Surface modification for optical biosensor applications

ETH Library

Surface modification for opticalbiosensor applications

Doctoral Thesis

Author(s):Hofer, Rolf

Publication date:2000

Permanent link:https://doi.org/10.3929/ethz-a-004040300

Rights / license:In Copyright - Non-Commercial Use Permitted

This page was generated automatically upon download from the ETH Zurich Research Collection.For more information, please consult the Terms of use.

https://doi.org/10.3929/ethz-a-004040300

http://rightsstatements.org/page/InC-NC/1.0/

https://www.research-collection.ethz.ch

https://www.research-collection.ethz.ch/terms-of-use

Diss. ETHNo. 13873

Surface Modification for Optical

Biosensor Applications

for the degree of

DOCTOR OF NATURAL SCIENCES

A dissertation submitted to the

SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH

presented by

ROLF HOFER

Dipl. ehem., University of Fribourg, Switzerland

bom February 3, 1961

citizen of Langnau i.E., Bern

accepted on the recommendation of:

Prof. Dr. N.D. Spencer, examiner

Dr. M. Textor, co-examiner

Dr. M. Ehrat, co-examiner

Zürich, 2000

Oui tfc

3 i *•

1 111 ..:

Contents

Abstract 9

Zusammenfassung 12

I Introduction 15

II Theory and Background 18

1 Biosensor Technology 18

1.1 Recognition Elements 19

1.2 Immobilization of Recognition Elements 20

1.2.1 Adsorption 21

1.2.2 Covalent Binding 22

1.2.3 Binding through Biotin-Streptavidin-Link 22

1.3 Transducer Technique 23

1.3.1 Detection Methods 24

1.4 Specific and Non-specific Binding 24

2 Optical Sensors 25

2.1 Principle of Optical Waveguide 26

2.2 The Evanescent Field 28

2.3 Analytical Methods with Optical Waveguide 30

2.3.1 Change of Refraction Index 30

2.3.2 Fluorescence Measurement 31

HI Material and Methods 32

1 Substrate Material and Chemical Products 32

1.1 Substrate 32

1.1.1 Metal Oxides 32

1.1.2 Metals 34

1.1.3 GoldAdlayer 34

1.1.4 Planar Waveguide Chips 35

1.2 Adsorbents 36

1.2.1 Organic Phosphoric and Phosphonic Acids 36

1.2.2 Poly-(L)-Lysine-Poly(Ethylene Glycol) (PLL-PEG) 45

1.2.2.1 Synthesis of PLL-g-PEG Derivatives 46

1.3 Biochemical Substances 53

1.4 Solvents 54

1.5 Further Substances 55

2 Apparatus 56

2.1 Apparatus for the Cleaning and Coating of the Substrates 56

2.2 Surface Characterization 58

2.2.1 X-ray Photoelectron Spectroscopy (XPS) 58

2.2.2 Time-of-Flighi Secondary-Ion Mass

Spectrometry (ToF-SIMS) 59

2.2.3 Contact Angle 60

2.2.4 Microdroplet Density (u.dd) 61

2.2.5 Atomic Force Microscopy (AFM) 64

2.3 Determination of Adsorption 65

2.3.1 Grating Coupler 65

2.3.2 Optical Waveguide Lightmode Spectroscopy (OWLS) 66

2.4 Determination of Bioactivity 68

2.4.1 Determination of Enzyme Activity 68

2.4.2 Fast Optical Biospecific Interaction Analysis (FOBIA) 68

IV Results and Discussion 72

1 Cleaning of the Substrate Surfaces 72

1.1 Cleaning Procedures 72

1.2 Results 73

1.3 Influence on Further Surface-Modification Steps 78

1.4 Influence on the Waveguide Properties 80

2 Coating of the Substrate Surfaces 85

2.1 Self-Assembled Monolayers (SAMs) 85

2.1.1 Octadecyl Phosphate (ODP04) 86

2.1.1.1 Influence of Different Solvents 86

2.1.1.2 Adsorption Kinetics and SAM Stability 88

2.1.1.3 Characterization ofODP04 SAMs on Ta205 97

2.1.1.4 Summary, Conclusions and Suggestions 114

2.1.1.5 Molecular Model of ODP04 SAM on Ta205 122

2.1.2 Mono- and Bifunctional Phosph(on)ates 131

