000 at 6 mg/ml. We also did not observe any crystallization of the polymer on the surface.  117                   Figure 5.6 AFM images showing the surface morphology of bare gold surface (a), linear mPEG- 5000 grafted surfaces (b), low molecular weight HPG-SH-L grafted surfaces (c), and high molecular weight HPG-SH-H grafted surfaces (d) at 6 mg/ml incubation concentration. For a given sample the 3D image (upper) and section plot (lower) are shown.    (d)  (a) 2μm  (b) (c)  118 5.3.6 Protein Adsorption After a thorough characterization of the polymer coated surface, the ability of the polymer coatings to resist the protein adsorption was investigated. Because of the uniformity of surface grafted HPG and linear PEG, and similarities in their repeating units, they may be good models to study the mechanistic aspects of protein adsorption on linear and branched polymer grafted surfaces. However, in the present case there are differences in the nature of the end groups (OCH3 vs OH), therefore, an absolute comparison may be difficult. Evidences from the literature suggest that polymer graft density on the surface is one of the important parameters which determines the protein adsorption on surfaces grafted with linear polymer chains [31.63].  Also there are theoretical predictions and experimental evidences which suggest the enhanced protein repulsion by branched polymers on the surface [54,64,65]. This is believed to be due to the increased surface coverage offered by the branched polymers for a given graft density compared to linear polymers. We used two fluorescently labeled proteins, BSA and IgG in this study. Fluorescent intensity of the polymer coated surface was compared with that of bare gold after incubation with protein solutions. Representative fluorescent micrographs are shown in Figure 5.7(a)~(e) and Figure 5.8(a)~(d). The polymer coating significantly reduced the protein adsorption as seen from the decrease in green fluorescence on the surface. The fluorescence intensity of the protein (BSA) adsorbed surface was in the order of bare gold > HPG-SH-L~mPEG-2000 > mPEG-5000 > HPG-SH-H. Fig. 5.7(f) and 5.8(e) show the relative fluorescent intensity of polymer coated gold surface after the adsorption of BSA and IgG at different graft densities. The data given is the average from 6 photos for one sample done in duplicates. Approximately 75% reduction in the adsorption of BSA on HPG-SH-H modified surface (graft density ~0.57 chains/nm2) compared to ~62 % for  119 linear mPEG-5000 (~0.83 chains/nm2), ~50 % for mPEG-2000 (~1.38 chains/nm2) and ~49% for HPG-SH-L (~1.6 chains/nm2) (Fig. 5.7(f)). In the case of IgG, the percentage reduction in the protein adsorption under identical conditions was 72, 51, and 58 % respectively for HPG-SH-H, HPG-SH-L and linear mPEG-5000 (Fig. 5.8(e)). Although HPGs 21 showed decrease in protein adsorption with increase in graft density in the case of IgG, mPEG-5000 did not show any significant dependence on graft density. The total protein adsorption decreased upon mPEG- 5000 grafting. The relative magnitude of the protein adsorption was different for linear PEG and HPGs. For a given graft density, the high molecular weight HPG-SH-H is more protein resistant than linear mPEG-5000 or low molecular HPG-SH-L. Even at very low polymer incubation concentration (i.e. at low graft densities) there is some reduction in the protein adsorption compared to bare surface. IgG adsorption also followed a similar pattern as BSA (Fig. 5.7(e) and Fig. 5.8(e)) in the case of HPGs but not for mPEG-5000. The difference between HPG-SH-H and linear mPEG- 5000 was smaller in this case compared to BSA adsorption. Our results demonstrated that higher molecular weight HPG has better ability to resist protein adsorption than linear mPEG-5000. Both mPEG-2000 and low molecular weight HPG-SH-L behaved similarly. Our results also proved that the molecular weight of the branched polymer is important in the development of non-fouling surfaces. Although low molecular weight HPG-SH- L grafted surface has higher graft density than the high molecular weight HPG-SH-H, more proteins adsorbed to that surface. This suggests that along with surface coverage, flexibility of the chains is also contributing to the protein repulsion characteristics. Increased chain extension and flexibility of this surface films is evident from the larger hydrated thickness of HPG-SH-H  120 (at 0.57 chains/nm2) compared to HPG-SH-L (~1.6 chains/nm2) under aqueous conditions (Table 5.1).           Figure 5.7 Fluorescence photographs of BSA-adsorbed (a) bare gold and (b) linear mPEG-2000-, (c) linear mPEG-5000-, (d) low molecular weight HPG-SH-L-, and (e) high molecular weight HPG-SH-H-grafted surfaces. Polymer films were produced by incubating the gold surface in polymer solution at 6 mg/ml for 16 h. (f) Effect of the graft density on the BSA adsorption of mPEG-, HPG-SH-L-, and HPG-SH-H-grafted surfaces. (b) (d) (a) (c) (e) 0.0 0.4 0.8 1.2 1.6 2.0 0 30 60 90 120 R el at iv e In te ns ity  Chain Density (chains/nm2) mPEG-5000 HPG-SH-L  HPG-SH-H   mPEG-2000(f)  121             Figure 5.8 Fluorescence photographs of IgG-adsorbed (a) bare gold and (b) linear mPEG-5000-, (c) low molecular weight HPG-SHL-, and (d) high molecular weight HPG-SH-H-grafted surfaces. Polymer films were produced by incubating the gold surface in polymer solution at 6 mg/ml for 16 h. (e) Effect of the graft density on the IgG adsorption of mPEG-, HPG-SH-L-, and HPG-SH-H-grafted surfaces.  (b) (a) (c) (d) (e) 0.0 0.4 0.8 1.2 1.6 2.0 0 30 60 90   Chain Density (chains/nm2) R el at iv e In te ns ity  mPEG-5000   HPG-SH-L   HPG-SH-H  122 5.4 Conclusions We have reported a facile and robust method for the synthesis of mono thiol functionalized hyperbranched polyglycerols. Two different molecular weights of HPG thiols were synthesized and characterized. The surface adsorption characteristics of newly synthesized HPG thiols on gold were studied and compared with linear PEG thiols. Our results show that HPG thiols readily adsorbed to gold surface and gave uniform coating as evident from ellipsometric, ATR-FTIR and AFM topography measurements. Low molecular weight HPG thiol gave higher graft density compared to higher molecular weight HPG thiol or linear PEGs. The graft density of the chains on the surface increased with increase in polymer incubation concentration and became almost constant above 6 mg/ml polymer concentration. The dry film thickness of HPG thiols coated surfaces were lower compared to linear PEGs at all the polymer solution concentrations studied. There are finite differences in the topography of polymer layer formed by mPEG-5000 and HPG thiols. The branched polymers produced more uniform structures on the surface and low surface roughness. High molecular weight HPG thiol coated surface is more resistant to protein adsorption than low molecular weight HPG thiol or linear mPEGs. The protein repulsion of HPG coated surfaces increased with increasing graft density of chains on the surface. Our results show that HPGs with a single thiol group could be a good alternative to PEGs in the development of non-fouling surfaces with additional advantage of further surface functionalization.      123 5.5 Bibliography [1] a) M. B. Gorbet, M. V. Sefton, Biomaterials, 25 (2004) 5681; b)B. D. Ratner and S. J. Bryant, Annu. Rev. Biomed. Eng., 6 (2004) 41. [2] K. J. Kitching, V. Pan, and B. D. Ratner, in H. Biederman (Eds.), Plasma Polymer Films, Imperial College Press, London, 2004, p. 325-377. [3] N. Nath, J. Hyun, H. Ma, and A. Chilkoti, Surf. Sci., 570 (2004) 98. [4] K. F. Bohringer, J. Micromech. Microeng., 13 (2003) S1. [5] C. Mao, Y. Qiu, H. Sang, H. Mei, A. Zhu, J Shen, and S. Lin, Adv. Colloid Interfac. Sci., 110 (2004) 5. [6] H. Ma, J. Hyun, P. Stiller, and A. Chilkoti, Adv. Mater., 16 (2004) 338. [7] Z. Zhang, S. Chen, Y. Chang, and S. Jiang, J. Phys. Chem. B, 110 (2006) 10799. [8] I. Szleifer, Physica A, 244 (1997) 370. [9] W. Feng, S. Zhu, K. Ishihara, and J. L. Brash, Biointerphases, 1 (2006) 50. [10] W. K. Cho, B. Kong, I. S. Choi, Langmuir, 23 (2007) 5678. [11] L. Sofie J. P. Salvage, E. J. Lobb, S. P. Armes, N. C. Billingham, A. L. Lewis, G. W. Hanlon, and A. W. Lloyd, Biomaterials, 25 (2003) 1195. [12] S. Tosatti, S. M. De Paul, A. Askendal, S. V. Vondele, J. A. Hubbell, P. Tengvall, and M. Textor, Biomaterials, 24 (2003) 4949. [13] S. Pasche, S. M. De Paul, J. Voeroes, N. D. Spencer, and M. Textor, Langmuir, 19 (2003) 9216. [14] R. Schlapak, D. Armitage, N. Saucedo-Zeni, M. Hohage, and S. Howorka, Langmuir, 23 (2007) 10244. [15] H. Zhu, M. Snyder, Curr. Opin. Chem. Biol., 7 (2003) 55. [16] S. Sharma, R. W.  Johnson, and T. A. Desai, Biosens. Bioelectron., 20 (2004) 227. [17] A. L. Klibanov, V. P. Torchilin, and S. Zalipsky, in V. P. Torchilin and V. Weissig (Eds.), Liposomes, 2nd ed., Oxford University Press, Oxford, 2003, p. 231-265. [18] I. Tsujino, J. Ako, Y. Honda, and P. J. Fitzgerald, Expert Opin. Drug Deliv., 4 (2007) 287. [19] H. M. Burt, W. L. Hunter, Adv. Drug Deliver. Rev., 58 (2006) 350. [20] D. Kim, S. Park, J. H. Lee, Y. Y. Jeong, and S. Jon, J. Am. Chem. Soc., 129 (2007) 7661. [21] P. Caliceti and F. M. Veronese, Adv. Drug Deliver. Rev., 55 (2003) 1261. [22] E. Lilian, van Vlerken, K. V. Tushar, and M. A. Mansoor, Pharm. Res., 24 (2007) 1405.  124 [23] L. D. Unsworth, H. Sheardown, and J. L. Brash, J. L., Biomaterials, 26 (2005) 5927. [24] B. Zdyrko, S. K. Varshney, and I. Luzinov, Langmuir, 20 (2004) 6727. [25] C. E. McNamee, S. Yamamoto, and K. Higashitani, Biophys. J., 93 (2007) 324. [26] R. Bartucci, M. Pantusa, D. Marsh, and L. Sportelli, BBA-Biomembranes, 1564 (2002) 237. [27] B. M. Steels, J. Koska, and C. A. Haynes, J. Chromatogr. B Biomed. Sci. Appl., 743 (2000) 41. [28] N. Singh, X. Cui, T. Boland, and S. M. Husson, Biomaterials, 28 (2007) 763. [29] M. A. Carignano and I. Szleifer, Colloid Surface B, 18 (2000) 169. [30] J. Satulovsky, M. A. Carignano, and I. P. Szleifer, Natl. Acad. Sci. USA, 97 (2000) 9037. [31] A. Halperin, Langmuir, 15 (1999) 2525. [32] J. M. Harris, in J. M. Harris (Eds.), Poly(ethylene glycol) Chemistry, Biotechnical and Biomedical Applications, Plenum Press, New York, 1992, Chapter 1. [33] L. G. Cima, J. Cell. Biochem., 56 (1994) 155. [34] J. K. Tessmar and A. M. Göpferich, Macromol. Biosci., 7 (2007) 23. [35] P. Kingshott, J. Wei, D. Bagge-Ravn, N. Gadegaard, and L. Gram, Langmuir, 19 (2003) 6912. [36] L. D. Unsworth, Z. Tun, H. Sheardown, and J. L. Brash, J. Colloid Interf. Sci., 281 (2005) 112. [37] S. Sharma, R. W. Johnson, and T. A. Desai, Appl. Surf. Sci., 206 (2003) 218. [38] S. T. Milner, Science, 251 (1991) 905. [39] V. Stadler, M. Beyer, K. Koenig, A. Nesterov, G. Torralba, V. Lindenstruth, M. Hausmann, F. R. Bischoff, and F. Breitling, Proteome Res., 6 (2007) 3197. [40] P. K. Ajikumar, J. K. Ng, Y. C. Tang, J. Y. Lee, G. Stephanopoulos, and H. P. Too, Langmuir, 23 (2007) 5670. [41] S. Pathak, A. K. Singh, J. R. McElhanon, and P. M. Dentinger, Langmuir, 20 (2004) 6075. [42] J. W. Park, Y. Jung, H. Young, J. S. Seo, and Y. B. Lee, Kor. Chem. Soc., 25 (2004) 1667. [43] R. K. Kainthan and D. E. Brooks, Biomaterials, 28 (2007) 4779. [44] R. K. Kainthan, J. Janzen, E. Levin, D. V. Devine, and D. E. Brooks, Biomacromolecules, 7 (2006) 703. [45] R. K. Kainthan, M. Gnanamani, M. Ganguli, T. Ghosh, D. E. Brooks, S. Maiti, and J. N. Kizhakkedathu, Biomaterials, 27 (2006) 5377.  125 [46] C. Siegers, M. Biesalski, and R. Haag, Chem. Eur. J., 10 (2004) 2831. [47] A. Sunder, R. Hanselmann, H. Frey, and R. Muelhaupt, Macromolecules, 32 (1999) 4240. [48] S. Chen, J. Zheng, L. Li, and S. Jiang, J. Am. Chem. Soc., 127 (2005) 14473. [49] a) G. L. Ellman, Arch. Biochem. Biophys., 74 (1958) 443; b) G. L. Ellman, Arch. Biochem. Biophys., 82 (1959) 70. [50] S. Sharma, R. W. Johnson, and T. A. Desai, Langmuir, 20 (2004) 348. [51] F. Xiaowu, L. Lijun, and B. M. Phillip, Biomacromolecules, 7 (2006) 2443. [52] C. Tang, L. Feller, P. Rossbach, B. Keller, J. Vőrős, S. Tosatti, and M. Textor, Surf. Sci., 600 (2006) 1510. [53] K. L. Prime and G. M. Whitesides, Science, 252 (1991) 1164. [54] S. J. Sofia, V. Premnath, and E. M. Merrill, Macromolecules, 31 (1998) 5059. [55] W. A. Ducker, T. J. Sendan, and R. M. Pashley, Langmuir, 8 (1992) 1831. [56] P. Y. Yeh, J. N. Kizhakkedathu, J. D. Madden, and M. Chiao, Colloids Surf. B Biointerfaces, 59 (2007) 67. [57] J. C. Love, L. A. Estroff, J. K. Kriebel, R. G. Nuzzo, and G. M. Whitesides, Chem. Rev., 105 (2005) 1103. [58] N. A. Alcantar, E. S. Aydil, and J. Israelachvili, J. Biomed. Mater. Res., 51 (2000) 345. [59] B. Zhu, T. Eurell, R. Gunawan, D. Leckband, J. Biomed. Mater. Res., 56 (2001) 406. [60] V. V. Tsukruk, F. Rinderspacher, and V. N. Bliznyuk, Langmuir, 13 (1997) 2171. [61] V. Chechik and R. M. Crooks, Langmuir, 15 (1999) 6364. [62] A. Hierlemann, J. K. Campbell, L. A. Baker, R. M. Crooks, and A. J. Ricco, J. Am. Chem. Soc., 120 (1998) 5323. [63] W. Norde and D. Gage, Langmuir, 20 (2004) 4162. [64] a) D. J. Irvine, M. A. Mayes, and L. Griffith-Cima, Macromolecules, 29 (1996) 6037;  b) D. J. Irvine, M. A. Mayes, S. K. Satija, J. G. Barker, S. J. Sofia-Allgor, and L. G. Griffith, J. Biomed. Mater. Res., 40 (1998) 498. [65] J. Groll, Z. Ademovic, T. Ameringer, D. Klee, and M. Moeller, Biomacromolecules, 6 (2005) 956.     126 CHAPTER 6 A SILICON-BASED MICROFLUIDIC ION EXCHANGE PROTEIN CHROMATOGRAPHY DEVICE GRAFTED WITH CARBOXYL FUNCTIONALIZED HYPERBRANCHED POLYGLYCEROLS 517  6.1 Introduction Recent developments in microfluidic systems have provided important platforms for biomedical applications such as liquid chromatography (LC) [1-3], proteome profiling [4-6], cell sorting [7], and biochemical analysis [8]. Microfluidic systems use microfabricated channels that are easy to fabricate and have the advantages of reagent reduction and increased accuracy [9]. Although different LC applications with miniature pillar array columns such as affinity and reverse phase chromatography have been adapted successfully for microfluidic systems [1,2,10- 12], the concept of ion-exchange chromatography with a miniature design has received very little attention. Compared to monolith polymers or filled-in particles, the relatively small effective surface area of the miniature design has been a hindrance to its successful development. The non-specific adsorption of proteins has also caused problems. Poly(dimethylsiloxane) (PDMS) is one of the most popular materials used in microfluidic devices  due to its low cost and easy fabrication. PDMS is highly resistant to chemicals and can be bonded to glass and other polymeric materials using appropriate surface treatments. However, PDMS is highly hydrophobic,  and it is the hydrophobic interactions between PDMS surfaces  51A version of this chapter had been submitted for publication. Yeh, P. Y, Rossi, N. A. A., Kizhakkedathu, J. N. and Chiao, M. A Silicone-based Microfluidic Ion Exchange Protein Chromatography Device Grafted with Carboxyl Functionalized Hyperbranched Polyglycerols.  127 and biological reagents, such as DNA and proteins, that can cause reagent loss and subsequent biofouling [13-16]. When manipulating small quantities of reagents, any loss of the sample and surface biofouling will cause critical errors and malfunction of the device. For biological analysis, it is paramount that the surface fulfills certain criteria, such as wettability, compatibility with bio-fluids, and resistance to nonspecific adsorption. Furthermore, when biological reagents adsorb onto the PDMS surface, the electrical double layer will be perturbed by the charged analytes, leading to heterogeneous zeta potentials present on the PDMS surface. Since  the velocity of electroosmotic flow (EOF) depends on the zeta potential [17], a non-uniform EOF will be generated, causing analytical irreproducibility and poor resolution for biochemical analysis and electrophoresis. Recently, different surface modification approaches have been applied to PDMS microfluidic channels to increase surface hydrophilicity and reduce reagent adsorption to stabilize the EOF. Among the reported techniques, embedding [18,19] and grafting [20,21] of polymer “brushes” onto PDMS surfaces are popular methods for permanent surface modification. In these cases, the polymers must contain hydrophilic chains, similar interfacial free energies compared to water, and large degrees of mobility [22]. In the case of neutral hydrophilic polymer brushes, the steric barrier, caused by high conformational entropy of the anchored chains is the main contributing factor towards attenuation of non-specific adsorption [23-25]. Several factors, including pH, molecular weight, and grafting density of the polymer chain conformation, may affect the final performance of polymer-coated microfluidic channels [26-28]. One of the commonly used anti-fouling polymers is PEG [29-31]. The main disadvantage of PEG is its susceptibility to oxidation and subsequent degradation [32]. Moreover, PEG has limited functionality due to its linear nature. When anchored at one end, the surface density and  128 functionality decreases with increasing molecular weight [33]. In order to reduce non-specific adsorption, issues associated with limited functionality may not be as important as other parameters. However, for the development of functional surfaces for devices such as biosensors and ion exchange chromatography instruments, a decrease in surface functionality is a major disadvantage. Therefore, the development of a dual purpose surface which not only reduces non- specific adsorption, but also imparts chemical functionality, is of great interest to the field of microfluidics [34]. The use of dendritic or hyperbranched polymers with multiple peripheral functional groups has gained significant attention in recent years. It has been demonstrated that such polymers can be modified with a broad range of functional groups and biomolecules, such as peptides, antibodies, and DNA, for various biomedical applications [35,36]. Recently, our group showed that hyperbranched polyglyerols (HPGs) are highly biocompatible, resistant to protein non- specific adsorption, and can be functionalized to various degrees due to the presence of large amount of reactive hydroxyl groups [36-40]. As a result, grafting HPGs to materials surface will likely result in a functional and non-fouling interface. It is also important to note that, unlike dendrimers, HPGs can be easily synthesized in significant quantities and with good control over polymer properties [41]. Here, the use of the ion-exchange concept in a microchannel column for efficient protein separation is reported (Figure 6.1). Surface grafted HPG was used both as a cationic ion exchanger as well as a layer to prevent non-specific protein adsorption onto the PDMS surface. The microchannel columns were fabricated on PDMS with uniform pillar arrays; the channel surfaces were modified using a functional HPG for capture and release of biological reagents, while the non-specific adsorption was significantly reduced.  129               Figure 6.1 Illustration of a PDMS based microfluidic device for the selective capture of relevant proteins.  6.2 Materials and Methods 6.2.1 Materials Bovine serum albumin-fluorescein isothiocyanate (BSA-FITC, A9771) and avidin- Rhodamine conjugate (R01) were purchased from Sigma-Aldrich Corp. and Biomeda Corp. respectively, and used without further purification. Phosphate buffered saline (PBS) has a concentration of 0.067M PO4 and 0.15M NaCl. Inorganic salts were purchased from Fisher Scientific except NaH2PO3.H2O which was procured from EMD Bioscience Inc. RTV 615 polydimethylsiloxane (PDMS) pre-polymer and curing agent were purchased from GE Silicones. Pillar array Grafted polymer contains  COO- at pH 7.4 Positive charged Avidin at pH 7.4 Flow  130 All other reagents for the synthesis and modification were purchased from Aldrich Chemical Corp.  6.2.2 Microfabrication The PDMS prepolymer (RTV 615A) was mixed thoroughly with a curing agent (RTV 615B) at 10:1 (v/v), and degassed with a vacuum pump. PDMS microchannels were fabricated using soft lithography [42]. A photosensitive polymer, SU8 2025 (NanoTM SU8-2025, Microchem) was patterned by a conventional lithography tool (UV 405 nm, Canon mask aligner). The height of SU8 mold on the silicon substrate was 30 μm after spinning at 3000 rpm for 30 s and baked using the following protocol: SB (soft bake): 65ºC for 2 min, and 95 ºC for  5 min; PEB (post exposure bake):  65 ºC for 1 min, and 95 ºC   for 3 min. The lithography mask was designed by L-editTM (Tanner Research, Inc.) and printed on a transparency sheet (Output City CAD/Art Services, Inc.) with a resolution of 20,000 dpi and minimum feature size of 10 μm. The microchannel cover was a glass slide coated with PDMS spun at 1000 rpm for 1 min. The PDMS solution was poured over the SU-8 mold. Both the PDMS microchannel and the cover were cured at 80 ºC for 1 hr. After peeling off from the SU-8 mold, holes with a diameter of 1.2 mm were mechanically punched as inlet and outlet of device on the PDMS microchannel. The PDMS microchannel and cover were bonded by O2 plasma treatment (150W, 300mTorr, 10 s, Trion PECVD) under finger pressure.  6.2.3 Polymer Synthesis and Grafting Protocol Hyperbranched polyglycerol was synthesized from glycidol according to published procedures [41]. Functionalized polyglycerols containing reactive succinimidyl ester groups  131 were synthesized according to methods published previously [43]. Different carboxyl functionalized HPGs were synthesized by changing the molecular weight of HPG and number of carboxyl groups per HPG (Table 6.1).  Table 6.1 The properties of HPGs and number of COOH groups before and after grafting to PDMS surfaces.  Molecular Weight (g/mole) Number of active acid groups Number of remaining acid groups HPG-P1 2500 3 0~1 HPG-P2 8000 5 2~3 HPG-P3 8000 8 5~6   PDMS plates were used initially for developing surface modification protocols and were later adapted for microchannel modification. PDMS surfaces were treated with plasma in a M4L plasma reactor (PVA TePla Corp.) before polymer grafting. The plasma power was delivered at 13.56. PDMS plates and devices were treated using the following protocol: Ar plasma of 75W, 50 sccm, 350 mTorr for 1 min, and allylamine plasma at 55W, 80sccm, 540 mTorr for 10 min. The optimization of allylamine plasma modification was described in the Appendix E. After plasma treatment, the modified surfaces were washed with soap, water, methanol, and Mili-Q water (Milipore, Inc) and were stored in Mili-Q water until polymer grafting. Synthesized polymers were dissolved in Mili-Q water just before contacting the PDMS surfaces. The concentration of each polymer solution used was 1.4 mg/mL. PDMS surfaces were immersed in polymer solution for 1 hr with stirring. For the microchannels, the polymer  132 solutions were injected into microchannels every 10 min for 1 hr. After reaction, the surface was sonicated in Mili-Q water for 10 min and then flushed with water to remove unattached polymer from surface. The surfaces were treated with allylamine plasma (NH2) and grafted with HPG-P1 (2.5 kDa), HPG-P2 (8 kDa), and HPG-P3 (8 kDa, 5 COOH groups per molecule).  