2.1.2.1 Applied Methods 131

2.1.2.2 Influence of Different Solvents 132

2.1.2.3 SAM Orientation and Order 136

2.1.2.4 Chemical Bonding 146

2.1.2.5 Summary 149

2.1.2.6 Toxicology 151

2.1.3 Ammonium Salt of Alkyl Phosphates 153

2.1.3.1 Applied Methods 154

2.1.3.2 Sample Preparation 155

2.1.3.3 Investigation of Surface Properties.... .....156

2.1.3.4 Exposure to Inorganic Phosphate Solution 169

2.1.4 Poly-(L)-Lysine Poly-(Ethylene Glycol) (PLL-PEG) 173

2.2 Adsorption of Thiols onto Gold Interfaces 178

2.2.1 Coating ofTa205 with Gold Chloride 178

2.3 Stability of Modified Surfaces 190

2.3.1 Stability of Hydrophobic Phosph(on)ate SAMs in Water 190

2.3.2 Stability of Phosph(on)ate SAMs and PLL-PEG

Adlayers in Different Buffer Systems 191

3 Immobilisation of Biomolecules and

Quantification of Target Molecules 197

3.1 Enzymatic Activity of Horseradish Peroxidase 198

3.1.1 Peroxidase Activity on PLL-PEG Coated Chips 200

3.1.2 Peroxidase Activity on Phosph(on)ate Coated Chips 204

3.2 Optical Biosensor Application 208

3.2.1 Labeling of Biomolecules 208

3.2.2 Specific and Non-Specific Binding of Target Molecule 210

3.2.3 Specific and Non-Specific Binding of Streptavidin 215

3.2.4 Detection Limit of Target Molecules 218

3.2.5 Optimization of Biotin Concentration in

PLL-PEG Derivatives 224

3.2.6 Immunoassay Design 227

3.2.7 Application of the Sensor for Serum Samples 232

3.2.7.1 PLL-PEG Sensor Interfaces 234

3.2.7.2 Hydrophobic Alkyl PhosphateSensor Interfaces 235

3.2.7.3 Comparison of PLL-PEG and Alkyl

Phosphate Sensor Interfaces 236

3.2.8 Mixed SAMs for Biosensor Applications 237

3.2.9 Summary of the Immunoassay Performance 240

V Conclusions and Outlook 241

VI Literature 245

VII Appendix 252

Curriculum Vitae 298

Acknowledgements 302

il)*19* y il

Seht Ihr den Mond da stehen

Er ist nur halb zu sehen

Und ist doch rund und schön

So gibts noch viele Sachen

Die wir getrost belachen

Weil unsre Augen sie nicht sehn

Aus dem Lied, "Der Mond ist aufgegangen"

von Matthias Claudius (1740-1815)

Meiner Familie

Seite Leer /

Blank leaf

Abstract

Abstract

The thesis covers the development of two novel surface modification

systems based on monomolecular assembly of functionalized molecules on

oxide surfaces and their successful application for optical bioaffmity sensor

chips. One or both of the two techniques developed will be used in the near

future to fabricate optical waveguide chips on a commercial scale by the

industrial project partner, thanks to the combination of technical

performance, ease of application and cost-effectiveness of the developedtechnology.

Bioaffmity sensors typically use the concept of specific, biologicalinteractions between immobilized recognition elements (antibodies,enzymes, nucleic acids, ...) and their corresponding target molecules

(antigen, substrate, complementary nucleic acids, ...) in an analyte solution

for application in medical diagnostics, drug development and highthroughput screening. The analytical requirements, in particular the

sensitivity and selectivity of an particular affinity assay, depend on both the

type of detection technique and the architecture of the transducer/analyteinterface. The latter is particular critical to the sensor performance and

reliable, cost-effective surface functionalization techniques are urgentlyneeded to satisfy the needs of a rapidly growing market.