6.2.4 Contact Angle Measurements The static contact angles of native and the surface modified PDMS plates were measured by a sessile drop contact angle apparatus (Meulchen Mark I). A 30 μL drop of Mili-Q water was placed on the surface of PDMS plates and then the images were taken. The water contact angle on PDMS plates were measured at least 8 places and the average value is reported.  6.2.5 FTIR Measurements ATR-FTIR absorption spectra were collected using a Nexus 670 FT-IR ESP (Nicolet Instrument Corp., Waltham, MA) with a MCT/A liquid nitrogen-cooled detector, KBr beam splitter, and a MkII Auen Gate Single Reflection attenuated total reflectance (ATR) accessory (Specac Inc., Woodstock, GA). The sample stage contains a diamond window and a sapphire anvil on a torque limiting screw set to deliver 80 lbs. of pressure. IR spectra of the surfaces were recorded from 800–4000 cm−1 at room temperature.  6.2.6 EOF Measurements The PDMS microchannel was filled with PBS buffer for at least 3 h before EOF measurement. The voltage was applied on two ends of the microchannel filled with buffer, and  133 the velocity of flow was measured for four channels, native, NH2, HPG-P2, and HPG-P3 grafted PDMS.  6.2.7 Non-specific Protein Adsorption The PDMS plates were incubated and the microchannels were filled initially with PBS buffer for at least 3 h and then with 2 mg/mL BSA in PBS buffer at room temperature for 1 hr to evaluate the performance of attenuation of non-specific protein adsorption. Following incubation, the surfaces were rinsed thoroughly with PBS to remove loosely adsorbed proteins (PDMS plates were washed 5 times, while microchannels were injected with PBS 3 times). Fluorescent images of the PDMS surfaces were taken by a fluorescence microscope (Nikon eclipse TE 2000-U with X-Cite 120 fluorescence illumination system, FITC filter, DAPI-FITC-TRITC filter, and DS-U1 suit digital camera). The amount of adsorbed BSA is linearly related to the fluorescence intensity [37]. The fluorescence intensity was analyzed by Adobe PhotoshopTM 6.0. The relative fluorescent intensity on the surface was normalized by a calibration scale in which the intensity of native PDMS surface without protein was set to 0% and the intensity of protein exposed native PDMS surface was set to 100%.  6.2.8 Ion-exchange Chromatography and Protein Separation in Microchannels The PDMS microfluidic devices with pillar arrays were incubated with 1 mg/ml BSA, or 1 mg/ml avidin solutions, or avidin-BSA mixtures (wt ratio: 1:1) of 0.2, 0.066, 0.02 mg/ml at room temperature for 3 min to evaluate the capture efficiency. The microchannel surfaces were then washed by injecting PBS 3 times into the channel.  134 The fluorescent images of the protein adsorbed surfaces were taken by a fluorescence microscope as with a DAPI-FITC-TRITC filter only, and the images were taken in RGB mode. The images were then split into three separate color channels. The BSA-FITC was excited using a wavelength of 495 nm, and emission fluorescence was monitored at 520 nm. The avidin- Rhodamine Red was excited at 570 nm while emission was monitored at 590 nm. Hence, the avidin-Rhodamine was seen as red color while BSA-FITC was monitored as green color under the DAPI-FITC-TRITC filter. The adsorbed amount of proteins was linearly related to the fluorescence intensity of the images and was normalized. The performance of the ion-exchange column was evaluated by measuring the capture efficiency (% per 3 min) and selectivity. The capture efficiency was defined as the relative capture percentage per 3 min and the selectivity was defined as relative quantity (fluorescence intensity) of avidin to BSA.  6.2.9 Optical Profiler A Wyko NT1100 Optical Profiling System uses interferometry to measure surface profiles. The SU8 mold was focused under a 20 x objective lens. The silicon substrate was the focus plane and the height and topography was constructed by a back scan of 40 μm, the stage tilt was compensated by software.  6.2.10 Scanning Electron Microscope (SEM) Surfaces were imaged using a Hitachi S-3000N scanning electron microscope with a tungsten electron source. The accelerating voltage can be varied over the range of 0.5-30 kV. The device was glued on a carbon stick and coated with a thin gold by a plasma generator. A conductive  135 carbon paste was connecting the gold film to the bottom of the carbon stick for reducing electon accumulation.  6.3 Results and Discussions Functional surfaces with controlled properties have attracted substantial research interest recently. Potential applications include self-cleaning surfaces [44], devices of intelligent interfaces for biological separation [45] and assays (electrophoresis-based1 and pressure-based chromatography [2,3]), and tissue engineering [46]. In the present work, HPG was used as scaffold for creating functional and non-fouling surfaces in microfluidic channels. HPG was functionalized with carboxylic acid groups by reacting a fraction of the available hydroxyl groups (for example, HPG 8k contains approximately 100 available OH groups) with succinic anhydride [43]. The number of acid groups generated on the surface of the HPG samples was varied by using different molar ratios of OH groups to succinic anhydride. For example, with a molar ratio of 1:5, each HPG should theoretically have an average of 5 acid groups. The properties of the different HPGs, namely molecular weight, number of active succinimidyl ester groups, and number of carboxyl groups, are listed in Table 6.1. A percentage or all of the available carboxy groups were activated using N-hydroxysuccinimide and coupled to amine groups on PDMS surface.  A schematic representation of the modification is shown in Figure 6.2. Initially PDMS plates were used as a model to optimize the grafting protocol prior to the use of similar strategies to modify PDMS microchannels. This is due to the convenience of following the surface modification on a flat surface compared to microchannels with various surface analytical techniques. Argon plasma treatment before allylamine plasma treatment was necessary  136 for efficient amine modification. The amine functionalized PDMS surface and microchannels were later reacted with succinimidyl ester modified HPG to generate a HPG grafted surface. The reaction scheme is shown in Figure 6.2(b). Any remaining unreacted succinimidyl ester groups present on the grafted HPGs will be hydrolyzed to generate carboxyl groups on the surface since succinimidyl ester groups degrade in water to form N-hydroxysuccinimide and carboxylic acids (half-life of succinimidyl ester groups in water is ~16 min) [47]. This process generated a covalently coupled, anionic polymer layer in PBS buffer at pH 7.4 on the microchannel walls, as well as PDMS plates. Water contact angles of the PDMS plates were measured and are shown in Figure 6.3(a). They were 100.4 ± 4.4º, 89.2 ± 1.8º, 80.6 ± 3.5º, 73.2 ± 2.8º, 77.6 ± 1.2º for native, NH2 modified, HPG-P1, HPG-P2, and HPG-P3 grafted PDMS surfaces respectively. After plasma treatment, the surface became less hydrophobic due to the incorporation of more hydrophilic amine groups, as highlighted by the lower water contact angle. Upon HPG grafting, the PDMS surface became relatively hydrophilic due to the formation of a pre-wetting water layer on the HPG caused by hydrogen bonding [48,49]. A slight increase in the water contact angle for HPG- P3 grafted surface compared to HPG-P2 may due to the presence of more amounts of less hydrophilic carboxyl groups on HPG-P3. The surface modification was further investigated by ATR-FTIR spectroscopy. As shown in Figure 6.3(b), verification of HPG grafting on the PDMS plate is evidenced by characteristic C- O-C stretching around 1100 cm-1, C-H stretching at 2875 cm-1, and a small broad peak around 3400 cm-1 due to hydroxyl stretching [37] of HPG. Similar ATR-FTIR profiles were shown when different HPG preparation was used for surface modification (Fig. 6.3(b)). The current grafting method was also compared with other grafting chemistries such as the use of aldehyde  137 chemistry; results are described in the supporting information. The current method produced maximum HPG grafting on the PDMS surface.                 Figure 6.2 (a) Representation of the chemical structure of HPG containing functional succinimidyl ester groups (molecular weights 2.5 kDa or 8 kDa) and (b) the reaction scheme involving the grafting process of functionalized HPG to amine modified PDMS surfaces.     1 h O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS N O O O O O NHS = O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS H2N H2N H2N H2N H2N H2N H2N O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS H2N H2N O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS N S NHS O O O OH O O O HO O O OHHOO OH OH O O O OH HO HO O OH O O OH O O O OH O HO O HO OH O OH O O O OH O O O HO OH O O HO HO O OH HO HO HO O O O OO O O HO OH O OH OH HO O HO HO HO O O OH O O O OH HO O HO HO OHO O OHHO HO HO HO O NHS NHS NHS A B (a) (b)  138                    Figure 6.3 (a) Contact angles of native, amino (NH2), HPG-P1, HPG-P2, and HPG-P3 modified PDMS surfaces (error bars extrapolated from standard deviations from total of 8 measurements), and (b) FTIR spectra of HPG-P1, HPG-P2, and HPG-P3 modified PDMS surfaces. The surfaces were dried before contact and FTIR measurements. 4000 3500 3000 2500 2000 1500 1000 500 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20  HPG-P3  HPG-P2  HPG-P1  A ds or ba nc e Wavelength number (cm-1) (b) (a) Native PDMS NH2 HPG-P1 HPG-P2 HPG-P3 0 5 60 80 100  C on ta ct  A ng le  (O )  139 To verify HPG grafting and the presence of carboxylic acid groups within the PDMS microchannels, electroosmotive flow (EOF) measurements were performed for native, NH2- modified, HPG-P2, and HPG-P3 grafted microchannels. The mobility of EOF (μEOF) is defined as VEOF x E-1, where VEOF and E are velocity and electric field, respectively [50]. E is equal to the applied voltage divided by the distance between the electrodes (E = 4*103 and 16*103 V/m for native, HPG-P2, HPG-P3, and NH2 modified microchannels). Alternatively, μEOF can be defined as εξ/η, where ε and η are the permittivity and viscosity of the solution, respectively. ξ is the zeta potential, which depends linearly on the surface charge density [49]. Since all measurements were performed in PBS buffer, variations in the velocity of EOF in the PDMS microchannels can only be due to the different surface treatments and changes in the zeta potential. A comparison of the EOF mobilities measured for native, NH2, HPG-P2, and HPG-P3 grafted microchannels are shown in Figure 6.4. The μEOF measured for the native PDMS microchannel after O2 plasma treatment is 3.84 ± 0.16*10-4 cm2/V.s (mobility is a vector, which has the same direction as electric field). The value of EOF velocity shows that the PDMS surface is negatively charged; this can be correlated to the formation of silanol groups or carboxyl groups produced during O2 plasma treatment. The μEOF of the NH2 modified PDMS microchannel is (-0.86 ± 0.29)*10-4 cm2/V.s. The NH2 modified surface poses a net positive charge and hence the direction of EOF is reversed compared to the O2 plasma treated PDMS surface. The HPG-P2 grafted microchannel has a mobility of 1.6 ± 0.52*10-4 cm2/V.s. The complete reversal of mobility compared to the amine treated surface is further evidence for the presence of grafted HPG-P2 in the microchannel. For HPG-P3 grafted microchannels, the value of EOF ((4.19 ± 0.07)*10-4 cm2/V.s) increased compared to HPG-P2 grafted channels. This is due to the presence of a higher percentage of peripheral carboxylic acid groups on HPG-P3 (see  140 Table 6.1), which causes a higher zeta potential closer to the surface. The higher water contact angle of HPG-P3 grafted surface given earlier also supports this observation.    Native PDMS NH HPG-P2 HPG-P3 -1 0 1 2 3 4 5   μE O F( 10 -4 cm 2 / V. s)  Figure 6.4 Values of measured μEOF of native, amino (NH2), HPG-P2, and HPG-P3 modified PDMS microchannels. The applied electric fields for native, HPG-P2, and HPG-P3 modified microchannels are 4*103 V/m. The applied electric field for amino modified microchannels is 16*103 V. The average velocity and error bars are from 3 independent measurements.      2  141 Non-specific protein adsorption onto both the PDMS plates and microchannels was investigated using FITC labeled BSA. The relative fluorescence intensities of FITC-BSA incubated PDMS plates and microchannels are shown in Figure 6.5(a); the values for NH2 modified, HPG-P1, HPG-P2, and HPG-P3 grafted PDMS surfaces compared to the native PDMS surface are 79.24 ± 4.76%, 36.10 ± 5.68%, 27.38 ± 2.35%, and 30.58 ± 2.88%, respectively. The higher hydrophilicity of NH2 modified surface may have caused a reduction in BSA adsorption to this surface. Furthermore, surfaces grafted with HPGs strongly attenuated the BSA adsorption. However, there were no significant differences between the relative fluorescent intensities on surfaces grafted with different HPG preparations. In the case of the microchannels, the relative fluorescent intensities in the channels were 66.54 ± 9.41%, 24.01 ± 6.55%, 24.77 ± 8.64%, and 17.25 ± 12.4% respectively for NH2 modified, HPG-P1, HPG-P2, and HPG-P3 (Figure 6.5(b)). Although HPG-P3 has more carboxyl groups present, it did not show any significant change in non-specific protein adsorption suggesting that a functional and non-fouling surface is produced on PDMS plates and in the microchannels. The only difference between HPG-P2 and HPG-P3 modification is the presence of additional carboxyl groups on HPG.          142                   Figure 6.5 (a) Fluorescence intensities of native, amino (NH2), HPG-P1, HPG-P2, and HPG-P3 modified PDMS surfaces and microchannels after FITC labeled BSA adsorption (the error bars in the figure are the standard deviation of 6 measured data points). The BSA solution was incubated with PDMS surfaces for 1 hr at room temperature. After adsorption, PDMS surfaces and microchannels were washed with PBS buffer 5 and 3 times, respectively. (b) Fluorescence photos of the native, amino (NH2), HPG-P1, HPG-P2, and HPG-P3 modified PDMS microchannels shown in Fig. 6.5 (a). The white dash lines are the boundaries of the microchannels. (a) NH2 PDMS HPG-P3 HPG-P1 HPG-P2 (b) Native PDMS NH HPG-P1 HPG-P2 HPG-P3 0 20 40 60 80 100 120  R el at iv e Fl uo re sc en ce  Surface  Microchannel 2  143 Having grafted a non-fouling and functional polymer layer to the microchannels, the potential application of this coating as an ion exchange liquid chromatographic material is tested. The concept (Fig. 6.1) involves the use of a PDMS ion exchange column grafted with HPG to selectively capture a specific reagent from a mixture. The inlet and outlet are connected to a syringe to generate a fluid flow.  Polymer coatings with negatively charged stationary phases are used to capture positively charged proteins, and vice versa. The microfluidic device is comprised of 300 μm wide channels, connecting an inlet and outlet to a separating column (1350 μm wide and 5000 μm long).  The topography of the SU8 mold was profiled by an optical profiler (Figure 6.6(a)). The height of the SU8 pillar mold is around 30 μm. The SEM images of the PDMS replica peeled off from the SU8 mold are shown in Figures 6.6(b) and 6.6(c). Although the heights of the PDMS channel walls and pillars should be around 30 μm, the cross section of PDMS pillar as measured by SEM after curing is approximately 15 μm x 15 μm due to lithography error and shrinking of the  (the cross section in mask design is 20 μm x 20 μm). The surface area to volume ratio of this column design with pillar arrays was 0.11 (1/μm) compared to 0.066 for those without pillar arrays, corresponding to an increase of ~67 %. A further increase in this ratio can be made possible by using molds with higher stationary phase densities and lower aspect ratios (i.e. by lowering the height of the channel). The effective volume and surface area of the current column are 1.7415*10-4 cm3 (5000 μm* 1350 μm*30 μm*((402- 152)/402)) and 1.9514*10-1 cm2, respectively.      144         Figure 6.6 (a) The topography SU8 mold on a silicon substrate measured by an optical profiler (part of the ion exchange column, scale is shown in the figure). The SEM photos of (b) x 400, and (c) x 1000 PDMS pillar arrays. The dimensions of the cross section of the pillar arrays are 15 μm x 15 μm.   Two proteins, BSA and avidin, are used as model proteins for the separation experiments since they are negatively and positively charged respectively in PBS, pH 7.4 (isoelectric points (IP) of BSA and avidin are 4.7-5.0 [51] and 10.5 [45]). The HPG-P3 grafted channels were filled with a solution of proteins at a concentration of 1 mg/mL and washed after 3 min. Fluorescence images of channels filled with 1 mg/mL Avidin and BSA at the onset are shown in Figure 6.7(a) and (c), and were set as 100 %. Fluorescence images of channels incubated with 1 mg/mL avidin and BSA for 3 min and then washed 3 times with PBS buffer are shown in Figure 6.7(b) and (d), and represent proteins on the surface remained due to their affinity with the grafted HPG. The relative fluorescent intensity of the channel due to avidin adsorption is 21.17 ± 2.78% (Fig. 6.7(b)) compared to 8.84 ± 1.23 % for BSA (Fig. 6.7(a)). This suggests that there is a preference for the absorption of positively charged avidin compared to negatively charged BSA to this channel under the conditions used. The relatively low separation efficiency might be due to the (b) (c) (a)  15 0 -15 μm  145 saturation of binding sites caused by a high concentration of proteins. Hence, lower concentrations (0.02, 0.066, 0.2 mg/mL, wt: 1:1) of the protein mixtures were employed to further evaluate the column selectivity. Fluorescence images of the column at the start of the experiment at 0.2 mg/mL total protein concentration (0.1 mg/mL of avidin and BSA each) and the end of the experiment (3 min incubation and washing) are shown in Figure 6.8(a) and (b), respectively. The fluorescence intensity ratio of avidin/BSA at the onset of filling of the protein mixture, followed by a 3 min incubation period and washing 3 times with PBS is shown in Figure 6.8(c). There is a marked difference in the color (red to green) of the column after protein mixture incubation and washing, indicating that more avidin was adsorbed from the protein mixture compared to BSA. The relative fluorescent intensity of avidin to BSA changed dramatically; from 0.96 to 2.36. Hence, the selectivity based on the percentages of relative amount of protein (based on the fluorescent intensity) is 2.46 (2.36/0.96) and it translates to 71.1 % of avidin and 28.9 % of BSA remaining in the column. The high selectivity of avidin indicates that the HPG layer considerably prevented non-specific adsorption, while enhancing specific electrostatic interaction. The capture efficiency of the column upon filling with a 0.2 mg/mL (wt: 1:1) protein solution is 47.7 % after 3 min. Previous work [45] using similar concentrations of a protein mixture (avidin and streptavidin wt: 1:1) reported the capture efficiency at around 86 % after 30 min. A higher capture efficiency is achieved here, possibly due to an increase in the surface to volume ratio (0.11 as opposed to 0.04 reported previously [45]) caused by pillar arrays in the column. The captured avidin was also released using a higher ionic strength buffer solution (10 times Na+ ion strength). After 3 times 10 times concentrated PBS washing, ~ 45.5 % of captured avidin is removed from the column.   146   Figure 6.7 Fluorescence images of columns (a) filled with avidin, (b) incubated with avidin for 3 min and washed 3 times  with PBS buffer (the fluorescent intensity is 21.17 ± 2.78 % compared to (a)), (c) filled with BSA, (d) incubated with BSA for 3 min and washed 3 times with PBS buffer (the fluorescent intensity is 8.84 ± 1.23 % compared to (c)). The concentration of BSA and Avidin was 1 mg/ml, and the measurement was performed at room temperature.         Avidin BSA (a) (c) (b) (d)  147               Figure 6.8 Fluorescence images of columns at (a) 0 min after incubation with protein mixture (0.1 mg/ml avidin and 0.1 mg/ml BSA), (b) After 3 min incubation and 3 times wash with PBS buffer and (c) fluorescence intensity ratio of avidin/BSA at the onset and after 3 min of incubation (followed by washing 3 times with PBS buffer) with the protein mixture.   The column performance upon incubation with lower concentrations of the protein mixtures (0.066 and 0.02 mg/mL) are also investigated. The selectivity of the columns and the avidin capture efficiency upon incubation for 3 min with various concentrations of avidin-BSA mixtures are shown in Figure 6.9. With decreasing protein concentration, the capture efficiency of avidin increased up to 95.4 ± 4.7%. However, selectivity decreased with decreasing protein concentration (Fig. 6.9). Furthermore, the percentage of BSA adsorption also increased as protein concentration decreased. (b) (a) 0 minute 3 minutes 0.0 0.5 1.0 1.5 2.0 2.5   Fl uo re sc en ce  o f A vi di n / F lu or es ce nc e of  B SA  (c)  148 In order to confirm the selective capturing property of the HPG-P3 layer, a control experiment was performed using a HPG-P2 grafted column. The same experiment as described above was repeated using a 0.02 mg/mL avidin-BSA mixture. It was found that the percentages of avidin and BSA present in the column after PBS washing were 60.8 % and 55.3 % respectively with a selectivity of 1.1. In contrast to the HPG-P3 grafted layer, HPG-P2 did not have a preference for any of the proteins. Under similar conditions, the HPG-P3 grafted channel captured ~95.4 ± 5.3 % of avidin. The residual protein adsorption observed here may be due to the non-specific adsorption of protein, which was minimized but not completely eliminated. Based on our previous experiments [37], the grafting density of HPG-P3 inside the microchannels was increased to investigate column performance by increasing the HPG incubation concentration (3 mg/mL and 1.4 mg/mL).  