The thesis focusses on detailed studies of spontaneously formed,organized monolayers of long-chain alkanephosphates and of poly(ethyleneoxide) grafted poly-lysine copolymers on optically transparent, high-refractive-index thin films such as tantalum, niobium and titanium oxides,the study of their mterfacial chemistry and physico-chemical behavior, and

their quantitative performance in model bioaffmity sensor assays. The

requirements and objectives for the biosensor surfaces were: high opticaltransparency, high sensitivity, high specificity and general applicability for

the reproducible immobilization of recognition elements and detection of the

corresponding target molecules.

9

Abstract

a) The high optical transparency was a fundamental condition, since the

biosensor surfaces are to be part of an evanescentfield waveguide opticalsensor system. The apparatus used for testing the performance of the

adlayers consists of a planar waveguide. The optical signal for the analytedetection is measured through the sensor surface and its support. As

waveguiding layers, highly transparent, high refractive index (tantalumpentoxide) coated glass chips were used.

b) The high sensitivity is obtained by selective excitation of fluorescent

chromophores of tracer molecules immobilized at the sensor/analyteinterface and excited by the evanescent field of the planar waveguide.This allows the detection of surface species with almost no contribution

from the bulk solution.

c) The high specificity is reached by minimization of non-specific bindingof the target molecules and tracers at the sensor surface and optimizationof specific binding by the optimized immobilization of recognitionmolecules (e.g. antibodies). As a model bioaffmity assay system, an

antigen/antibody reaction based on immunoglobulin (rabbit-IgG/anti-rabbit-IgG) was selected.

d) Two general sensor interface platforms were developed and tested

according to c): The first consists of the formation of self-assembledmonolayers (SAMs) of alkyl phosphates and phosphonates at the metal

oxide surfaces. Such SAMs form two-dimensionally ordered, highlyhydrophobic surfaces that strongly adsorb albumin, thus reducing non¬

specific adsorption to very low values. The binding of biotin-conjugatedbovine serum albumin (BSA-biotin) proved to be a viable technique to

produce biotin-activated surfaces that can be further functionalized usingstreptavidin. The second development covers the modification of the

oxide surface with poly(ethylene glycol)-grafted copolymers.Polyethylene glycol)-grafted poly-L-lysine (PLL-g-PEG), chargedpositively at physiological pH, adsorbs to negatively charged oxide

surfaces by electrostatic interactions. The polymer architecture was

optimized in terms of the molecular weights of the PLL and PEG

constituents and of the grafting ratio PEG/PLL. Extremely low values of

non-specific adsorption in full serum (generally less than 3 ng-cm"2

10

Abstract

corresponding to typically less than 0.1% of a typical protein monolayer)could be realized. Again, functionalization via biotin conjugation(terminal PEG-chain functionalization) was used to performbiotin/straptavidin-based bioaffmity model assays and to determine the

comparative sensing performance of the two surface modification

technqiues developed.

It was demonstrated that both surface modification routes provideexcellent general platforms on which streptavidin-conjugated recognitionmolecules can be successfully immobilized. In addition it is also possible to

form a streptavidin adlayer on these platforms followed by immobilization

of biotin-conjugated recognition molecules. Within the framework of the

thesis it has been shown that effective target molecule concentrations in the

femtomolar region could be easily detected, even in total serum samples and

without any sample pretreatment such as extraction or purification.

The combination of optimized interfacial chemistry, the opticaltransducer, and the assay architecture resulted in a optimum performancesystem that has a high potential for immunoassay applications in the future.

11

Abstract

Zusammenfassung

Die Doktorarbeit beschreibt die Entwicklung zweier neuartigerOberflächenmodifikations-Systemen, welche auf der Bildung von

selbstorganisierten, monomolekularen Adsorptionsschichten an Metalloxid¬

oberflächen und deren erfolgreichen Anwendung auf dem Gebiete der

optischen Biosensorik basieren. Eine (oder evtl. beide) der entwickelten

Techniken wird in nächster Zukunft Anwendungen zur Herstellung optischerWellenleiter finden, welche in kommerziellem Massstab hergestellt werden

sollen. Diese, durch den industriellen Projektpartner angestrebte Anwendungwird durch die optimale Kombination technischer Leistungsfähigkeit,einfacher Anwendbarkeit und Kostengünstigkeit der entwickelten

Technologie ermöglicht.