The avidin capture efficiency and selectivity of the column upon incubation with 0.066 and 0.020 mg/mL avidin-BSA mixtures were 68.1 % & 1.74 and 93.5 % & 1.74, respectively. Increasing the polymer graft density did not cause any observed column performance improvement. However, an increase in polymer density did cause a decrease in non-specific BSA adsorption. It is important to note that the functional characteristics of the polymer coatings can be optimized to achieve maximum selectivity and minimum non-specific adsorption. In order to study the effects of the surface area to volume ratio inside the column on the performance of the ion-exchange column, microfluidic channels with and without pillar structures grafted with HPG-P3 in 3 mg/mL polymer solutions were investigated. The in-situ selectivities of the columns at two protein concentrations and various incubation times (30 to 90 s) were also studied (Figure 6.10). Selectivity varied with time; the avidin selectivity increased in the case of the pillar array column compared to the column without pillar arrays. The increased  149 selectivity of the pillar array column is due to the increased surface area to volume ratio (see Fig. 6.6), which was ~1.67 times more than those without pillar arrays. Hence, further increases in the surface area to volume ratio are also essential if an efficient and high-loading chromatography column is to be developed in the future.  0.00 0.05 0.10 0.15 0.20 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 A vidin C apture E fficiency (relative %  per 3 m inutes) S el ec tiv ity Protein Mixture Concentration (mg/ml) 30 40 50 60 70 80 90 100   Figure 6.9 Selectivity (left) of columns and capture efficiency (right) of avidin with various concentrations (mg/ml) of avidin and BSA mixtures (wt ratio 1:1). The selectivity was defined in terms of the relative quantities of avidin to BSA based on the fluorescent intensity.  The capture efficiency of avidin is relative total capture of proteins after 3 min.  150 0 20 40 60 80 100 120 1.0 1.2 1.4 1.6 1.8 2.0 2.2 Time (sec) Se le ct iv ity  Conc. 0.02 mg/ml, With pillar arrays  Conc. 0.02 mg/ml, Without pillar arrays  Conc. 0.066 mg/ml, With pillar arrays  Conc. 0.066 mg/ml, Without pillar arrays   Figure 6.10 In-situ selectivity of columns vs. time at two protein concentrations (0.02 and 0.066 mg/ml) of avidin and BSA mixtures (wt ratio 1:1) with or without pillar arrays. Incubations of columns with pillar arrays with similar concentrations of the avidin and BSA mixture showed higher selectivity.          151 6.4 Conclusions A novel non-fouling, functional surface coating consisting of HPGs was developed on a PDMS surface as well in a PDMS based microchannel. The carboxyl groups on the functionalized HPGs were used as a stationary support for ion-exchange chromatography for the selective capture of positively charged proteins in PBS buffer at pH 7.4.  The grafting process of HPGs onto the surface was followed using contact angle, ATR-FTIR, and EOF measurements. The new microfluidic ion-exchange chromatography device was tested to determine its ability to separate model proteins from a mixture. The capture of positively charged avidin was observed at various avidin-BSA mixture concentrations. In addition, it was shown that the selectivity of protein capture was enhanced with increasing number of pillar arrays in the column due to the increased surface to volume ratio. To enhance the efficiency of the protein separation device, several essential parameters were identified and will need to be improved upon in future – these include increasing the surface to volume ratio of the PDMS column, optimizing the HPG graft density, grafting HPGs with higher numbers of carboxylic acid groups, and increasing the size (chain length) of the HPGs.           152 6.5 Bibliography [1] B. He, N. Tait and F. Regnier, Anal. Chem., 70 (1998) 3790. [2] W. De Malsche, H. Eghbali, D. Clicq, J. Vangelooven, H. Gardeniers and G. Desmet, Anal. Chem., 79 (2007) 5915. [3] I. M. Lazar, P. Trisiripisal and H. A. Sarvaiya, Anal. Chem., 78 (2006) 5513. [4] V. Thongboonkerd, N. Songtawee and S. Sritippayawan, J. of Proteome Res., 6 (2007) 2011. [5] Y. C. Wang, M. H. Choi and J. Han, Anal. Chem., 76 (2004) 4426. [6] J. D. Ramsey, S. C. Jacobson, C. T. Culbertson and J. M. Ramsey, Anal. Chem., 75 (2003) 3758. [7] B. Yao, G. Luo, X. Feng, W. Wang, L. Chena and Y. Wang, Lab Chip, 4 (2004) 603. [8] M. A. Burns, B. N. Johnson, S. N. Brahmasandra, K. Handique, J. R. Webster, M. Krishnan, T. S. Sammarco, P. M. Man, D. Jones, D. Heldsinger, C. H. Mastrangelo and D. T. Burke, Science, 282 (1998) 484. [9] P. Kim, H. E. Jeong, A. Khademhosseinib and K. Y. Suh, Lab Chip, 6 (2006) 1432. [10] C. M. Harris, Anal. Chem., 75 (2003) 64A. [11] N. Malmstadt, A. S. Hoffman, P. S. Stayton, Lab Chip, 4 (2004) 412. [12] N. Malmstadt, P. Yager and A. S. Hoffman, Anal. Chem., 75 (2003) 2943. [13] Y. Xiao, X. D. Yu, J. J. Xu and H. Y. Chen, Electrophoresis, 28 (2007) 3302. [14] C. S. Effenhauser, M. G. J. Bruin, A. Paulus and M. Ehrat, Anal. Chem., 69 (1997) 3451. [15] A. Muck and A. Svatos, Talanta, 74 (2007) 333. [16] H. Bi, S. Meng, Y. Li, K. Guo, Y. Chen, J. Kong, P. Yang, W. Zhong and B. Liu, Lab Chip, 6 (2006) 769. [17] J. Liu and M. L. Lee, Electrophoresis, 27 (2006) 3533. [18] B. Huang, H. Wu, S. Kim, B. K. Kobilka and R. N. Zare, Lab Chip, 6 (2006) 369. [19] K. Yu and Y. A. Han, Soft Matter., 2 (2006) 705. [20] D. Wu, B. Zhao, Z. Dai, J. Qin and B. Lin, Lab Chip, 6 (2006) 942. [21] D. J. Guo, H. M. Han, J. Wang, S. J. Xiao and Z. D. Dai, Colloid Surface A, 308 (2007) 129. [22] W. Norde, J. Buijs and J. Lyklema, in J. Lyklema (Eds.), Fundamentals of Interface and Colloid Science Vol. V, Academic Press, London, 2005, Chapter 3. [23] C. E. McNamee, S. Yamamoto and K. Higashitani, Biophys. J., 93 (2007) 324.  153 [24] N. Singh, X. Cui, T. Boland and S. M. Husson, Biomaterials, 28 (2007) 763. [25] J. Satulovsky, M. A. Carignano and I. P. Szleifer, Natl. Acad. Sci. USA, 97 (2000) 9037. [26] A. J. Wang, J. J. Xu and H. Y. Chen, J. Chromatogr. A, 1147 (2007) 120. [27] S. Hu, X. Ren, M. Bachman, C. E. Sims, G. P. Li and N. Allbritton, Anal. Chem., 74 (2002) 4117. [28] W. Thormanna, J. Caslavska and R. A. Mosher, J. Chromatogr. A, 1155 (2007) 154. [29] J. M. Harris, in J. M. Harris (Eds.), Poly(ethylene glycol) Chemistry, Biotechnical and Biomedical Applications, Plenum Press, New York, 1992, Chapter 1. [30] L. G. Cima, J. Cell. Biochem., 56 (1994) 155. [31] J. K. Tessmar and A. M. Göpferich, Macromol. Biosci., 7 (2007) 23. [32] S. Sharma, R. W. Johnson and T. A. Desai, Appl. Surf. Sci., 206 (2003) 218. [33] S. T. Milner, Science, 251 (1991) 905. [34] V. Stadler, M. Beyer, K. Koenig, A. Nesterov, G. Torralba, V. Lindenstruth, M. Hausmann, F. R. Bischoff and F. Breitling, J. Proteome Res., 6 (2007) 3197. [35] P. K. Ajikumar, J. K. Ng, Y. C. Tang, J. Y. Lee, G. Stephanopoulos and H. P. Too, Langmuir, 23 (2007) 5670. [36] S. Pathak, A. K. Singh, J. R. McElhanon, P. M. Dentinger, Langmuir, 20 (2004) 6075. [37] P. Y. J. Yeh, R. K. Kainthan, Y. Zou, M. Chiao and J. N. Kizhakkedathu, Langmuir, 24 (2008) 4907. [38] R. K. Kainthan, M. Gnanamani, M. Ganguli, T. Ghosh, D. E. Brooks, S. Maiti and J. N. Kizhakkedathu, Biomaterials, 27 (2006) 5377. [39] R. K. Kainthan, J. Janzen, J. N. Kizhakkedathu, D. V. Devine and D. E. Brooks, Biomaterials, 29 (2008) 1693. [40] J. G. Zhang, O. B. Krajden, R. K. Kainthan, J. N. Kizhakkedathu, I. Constantinescu, D. E. Brooks and M. I. C. Gyongyossy-Issa, Bioconjugate Chem., 19 (2008) 1241. [41] R. K. Kainthan, E. B. Muliawan, S. G. Hatzikiriakos and D. E. Brooks, Macromolecules, 39 (2006) 7708. [42] Y. Xia and G. M. Whitesides, Annu. Rev. Mater. Sci., 28 (1998) 153–184. [43] N. A. A. Rossi, R. K. Kainthan, M. D. Scott, D. E. Brooks and J. N. Kizhakkedathu, Polym. Prepr., 49 (2008) 475.  154 [44] R. Wang, K. Hashimoto, A. Fujishima, M. Chikuni, E. K. Kojima, A. Kitamura, M. Shimohigoshi and T. Watanabe, Nature, 388 (1997) 431. [45] L. Mu, Y. Liu, S. Cai and J. Kong, Chem. Eur. J., 13 (2007) 5113. [46] T. Shimizu, M. Yamoto, A. Kikuchi and T. Okano, Tissue Eng., 7 (2001) 141. [47] M. J. Roberts, M. D. Bentley and J. M. Harris, Adv. Drug Deliv. Rev., 54 (2002) 459. [48] K. L. Prime and G. M. Whitesides, J. Am. Chem. Soc., 115 (1993) 10714. [49] D. Xiao, H. Zhang and M. Wirth, Langmuir, 18 (2002) 9971. [50] T. M. Squires and S. R. Quake, Rev. Mod. Phys., 77 (2005) 977. [51] C. L. de Vasconcelos, P. M. Bezerril, T. N. C. Dantas, M. R. Pereira and J. L. C. Fonseca, Langmuir, 23 (2007) 7687.    155 CHAPTER 7 CONCLUSION 7.1 Summary The reduction of biofouling is critical to the success of implantable medical devices. To reach this goal, several key methods have been used. The physical method relies on surface charge and/or vibration, initiated by a piezoelectric material, while the chemical method uses a hydrophilic polymer. In this thesis, both methods are evaluated based on their ability to attenuate protein adsorption (i.e., for BSA, IgG, and/or plasma proteins). As described in Chapter 2, we investigated the mechanism for attenuating protein adsorption for the first time. We described a concept for attenuating the adsorption of BSA and IgG by using surface charges. The initiating electrostatic interaction on the surface repelled the proteins from the surface. The magnitude of the electrostatic interaction is important for preventing the attractive interactions, for example, VDW and hydrophobic interactions (which persist in most cases). The theoretical calculations for the surface interactions matched with the experimental results, so that the roles of the interactions in protein adsorption were verified. In addition, protein adsorption can be mediated by flow shear stress caused by vibration, though the application of this principle faces technical challenges. First, proteins are not simple particles with a regular shape allowing them to be entered into an equation. Proteins can change their conformations in different environments, based on the ion strength, pH, of buffer characteristics. Hence, the equations that are used in Chapter 2 are not likely suitable for all proteins in every environment. BSA, for example, is a globular protein that can be modeled as a particle or a sphere. In addition, some proteins may change their conformation after being adsorbed, and the adsorption is usually irreversible. The outward-facing  156 peripheral areas of some proteins may be composed of different pitches, with charged groups, hydrophobic groups, etc. Although the hydrophobic interaction is the main factor in protein adsorption on the surface, electrostatic interactions may also play a role. The totality of these effects may explain why a strong attenuation of protein adsorption on the surface is not achieved. Nevertheless, the method in Chapter 2 illustrates the main concept to direct research for preventing surface deterioration due to protein adsorption. In Chapter 3, the concept from Chapter 2 is applied to the MEMS-based membrane. With a continuous supply of surface charges and flow near the surface, a corresponding amount of protein adsorption attenuation is achieved. This work is especially important for implantable devices, for two reasons: 1) the devices are of a similar size to the MEMS-based membrane, 2) attenuation of plasma proteins is also observed. The actual environment for most implantable devices will contain plasma proteins. In addition, because of the attenuation of the more complicated plasma proteins, the vibration-assisted antifouling mechanism becomes a more feasible system for implantable devices. The simulation experiments of the vibration of membrane were verified, and may