Bioaffinitäts-Sensoren basieren typischerweise auf dem Prinzip der

spezifischen, biochemischen Wechselwirkung zwischen immobilisierten

Erkennunkselementen (Antikörper, Enzyme, Nuclemsäuren, ...) und deren

entsprechenden Bindungspartnern (Antigen, Substrat, komplementäreNucleinsäuren, ...). Solche Sensoren finden unter anderem Anwendung in

der medizinischen Diagnostik, der Pharmaforschung sowie zur Bewältigungvon grossem Probendurchsatz in der Analytik. Das Erreichen der

analytischen Anforderungen, insbesondere die hohe Empfindlichkeit und

Selektivität, an die einzelnen Affinitätsassays, hängt von der

Detektionstechnik sowie von der Architektur der Transducer/Analyt-Grenzfläche ab. Letzteres ist insbesondere massgebend für die

Leistungsfähigkeit und Zuverlässigkeit der Sensoren. KostengünstigeOberflächenfunktionalisierungs-Techniken sind dringend gefragt, um der

rasch steigenden Nachfrage und dem wachsenden Markt zu genügen.

Das Hauptgewicht der Doktorarbeit liegt in den detailierten Untersu¬

chungen spontan gebildeter Monoschichten von langkettigenAlkylphosphaten und -phosphonaten sowie Polyethylenglycol-Polylysin -

Copolymeren an optisch transparenten Metalloxidoberflächen. Die

Oberflächenchemie sowie die physikalisch-chemischen Eigenschaftensolcher Adsorbatschichten wurden insbesondere an hochbrechenden

12

Abstract

Metalloxiden wie Tantal-, Niob- und Titanoxid untersucht und deren

Leistungsfähigkeit als Bioaffinitäts-Sensorplatform anhand eines

Modelassays geprüft. Die Anforderung an solche Biosensoroberflächen

waren: hohe optische Transparenz, hohe Empfindlichkeit, hohe Spezifitätund generelle Anwendbarkeit zur reproduzierbaren Immobilisierungverschiedener Erkennungseinheiten und Detektion der entsprechendenZielmoleküle.

a) Die hohe optische Transparenz war eine Grundbedingung, da die

Biosensoroberflächen auf ein optisches Analysensystem anwendbar sein

sollen, welches auf dem Prinzip der optischen Evaneszentfeld-Wellenleiter basiert. Das Gerät mit dem die Leistungsfähigkeit bezüglich

Bioassay untersucht wurden, verwendet planare Wellenleiter-Technik,

wobei die Detektion des Analyten durch die Sensoroberfläche und den

Träger hindurch erfolgt. Die hohe Transparenz wurde durch Verwendungdefinierter Tantalpentoxid-Beschichtungen von Glasträgern erreicht.

b) Die hohe Empfindlichkeit wird durch optische Anregung fluoreszieren¬

der Chromophore der sogenannten Tracer erzielt, was eine sehr starke

Reduktion von nichtspezifischem Grundsignal zur Folge hat. Die

Anregung durch das evaneszente Feld der planaren Wellenleiter, erlaubt

Untersuchungen von Oberflächenphänomenen mit weitgehender

Vermeidung von Bulksignal.

c) Die hohe Spezifität wird durch Minimierung der unspezifischen Bindungder Zielmoleküle und Tracer an die Sensoroberfläche bei gleichzeitigerErhöhung der spezifischen Bindung durch optimierte Immobilisierungvon Erkennungsmolekülen (Antikörper) erzielt. In dieser Arbeit wurde

ein Antigen/Antikörper-Modelsystem (Rabbit-IgG/anti-Rabbit-IgG)verwendet.

d) Zwei verschiedene Sensoroberflächen-Fiatformen wurden entwickelt

und nach den Gesichtspunkten die unter Punkt c) besprochen wurden

untersucht: Einerseits die Erzeugung selbstorganisierter, monomoleku¬

larer Adsorptionsschichten (= SelfAssembled Monolayers, SAMs) von

Alkylphosphaten und -Phosphonaten auf Metalloxidoberflächen. Solche

SAMs bilden zweidimensional geordnete, stark hydrophobeOberflächen, welche sich sehr gut eignen, Albumin stabil zu binden

13

Abstract

(bewirkt starke Reduktion von unspezifischger Proteinbindung). Durch

Verwendung von biotinyliertem Rinderserum-Albumin (= BSA-biotin),können auf diese Weise Biotin-aktivierte Sensoroberflächen hergestelltwerden, welche ihrerseits zur Oberflächenfunktionalisierung mittels

Streptavidin genutzt werden können. Eine zweite Oberflächenbeschich-

tungs-Variante besteht aus Adsorbatschichten von Polylysin - Poly-ethylenglycol-Copolymeren (PLL-g-PEG). Diese Polymere sind im

neutralen pH-Bereich positiv geladen und zeigen starke Adsorption an

negativ geladene Oxidschichten durch elektrostatische Wechsel¬

wirkungen. Die Polymerzusammensetzung wurde durch Anpassung der

Edukt-Molekulargewichte (Polylysin (PLL) bzw. Polyethylenglycol(PEG)) sowie der Bindungsverhältnisse (PEG/PLL) optimiert. Dadurch

konnten extrem tiefe Werte bezüglich nichtspezifischer Protein-

adsorption aus Vollserum (generell weniger als 3 ng/cm", was wenigerals 0.1 % einer Proteinmonolage entspricht), erreicht werden. Biotin-

Funktionalisierung mittels polymergebundener, endständig biotinylierterPEG-Ketten wurde verwendet um Biotin/Streptavidin-basierendeBioaffmitätsassays herzustellen und die Sensorik-Performance der

beiden Systeme zu vergleichen.

Es wurde gezeigt, dass beide Varianten ausgezeichnete Eigenschaftenin der Anwendung als generell verwendbare Oberflächen aufweisen, an

welche Streptavidin-konjugierte Erkennungsmoleküle immobilisiert oder,über eine Streptavidin-Zwischenschicht, Biotin-konjugierte Rezeptorengebunden werden können. Im Rahmen dieser Arbeit konnte gezeigt werden,dass mittels dieser definierten Oberflächen effektive Analyt-Konzentra-tionnen im femtomolaren Bereich nachgwiesen werden können. Auch in

Vollserum werden solche Nachweisgrenzen erreicht, ohne dass zusätzliche

Probenvorbereitungen wie Extraktion oder Fällungsreaktionen notwendigsind.

Die Kombination von optimierter Zwischenphasenchemie, optischemTransducer und der verwendeten Assay-Architektur ergeben ein hoch

leistungsfähiges System, das ein hohes Potential für die zukünftigeAnwendung auf dem Gebiet der Immunoassay-Sensorik hat.

14

I Introduction

I Introduction

Quantitative analytical methods in biomedical diagnosis require highsensitivity, high specificity and high reproducibility. Bioaffmity sensors are

tools which have the potential to fulfil these demands using the highlyspecific interaction of recognition molecules such as oligonucleotides, DNA-

sequences, enzymes or antibodies with their corresponding partners.

The trend in diagnostics is towards the early detection of health or

disease markers such as growth factors (e.g. intcrleukin) resulting in the

need for lower detection limits. The design and introduction of much more

efficient medicine replace drugs of higher dosage (such as the replacementof barbiturates by benzodiazepines in the past). For pre- and neo-natal

diagnostics, only a very small amount of biological material is at one's

disposal for analysis. Very low detection limits are the consequence of the

aforementioned facts of low target-molecule concentration and small

analytical material quantity. Biochemical recognition elements, such as

antibodies, enzymes or oligonucleotides are ideal functions for highlyspecific detection of target molecules. The oriented immobilization of such

recognition elements becomes increasingly important with the need of

miniaturization for the sensitive analysis of sample volumes in the microliter

scale.