serve as a stepping-stone for designing other effective vibration methods. These affect the flow pattern near the surface. In general, the attenuation of protein adsorption works to a degree, but more enhanced attenuation would be needed for clinical applications where powerful antifouling is required. In Chapter 4, we investigate a combination of method used in Chapters 2 & 3 and coating of hydrophilic polymer coating. The antifouling of piezoelectric surfaces, grafted by polymers with different grafting densities, was evaluated with and without initiating vibration. A much greater attenuation (~70% further attenuation compared to polymer coating alone) was found with vibration combined with lower polymer grafting density or with high polymer grafting density  157 alone. For surfaces with high polymer grafting densities, it is expected that the attenuation depends only on polymer graft densities. Because, at high grafting density (where the distance of polymer molecules is less than half of the Flory distance), the polymer is stretched out to several nm (for polymers having a MW of around 5000 Da), the surface charges induce electrostatic interaction decays which is incomparable with the steric interaction of the polymer. Nevertheless, the attenuation at low surface grafting density in combination with vibration gave encouraging results. Since a much lower polymer concentration can achieve a comparable antifouling performance when combined with vibration, the additional method can play an important role in antifouling, especially when the surface grafting density of the polymer is low (i.e., when the distance between polymer molecules is less than the size of the protein). This observation is important since some situations require the polymer grafting density to be lower so that polymers on the surface can bend and release. If the grafting density is too high, then the surface loses its flexibility. In addition, the use of polymers to attenuate protein adsorption may be applicable for biosensors. In some biosensor designs, the hydrophilic polymer is coated onto non-detect areas to reduce regent loss, while another polymer with modified end groups is grafted to a specific biomolecule. Polymer coating with vibration can reduce the adsorption of non-specific proteins, thus increasing the sensor sensitivity. In Chapter 5, we investigate the attenuation of protein adsorption by two polymer structures, linear (as discussed in Chapter 4) and branched types. The aim is to study the antifouling performance of branched polymers and compare it with linear polymers. The intention is to develop a branched polymer with similar or better antifouling performance than the linear polymers, and modifying the branched polymer so that in the future, a more functional surface will be created (i.e., for chromatography applications mentioned in Chapter 6). The branched  158 polymers have some advantages over the linear variety, in their insusceptibility to oxidation and subsequent degradation, more functionalities, and a linear relationship between surface grafting density and polymer molecular weight. The AFM used in this chapter is not only for the surface morphologies and polymer thickness, but also for investigating the detailed properties of polymers, to investigate the interaction between polymers and different surfaces (charged, specific agents, etc.), and to quantify surface interactions and polymer peripheral properties. In Chapter 6, we applied the branched polymer to PDMS-based microfluidic devices, i.e., ion exchange liquid chromatography. The concept of ion-exchange chromatography is to coat positive and negative charges on surfaces for capturing negative and positive charges, respectively. Branched polymers can be chemically modified to possess such charges, and in this chapter, we discuss the grafting of branched polymers onto a PDMS surface, to attenuate protein non-specific adsorption, while also selectively capturing proteins of interest from the protein mixture. Many parameters must be optimized for the chromatography, such as increasing the ratio of area to volume, designing the pillar arrays for the chromatography, determining the depth of device, increasing the polymer grafting density, increasing the polymer grafting efficiency, using a suitable protein mixture concentration, etc. At this stage, we have verified the feasibility of the ion exchange chromatography concept for a PDMS-based device. More research in this area will likely arise in the future.  7.2 Future Work Although a novel antifouling mechanism and an original PDMS-based ion exchange liquid chromatography system were developed in this thesis, other new and promising directions may arise in the future.  159 1. The antifouling evaluation in vivo is a crucial step. Although we have tested the antifouling performance using plasma proteins, in vitro, the real environment of the biological body is the ultimate test for the implantable devices. 2. Patterns of proteins or cells may be possible with the aid of the polymers. While polymer coating helps to reduce non-specific protein adsorption, improved and specific protein adsorption, as well as cell adhesion, may also be possible. By modifying the polymer end groups to make them adsorb specific biomolecules, while reducing nonspecific adsorption, a contrast in the pattern of proteins or cells may be able to enhance the sensitivity of the final devices, as when they are used in biosensors. 3. Further investigation of the physical properties of branched polymers by AFM seems to be warranted. This can also aid the investigation of fundamental aspects of bio-physics. 4. From the results presented in this thesis, more research into ion exchange liquid chromatography may be a rewarding direction. At this stage, we have developed a device having an acceptable performance. Additional modifications for the device design and experimental parameters may lead to faster and more efficient protein separations. Even greater device versatility may come from the use of reversible coatings (i.e., where a polymer is grafted onto a gold-coated PDMS device and released by voltage).   160 Appendix A Experiment details of chapter 2  A.1 PZT plate performance and simulation The vibration amplitude is a linear function of applied voltage and is shown in Figure A1. The amplitude is smaller when PZT is vibrating in PBS with loads on both sides.    Figure A.1 The linear correction between the vibration amplitude of the center of PZT plate and the applied voltage.          Figure A.2 The amplitude-frequency spectrum of the center of PZT plate. The resonance frequency exhibits the largest vibration amplitude.  The resonance frequency of PZT plate is chosen to have largest vibration amplitude by 10 Vpp. The frequency is swept from 1 kHz to 100 kHz, the vibration amplitude is measured by LDV. 0 2 4 6 8 10 0.0 5.0x10-8 1.0x10-7 1.5x10-7 2.0x10-7 2.5x10-7 3.0x10-7 3.5x10-7 4.0x10-7 4.5x10-7  A m pl itu de (m ) Applied Voltage (V)  In Air  In PBS 10 12 14 16 18 20 1.0x10-7 2.0x10-7 3.0x10-7 4.0x10-7 5.0x10-7 6.0x10-7 7.0x10-7 8.0x10-7 9.0x10-7   A m pl itu de  (m ) Applied Frequency (kHz)  In Air  In PBS  161  The PZT plate vibration was simulated by ANSYSTM 8.0 and the result is shown in Figure A.3. The vibration amplitude is largest in the middle of plate and is smallest in two ends, and this is because that two ends were constrained by wires for applying voltage.      Figure A.3 The simulation of PZT plates activated by 10 Vpp at 16 kHz by ANSYSTM 8.0.       The experiment measurement of vibration amplitude on a PZT plates is shown in Figure A.4. The measurements match well with simulation, which verifies the blending mode of vibration. 0.0 0.2 0.4 0.6 0.8 0.0 2.0x10-7 4.0x10-7 6.0x10-7 8.0x10-7 1.0x10-6 1.2x10-6 1.4x10-6 1.6x10-6 0.0 0.1 0.2 0.3 0.4 0.5 A m pl itu de  (m ) Y Di men sion  (cm ) X Dim ension (cm ) Figure A.4 Experimental measurement of vibration amplitude across the PZT plate by LDV. The plate is vibrating at 16 kHz by a 10 Vpp.  162 The measured fluorescence intensity of a PZT plate after vibration for 5 min is shown in Figure A.5 (the BSA-FITC was used in the figure). The fluorescence intensities across the PZT plate were measured by Image J and were smoothed by Excel to remove spikes. From the figure, there is strong relationship between vibration amplitude and BSA adsorption (compared to the Fig. A.4). From the Fig. 2.2 (b), we found more BSA adsorption on PZT plate vibrated at resonance frequency but less applied voltage (hence less vibration amplitude). In this figure, the attenuation of BSA adsorption is found not uniform across the PZT plate.  0.0 0.2 0.4 0.6 0.8 20 40 60 80 100 0.0 0.1 0.2 0.3 0.4 0.5 Fl uo re sc en ce  In te ns ity Y Di men sion  (cm ) X Dim ension (cm )  Figure A.5 Measured fluorescence intensity after vibration for 5 min. The remained BSA after vibration depends on the vibration amplitude (Fig. A.4). The smaller vibration magnitude, the less antifouling performance can be achieved.  The fluorescence photos of surfaces charged with DC signal is shown in Figure A.6.  Figure A.6 Fluorescence photos of (a) positively and (b) negatively charged surfaces.  (a) (b)  163 A.2 Protein adsorption quantified by labeled with SYPRO ruby The SYPROTM ruby dyed was purchased from invitrogen (S-12001). The surfaces were incubated in SYPROTM ruby solution without further dilution overnight and then were washed by PBS 10 times. The dye will attach to BSA because of its peptide bonds. The dye can be excited by 450 nm light and emitted 610nm light. The fluorescence photos are shown in Figure A.7(a). After vibration for 5 min, less BSA adsorbed on the PZT plate. The corresponding fluorescence intensity is shown in Figure A.7(b). Around 45% less fluorescence intensity is observed after vibration, and the result is close to what measured by BSA-FITC. This shows that the vibration doesn’t break the BSA-FITC bond and the less fluorescence after vibration is because of the less adsorbed BSA but not less fluorescence probes.          Figure A.7 (a) Fluorescence intensity of Sypro ruby dyed BSA on bare and after vibrated PZT plates. (b) The modified BSA adsorption quantities from (a). Background                       After incubation                After vibration After incubation After Vibration 0 20 40 60 80 100   M od ifi ed  B S A  Q ua nt ity  (a) (b)  164 A.3 Vibration and temperature The application of vibration inevitably increases the temperature around the PZT plate. We measured the temperature at the center of PZT plate with different distance from surface by a digital thermometer with thermocouple. The temperature increases with the time and creases with the larger distance from surface as shown in Figure A.8. The point on surface has largest temperature increase of 37 oC. The temperature for a BSA to denature is around 60 oC [1], hence the temperature is not possible to denature BSA. 0 5 10 15 20 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0   0 cm  1 cm  2 cm Te m pe ra tu re  (o C ) Time (minutes) Figure A.8 Temperature increase tendency of points with different distance from PZT surface during application of vibration.   165 A4. Discussion on adsorption Isotherm of different molecules on different surfaces: Case studies: 1. BSA adsorption isotherm for 1 h incubation time (Figure A.9(a)). Conditions: Concentrations of BSA used were 0.1, 0.2, 0.4, 0.6, 0.8, 1, 2, and 4 mg/ml and the substrate was silver. Langmuir isotherm fitting: The fluorescence intensities were fitted to a (Langmuir isotherm equation. The fitting parameters are maximum relative adsorption quantity (fluorescence) and Langmuir equilibrium constant of adsorption (K).       Results: maximum fluorescence is 169.3, equilibrium constant is 5.1. 2. BSA adsorption isotherm for 30 min incubation time (Figure A.9(b). Conditions: Concentrations of BSA used were 0.1, 0.2, 0.4, 0.6, 0.8, 1, 2, and 4 mg/ml and the surface was silver.       Fitting results: maximum fluorescence is 105.3, equilibrium constant is 7.3.  Conclusion of 1 and 2: The maximum fluorescence at 30 min incubation time is smaller than that of 1 h incubation. But the equilibrium constants for the two cases are comparable. We have obtained good fitting (R2=0.9~0.94) in these cases. The equilibrium constants in Fig. A.9(a) and (b) are comparable which shows the similar adsorption affinity for BSA on silver surfaces with different incubation time.  3. BSA adsorption isotherm for 16 h incubation time on PEG modified surfaces (Figure A.9(c)). Conditions: Concentration of PEG used were 5000: 2*10-7, 2*10-5, 2*10-3, 0.1, 6 mg/ml and the surface was gold.       Fitting results: maximum thickness ~ 7.72 nm, equilibrium constant is 148.5. 4. Avidin adsorption isotherm for 3 min incubation time on HPG modified surface (Figure A.9(d), data not shown in the thesis). Conditions: Concentration of avidin used were 0.01, 0.033, 0.1, 1 mg/ml and the surface was HPG-P3 grafted PDMS surface. Fitting results: maximum amount of avidin adsorbed is 5.85*10-5 mg, equilibrium constant is 1.8.   166  0 1 2 3 4 0 40 80 120 160 200    conc. (mg/ml) R el at iv e Fl uo re sc en ce Equation y = (a*b*x^(1-c))/(1 + b*x ^(1-c)) Adj. R-Squar 0.90782 Value E a 169.2587 E b 5.10179 E c 0  0 1 2 3 4 0 20 40 60 80 100 120 R el at iv e Fl uo re sc en ce conc. (mg/ml) Equation y = (a*b*x^(1-c))/(1 + b*x^ (1-c)) Adj. R-Squar 0.94634 Value H a 105.3104 H b 7.3356 H c 0  0 1 2 3 4 5 6 0 2 4 6 8 Th ic kn es s (n m )    PEG Conc. (g/l) Equation y = (a*b*x^(1-c))/(1 + b*x^(1-c)) Adj. R-Sq 0.91589 Value C a 7.7268 C b 148.51 C c 0  0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 5.0x10-6 1.0x10-5 1.5x10-5 2.0x10-5 2.5x10-5 3.0x10-5 3.5x10-5 4.0x10-5 4.5x10-5   R el at iv e qu an tit y (m g) Avidin concentration (mg/ml) Equation y = (a*b*x^(1-c))/(1  + b*x^(1-c)) Adj. R-Sq 0.997 Value C a 5.84971E-5 C b 1.79032 C c 0  Figure A.9 Comparison of adsorption behavior of (a) 1 hr (Fig. 2.2 (a)) and (b) 30 min (Fig. 2.2 (b)) BSA incubation on silver surfaces; (c) 16 hr PEG incubation on gold surfaces (Fig. 4.3); (d) 3 min avidin incubation on HPG-P3 grafted surfaces (data not shown in thesis).  From the equilibrium constant comparison, it is clear that the surface affinity varies with different proteins and surfaces. Since the incubation time is different for each experiment for a given protein (Fig. A.9(a)~(d)), the equilibrium constant obtained can not be compared. But it is important to mention that the equilibrium constant varies with different proteins and surface suggesting that both surface chemistry and the type of protein influencing the adsorption.  (a) (b) (c) (d)  167 The kinetic adsorption behavior observed in Figure 3.5 in the thesis: In the case of Langmuir adsorption isotherm, the equilibrium constant defines how preferable the proteins adsorb onto the surface either in static or equilibrium conditions. However, in the case of kinetic adsorption process, the rate constant defines how fast the proteins adsorb onto the surfaces. The net rate of the adsorption reaction can be defined as the following: [2,3]  1 1(1 ) 1 l l k t l d kdt e θ θ θ − = − = −                                                               (A.1) where θl is the coverage ratio of BSA adsorbed on the surface, t is time, and k1 is the rate constant. By solving the differential equation, the relationship between θ and time is shown in Equation (A.1). By fitting this equation with the measured data (Figure A.10), the calculated k1 is 0.0653 and 0.13131 respectively for BSA and IgG, A two-step reaction has been used to describe the kinetic adsorption of proteins [2,3]. The first step is initial adsorption and the second is conformation change. In this experiment, two step equations were used to fit the data but the rate constants were found the same, implying that only one step was involved in the adsorption process. Nonetheless, the isotherm data follows an exponential growth, and this characterizes the protein layer growth on the surface.   Figure A.10 Kinetic adsorption isotherms of BSA and IgG on a SiO2 surface. The protein solution concentrations were 2 mg/ml for BSA and 0.1 mg/ml for IgG. After each different protein incubation time, the surface was washed with PBS 5 times before the fluorescence intensity measurements. Equ. (A.1) was used to fit the data.   [1] D. R. Jackson, S. Omanovic, and S. G. Roscoe, Langmuir, 16 (2000) 5449. [1] Y. Mao, W. Wei, J. Zhang, H. Peng and L. Wu, Microchemical Journal, 70 (2001) 133. [2] H. He, Q. Xie, Y. Zhang and S. Yao, J. Biochem. Biophys. Methods, 62 (2005) 191. 0 10 20 30 40 50 60 0 50 100 150 200 R2=0.96 R2=0.99   Fl uo re sc en ce  In te ns ity Incubation Time (mins)  IgG  BSA  168 Appendix B Experiment details of chapter 3  B.1 Detailed recipe for MEMS membrane fabrication in Figure 3.2 1. SOI wafer (4 inches): silicon holder-400 μm, SiO2-1 μm, silicon film-2 μm. 2. SiO2 coating by thermal oxidation (300~500 nm). 3. Lithography on silicon holder side: a. Photoresist coating. Adhesive HMDS-4000 rpm 40 s, dry for 30 s, AZ 4110-4000 rpm 40 s (thickness of photoresist ~ 1 μm). b. Soft bake on 100 oC hot plate for 10 min. c. Exposure with UV 405 nm (Canon aligner), with 12 on light integral. d. Develop using 1:4 of AZ 400K:DI water for 1~2 min. 4. BOE etching for 5~10 min till reaching silicon holder. 5. TMAH (20%, 80~90 oC) anisotropical etching for 8~9 hr till reaching SiO2 of SOI.  B.2 Simulation of device The parameters for vibration simulation of device by ANSYSTM 10.0 are shown in Table B.1. Table B.1 Parameters for simulation Stiffness Matrix [c] (1010 N/m2)         Dielectric Matrix [εr]  Piezoelectric MatrixT [eT] (C/m2)  SiO2 and Si membrane (1 μm/2 μm)        Young's module: 127GPa        Poisson's ratio: 0.242        Density: 2287 kg/m3 20.97 12.11 10.51 0 0 0 20.97 10.51 0 0 0 21.09 0 0 0 4.247 0 0 4.247 0 4.247 ⎡ ⎤⎢ ⎥⎢ ⎥⎢ ⎥⎢ ⎥⎢ ⎥⎢ ⎥⎢ ⎥⎣ ⎦ 8.33 0 0 0 8.33 0 0 0 8.84 ⎡ ⎤⎢ ⎥⎢ ⎥⎢ ⎥⎣ ⎦ 0 0 0 0 0 0.48 0 0 0 0 0.48 0 0.573 0.573 1.32 0 0 0 −⎡ ⎤⎢ ⎥−⎢ ⎥⎢ ⎥− −⎣ ⎦  169 The electromechanical constitutive equations for linear material behavior used in ANSYS are:  { } [ ]{ } [ ]{ } { } [ ] { } [ ]{ }T T C S e E D e S Eε = − = +                                                                                                                (B.1) where {T} = stress vector; {D} = electric flux; {S} = strain vector; {E} = electric field vector; [C] = elasticity matrix; [e] = piezoelectric stress matrix; and [ε] = dielectric matrix density vector Plugging the known permittivity, piezoelectric and elastic coefficient matrix (Table B.1) into the constitutive equations, the applied voltage vector can be transformed to corresponding strain vector. Beneath the piezoelectric plate is the membrane, and those needed properties for ANSYS simulation, such as poisson’s ratio, density, and Young’s mudule, are listed in Table B.1. After providing properties of material to ANSYS, the geometry of membrane and constrains are set. Since the membrane is fixed at its periphery, the constraint (no movement and no voltage applied) is set along periphery. The element types for membrane and for piezoelectric plate are SOLID 45 and SOLID 5, respectively. The solid model built for simulation in ANSYS is shown in Figure B.1(a). The center rectangle is the area of membrane. The simulated vibrating membrane at resonance frequency (the highest peak in Fig. 3.3) under 10 Vpp is shown in Figure B.1(b). The z movement across the membrane is shown in Fig. 3.4. The sequential motions are shown in Figure B.1(c)~(f).  (a) (b) (c) (d)  170   Figure B.1 (a) The solid model for simulation, (b) the deformation of membrane at resonance frequency by 10 Vpp, (c), (d), (e), and (f) are sequential motions of membrane at 0, ¼, ½, ¾ cycle of motion.  From (c)~(f), it shows the FPW mode of vibration. The wavelength of vibration is 2/3 length of membrane.  B.3 Streaming induced by vibration of membrane After activated the vibration of membrane, the induced streaming flow is generated. This flow is visualized by yeast cells and is shown in the Figure B.3(a)~(d). One yeast cell is moving by the time. When the vibration is stop, no yeast cells’ movement is observed. This verifies the streaming close to the surface and generated by the vibration. (c) (d) (e) (f)  171   Figure B.2 Sequence of moving of yeast after vibration is activated: (a) 1/30, (b) 1 1/30 (c) 2 1/30, and (d) 2 ½ s. The magnification of microscope is 200x.               (a) (b) (c) (d)  172 Appendix C Experiment details of chapter 4  C.1 Electrochemical Impedance Measurement Electrochemical impedance measurements were investigated to check the electrochemical properties of surfaces. The electrochemical impedance was performed on a Solartron 1260 frequency response analyzer and 1287 potentiostat at room temperature. The PZT plates were used as the working electrodes. A platinum plate 10 times larger than that of the working electrode was as the counter electrode. Using a large-area counter electrode is particularly important for impedance measurement, since the impedance of a counter electrode must be negligible compared with that of the working electrode. A standard Ag/AgCl electrode functioned as a reference electrode and PBS with 0.14 M NaCl and 0.01 M phosphate as the electrolyte. The measured Nyquist plot (Z”-Z’) was fitted by an equivalent electrical circuit model. The measurement was controlled by Z-Plot 2.6 and data analysis was performed by Z- View 2.6 from Scribner Associates, Inc. The operation conditions were frequency sweep from 106~0.1 Hz under constant 10mV voltage. And the data were collected 20 points per span. For vibration experiment, the procedure was mentioned in experiment section of this report.  Figure C.1 The Nyquist plot of PZT plates (gold coated PZT, PEG grafted on gold coated PZT plates). The measurement performed by a three electrodes setting in PBS. The curve is fitted by the inset equivalent circuit. The value of RΩ signifies the resistance of electrolyte, R and CPE signifies the surface resistance and capacitance of working electrode. 