We have demonstrate that capture molecules, such as antibodies or

enzymes, can be immobilized via hydrophobic-hydrophobic interactions

onto lipophilic surfaces. Alkyl phosphoric acids form hydrophobic, highlyordered self-assembled monolayers (SAMs) on planar waveguide (PWG)-surfaces. The specific binding of target molecules (e.g. antigen) to capturemolecules (e.g. monoclonal antibodies) and the subsequent specificadsorption of a tracer (e.g. fluorescent-labeled polyclonal antibodies) on

another moiety of the antigen is a widely used immunoassay setup. Within

the framework of this thesis, the immobilization of the recognitionmolecules such as antibodies or enzymes was performed in an oriented way,so as to avoid that a part of the recognition molecules is bound by the

bioactive moiety onto the surface. That would make the recognitionelements inaccessible for target molecules. Poor orientation of recognitionelements is believed to lower the sensitivity of the sensors. The non-specific

15

I Introduction

binding of target and/or tracer molecules onto the sensor surface can be

reduced by passivating the surface that is free of capture molecules bymeans of bovin serum albumin (BSA) [Wert, 1999].

The oriented binding of recognition molecules on the sensor surface is

the aim of many research groups. A straightforward way to oriented

immobilization of recognition molecules is to use the well known binding of

biotin to streptavidin [Anza,1993; Ahlu,1991]

We synthesized different derivatives of poly(L-lysine) grafted poly¬ethylene glycol) (PLL-g-PEG), which adsorb easily on negatively chargedsurfaces by Coulomb interaction due to the inherent positively chargedamino groups of the poly(L-lysine) backbone at neutral pH. BiotinylatedPLL-g-PEG (PLL-PEG/PEG-biotin) was synthesized by grafting of biotin

functionalized PEG. The copolymer has the advantage of imposing proteinresistance to the surface (very low non-specific binding of target molecules

and/or tracer) due to the PEG moiety [Hubb,2000] and allowing for defined

and oriented binding of either streptavidin conjugated recognition molecules

or biotin conjugated recognition molecules via a streptavidin interface. The

defined fraction of PEG-biotin gives us the possibility to find an optimum of

the binding density of recognition molecules at the surface.

The overall goal of this work is the design and development of a

biocompatible platform that fulfils defined performance criteria in a

bioanalytical assay. Specifically, the aim is to overcome existing limits for

ultra-sensitive detection without any loss of specificity. With the

introduction of such bioaffmity sensors, the time to result can be shortened

considerably, which is important for emergency analysis, for example. This

would also result in a much lower cost of routine analysis in human

diagnostics. The possibility of miniaturization of the detection area of such

sensor surfaces is needed for pre- and neo-natal diagnosis, biopsy samplesand in other fields where only small amounts of analytical material are at

one's disposal.

Selective detection of surface-bond fluorophores, specificallyrecognized by e.g. antibodies, can be achieved by evanescent field

detection. Introducing interface chemistries that allow oriented, reproducibleand stable immobilization of biochemical receptor molecules on top of

transducer surfaces increases the sensitivity of the bio-affinity sensor.

16

I Introduction

In terms of detection sensitivity, real-time monitoring and recognitionelements with high affinities are needed to detect analyte concentrations in

the femtomolar (10" ^M) range. The technique of fluorescence detection

using an optical evanescent field is shown to be suitable for detection of

target molecules in this concentration range. The system is not limited to

immunoassays. DNA/RNA detection is also possible when using recognitionunits based on oligonucleotides. This makes it an important generaltechnique for future bioanalytical research and application in genomics and

proteomics .

17

TT Theory

II Theory and Background

1 Biosensor Technology [Spic.1998]

The expression "biosensor" has not been clearly defined until now

[Jana, 1998]. The definition of functionality, components and types of

sensors were proposed by IUPAC (International Union of Pure and AppliedChemistry) in 1991 [Hula, 1991].