0 200 400 600 800 1000 0 -1000 -2000   Z'  gold in 6mg/ml PEG  gold in 6mg/ml PEG+vib 5mins  gold  gold+vib 5mims RΩ R  173 The measured Nyquist plot was shown in Figure C.1. The inset is the equivalent circuit to model the surface. In circuit, RΩ signifies the resistance of electrolyte between working and reference electrodes, R is the resistance from electrolyte to working electrode, and CPE, which is represented by the constant-phase element, signifies the capacitance of electric double layer at the electrode-electrolyte interface and PEG coating layer on the working electrode. When Z” is zero at high frequency, the impedance Z’ mainly comes from the resistance of electrolyte, and this will change slightly because of different distance between working and counter electrodes. Compared to the resistance of electrolyte/electrode, the electrolyte resistance is much smaller and without significant different between 4 plates which data collected in Table C.1. However, the PEG grafted surface after immersed in 6 mg/ml initial PEG solution for 16 hr shows significant larger R, which is alike to reports by many other researchers [1-4]. From Fig. C.1, the PEG grafted surfaces show larger semi-circle and from Table C.1, the R of gold, gold after vibration, PEG grafted gold, and PEG grafted gold after vibration surfaces are 3486 ± 2.45%, 3495 ± 2.11%, 19271 ± 9.50%, and 19570 ± 8.25% Ω, respectively. From Fig. C.1 and Table C.1, resistance of electrolyte/electrode increased because of another PEG layer, and has no significant difference before and after vibration application. Hence, there is no significant loss of grafted PEG from surface after vibration.  Table C.1 The resistances R and RΩ of gold and PEG grafted gold surfaces with or without vibration for 5 min.   Gold Gold + PEG   wo vibration w vibration wo vibration w vibration Rl (Ω,%) 60.10 0.34 57.42 0.50 51.91 0.75 71.15 0.79 RΩ (Ω,%) 3486 2.45 3495 2.11 19271 9.50 19570 8.25  C.2 Insulation layer SiO2 deposition method The SiO2 insulation layer in Illustration 4.1 is deposited by e-beam evaporator. Originally, the PECVD (Trion) was used to deposit more compact SiO2. However, PECVD grown SiO2 has fluorescence intensity, and the fluorescence intensity is believed coming from remanent carbon in the SiO2 after process [1,2]. The raw material, (C2H5)2SiH2 (DES), is more environment benign than SiH4, however, longer C2H5- chain may cause remanent carbon after deposition  174 process [1]. Those carbons can react with Si to form SiC, which have fluorescence bands of 460 and 535 nm and are close to 495 nm emitting band of FITC [2]. Compared to the thermal grown SiO2, the fluorescence of PECVD grown SiO2 increased with the increase of film thickness and was shown in Figure C.2. The fluorescence intensity comes from this layer make the background intensity too strong and can’t be differentiated with the fluorescence from protein. Hence, e- beam evaporated SiO2 was deposited as insulation layer.     Figure C.2 Fluorescence intensities of SiO2 by different deposition methods.       [1] A. R. Barron, Adv. Mater. Opt. and Electronics, 6 (1996) 101. [2] Dihu Chen, Z. M. Liao, L. Wang, H. Z. Wang, Fuli Zhao, W. Y. Cheung and S. P. Wong, Opt. Mater., 23 (2003) 65.           Thermal 500nm PECVD 200 nm PECVD 400nm 0 20 40 60 80 100 120  Fl uo re sc en ce  In te ns ity  175 Appendix D Experiment details of chapter 5  D.1 Supporting information of chapter 5 The description of Table D.1 and Figure D.1~D.6 are written in Chapter 5.  Table D.1 Dry thickness of disulfide polyglycerol and polyglycerol adsorbed onto gold surface measured by ellipsometry.    PG  PG-S-S-PG low MW  PG-S-S-PG high MW Conc. (mg/ml)  1  6  1  6  1  6 Thickness(nm)  0.117  0  0  0  0.132  0.0504 SD (nm)  0.047  0  0  0  0.220  0.0560  Figure D.1(a) 1HNMR spectra of low molecular weight HPG-O-CH2-CH2-S-S-CH2-CH2-O- HPG. (a)  176  Figure D.1(b) 1HNMR spectra of high molecular weight HPG-O-CH2-CH2-S-S-CH2-CH2-O- HPG.      Figure D.2 UV-Vis spectra of mono thiol, disulfide HPG after reacting with DTNB.  (b)  177  Figure D.3(a): MALDI-TOF spectra of HPG-SG-L (a)  178 Figure D.3(b): MALDI-TOF spectra of HPG-SG-H  (b)  179 4000 3500 3000 2500 2000 1500 1000 500 0.10 0.15 0.20 0.25 0.30 0.35 0.40 0.45  Ab so rb an ce Wavenumber (cm-1)  HPG-SH-L  HPG-SH-H  Figure D.4 FTIR spectra of HPG-SH-L and HPG-SH-H.                 (a) 1000 800 600 400 200 0 0.0 2.0x105 4.0x105 6.0x105 8.0x105 1.0x106 1.2x106 1.4x106 1.6x106 1.8x106 2.0x106 C P S Binding Energy (eV)  6  1  2e-3  Bare  180              Figure D.5 XPS spectra of surface grafted (a) HPG-SH-L, and (b) HPG-SH-H at different polymer incubation concentration.              Figure D.6 The HPG-SH-L film thickness and C/O ratio calculated from the XPS spectra. The intensities of Au peak from the bare gold surface and polymer coated surfaces were used for the calculation of film thickness. 1000 800 600 400 200 0 0.0 2.0x105 4.0x105 6.0x105 8.0x105 1.0x106 1.2x106 1.4x106 1.6x106 1.8x106 2.0x106 2.2x106 C P S Binding energy (eV)  6      mg/ml C/O=2.450  1      mg/ml C/O=2.499  Bare            C/O=4.216 (b)  10-3 10-2 10-1 100 101 0 1 2 3 Th ic kn es s (n m ) Bulk Polymer Conc. (g/l) 1 2 3 4  C /O  R atio  181 D.2 AFM images of topography of surfaces and force curve measurement The AFM images on more surfaces with different graft conditions are shown in Figure D.7. All the three polymers produced uniform surface coverage at concentrations studied as evident from both 2D scan and 3D images. There are finite differences between the polymer coated surface and bare gold surface. The topography of the polymer films obtained from linear PEG and hyperbrached polyglycerols was also different. The height and roughness of grains increase with the increase of surfaces polymer densities.                         2μm  (a) (c) (d) (b) (e) (g) (f)  182           Figure D.7 AFM images showing the surface morphology of bare gold surface (a); linear mPEG-5000-grafted surfaces (b), (c), (d); low molecular weight HPG-SH-L-grafted surfaces (e), (f), (g); and high molecular weight HPG-SH-H-grafted surfaces (h), (i), (j) at 2x10-5 mg/ml (b, e, h), 2x10-2 mg/ml (c, f, i ), 6 mg/ml (d, g, j) incubation concentration. For a given sample 3D image (upper) and section plot (lower) is shown.   The force curve of Si3N4 tip and bare gold substrate is obvious different with polymer grafted surface. The gold surface is rigid, when the tip is approaching to the very proximity of surface, the repelling force increased suddenly. For PEG coated surface, however, when tip is approaching the surface, the repelling force curve gradually increases as shown in Figure D.8(a). The buffer length to sense a rapid increased repelling force by AFM of polymer-grafted surfaces is close to 10-20 nm, which indicates the hydrated thickness of polymer in buffer solution and is higher than dry thickness because of swelling of polymer. For hyper branched polymer coated surface, the hydrated thickness is smaller compared to linear one, and is shown on Figure D.8. In addition, HPG-SH-L-grafted surface shows shorter buffer length compared to HPG-SH-H- grafted surface. For HPG-SH-L-grafted surface, the approaching curve shows the combination effect of polymer and rigid gold surface, which indicates the lowest grafting density of HPG-SH- L among three polymers.  (h) (i) (j)  183 The retracting force curves of bare gold, mPEG-5000-, HPG-SH-L-, and HPG-SH-H-grafted surfaces are shown in Figure D.8(b).  When the tip is retracted, there is a short range VDW attraction between tip and bare gold surface shows, which is usually within 10nm. On all three polymer-grafted surfaces, stronger attractions are detected. And compared to linear polymer- grafted surface, HPG-SH-H- grafted surface (similar molecular weight with linear mPEG 5000) shows stronger attraction force to AFM tip, which means more attracted points between polymer and tip because of the branched structure of polymer. 0 10 20 30 40 50 60 -0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4  Fo rc e (n N ) Distance (nm)  Gold  PEG 5000  PG 1500  PG 4000  0 10 20 30 40 50 60 -1.0 -0.5 0.0 0.5 1.0 1.5   Gold  PEG 5000  PG 1500  PG 4000 Fo rc e (n N ) Distance (nm)   Figure D.8 (a) Approaching force curve of tip and bare gold, mPEG-5000-, HPG-SH-L-, and HPG-SH-H-grafted surfaces. (b) Retracting force curve of tip and bare gold, mPEG-5000-, HPG- SH-L-, and HPG-SH-H-grafted surfaces.          (a) (b)  184 Appendix E Experiment details of chapter 6  E.1 Supporting information of chapter 6 control nonactive active 0 20 40 60 80 100 120    native PDMS  P2 on 40W plasma treated surface  P2 on 55W plasma treated surface C on ta ct  a ng le  (o )   Figure E.1 Water contact angle on different surfaces.  Control represents bare PDMS for native PDMS, and allylamine treated surfaces for HPG-P2 grafted PDMS surfaces. Nonactive represents the HPG-P2 polymer without reactive succinimidyl ester groups. Active represents the HPG-P2 polymer with reactive succinimidyl ester groups. After the plasma treatment (10 min), the surfaces were washed with soap, Mili-Q water, methanol, and Mili-Q water. Then the polymers were allowed to react with plasma modified PDMS surfaces for 1 hr at room temperature with gentle stirring. Then the surfaces were sonicated in Mili-Q water for 10 min to remove adsorbed polymer. As shown in Figure E.1, the plasma treatments (40, and 55 W) reduce the contact angle suggests an amine modification  185 PDMS surfaces. Even lower power (15W) showed very little change on contact angle (data not shown), and the surface lost some transparency under higher power (70W). So only 40 and 50 W plasma treated surfaces were proceed for further polymer graft.  The HPG-P2 polymer without active groups also showed some reduction in contact angle, however, more hydrophilicity was produced by grafting HPG-P2 polymer with active groups. In addition, after 55 W plasma treatment, the polymer seems were grafted more on PDMS surface. Hence, stronger plasma energy (55W) was used for the further experiment in this manuscript.   Control P1 P2 P3 P4 0 20 40 60 80 100 120   C on ta ct  a ng le  (o )  Figure E.2 Contact angles of different polymers and/or different grafting methods coated PDMS surfaces.  The terms except HPG-P1 and HPG-P2 mentioned in this report are the surfaces treated by: Control, 55W plasma treatment; P3, HPG 2500 aldehyde reacted with plasma treated PDMS surface for 1 hr; P4, HPG 2500 aldehyde reacted with plasma treated PDMS surface for 2 hr. The HPG-P3 and HPG-P4 were grafted on surfaces by firstly mixing polymer solution at 1.4 mg/ml concentration with NaCNBH4 and quenched with glycine. The reaction between mixture solution with PDMS surfaces were 1 hr (HPG-P3) and 2 hrs (HPG-P4). After reaction, the same amount of NaCNBH4 was added into solution for another 60 min. From Figure E.2, there is no  186 significant difference on the contact angle for 4 polymer grafted surfaces. Although both grafting methods are successful, method in HPG-P1 and HPG-P2 cases is easier for polymer coating for microfluidic devices.  E.2 Detailed recipe E.2.1 Plasma treatment The plasma treatment is three steps. 1. Ar plasma: 50 sccm flow rate, 350 mTorr chamber pressure, 75 W plasma power for 1 min. 2. Allylamine plasma: 80 sccm flow rate, 540 mTorr chamber pressure, 55 W plasma power for 10 min. The chamber temperature increased 3~5 oC after this step. 3. Oxygen plasma cleaning after the sample taking out: 600 sccm flow rate, 800 mTorr chamber pressure, 250 W plasma power for 10 min. This step repeats twice to fully clean remaining allylamine in the chamber after allylamine plasma treatment. E.2.2 Silicon mold fabrication for PDMS device 1. Start from silicon substrate (clean with RCA1 protocol). 2. Pour SU8 2025. 3. Spin substrate at 500 rpm for 10s and then 3000 rpm for 30 s. 4. Soft baking on 2 separated hotplate (65 and 95 oC) by: 2 min 65 oC, 4 min 95 oC, another 2 min 65 oC. 5. UV exposure 55 s. 6. Post exposure baking on 2 separated hotplate (65 and 95 oC) by: 1 min 65 oC, 3 min 95 oC, another 1 min 65 oC. 7. Develop around 5 min.

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0067165/manifest

Comment

Related Items