Living organisms possess exquisite recognition elements such as

enzymes, antibodies or gene probes, often referred to bioreceptors, which

allow specific interaction and detection of complex chemical or biologicalspecies. The basic setup of a biosensor includes such biological elements, as

well as transducer elements such as an electrode, an optical device or a

piezo-electric crystal. In a biosensor setup, the bioreceptor either converts an

analyte (enzymatic reaction) or recognizes the analyte (antibody-antigenreaction or DNA-probe). The transducer transforms either the primarybiochemical reaction (e.g. consumption of substrate by an enzyme, changeof current or increase in mass) or the secondary chemical reaction (e.g.further reaction of enzymatically decomposed substrate) into a quantifiableoptical or electrical signal. This signal is amplified, processed and displayedin a suitable form. Biosensors combine the specificity and sensitivity of

biological recognition systems with the optimized processing of data.

[Higg,1987]

The choice of recognition element and transducer depends on the

application, the selectivity and the sensitivity demands on the system as well

as on the physico-chemical properties of the samples (gas, liquid, pH, etc.).

18

II Theory

Biosensor applications range from clinical analysis in diagnostics[McNe,1993] through environmental [Roge,1992], agricultural [Desh,1993],and nutrition research [Desh,1994] to process controlling [Sehe, 1987].

1.1 Recognition Elements

Antibodies (Ab) are highly specific recognition biomolecules. From the

chemical point of view, they are proteins, being produced in living species(e.g. animals) by an immuno reaction against a substance that the organismhas been exposed to (e.g. via injection). Typically, antibodies are extracted

from blood, livers or hen eggs and purified by chromatography, for example.Their binding to the corresponding antigens is very complex and takes into

account the shape and the charge of the binding sites of the connectingmoieties. Antibodies can be divided in two groups: monoclonal and

polyclonal. Monoclonal antibodies have a recognition site that binds to one

specific part of the antigen, while the polyclonal consist of a mixture of

antibodies that bind to different parts of one target molecule species. The

antigens are proteins, polysaccharides or organic substances, such as drugsand their metabolites.

Enzymes are also proteins that decrease the reaction energy barrier and

allow (bio)chemical reactions to proceed under mild conditions. In other

words they are biocatalysts. Enzymes are used as recognition elements in

sensors where on-line monitoring or long-term measurements are needed.

The redox reaction is electronically detected. The most well-known exampleis the glucose oxidase (GO) enzyme, immobilized for biosensing [Situ,1998; Kuma, 2000]. Another application of enzymes in bioaffmity assays is

their use as tracers (e.g. in enzyme linked immunosorbent assay, ELISA),where surface-bound target molecules react with enzyme-conjugatedantibodies. The surface containing the specifically bound enzyme is rinsed

and brought in contact with enzyme-substrate. The converted amount of

substrate is then detected by an indicator. The most prominent tracer enzymeis horseradish peroxidase.

19

II Theory

The specificity of the binding of both the antibodies and the enzymes

with their binding partners can be visualized by the key-and-lock analogy.

The sensitivity (detection limit) is proportional to the affinity of a

recognition molecule for its target molecule. The affinity constant is the

product of the association (building of antigen-antibody complex) and

dissociation constants.

1.2 Immobilization of Recognition Elements

The immobilization of antibodies and enzymes [Situ, 1998] is not easy,

since the binding or reacting part of the recognition molecules must not be

negatively affected during the immobilization process and must be available

for the target molecules in a similar manner as in the non-immobilized form

[Naka,1996, Huan,1996; Alar,1990; Vikh,1998] (specific and/or non¬

specific binding of recognition molecules). The substrate, on which the

recognition molecules should be immobilized, must also not adsorb the

target molecule itself [Lin, 1991; Ahlu,1991] (specific and/or non-specificbinding of target molecules). In addition, the immobilization of Abs can

result in a deformation of the recognition molecules due to dense packing or

confining of flexibility (steric hindrance) or may lead to denaturation of the

recognition molecules resulting in a decreased activity [Lu,1996]

(Scheme 1). On the other hand, it is important to form a stable binding of the

biomolecules at the transducer surface, in order to make sure that the

recognition molecules do not desorb during the analytical assays.

20

II Theory

immobilization of capture molecules steric hindrance of capture molecules

v\x

\ V

s> tow sensitivity => reduced accessibility, low sensitivity

specific bmdinq non specific binding

adsorption of target molecules denaturation of capture molecules

\ »

=> bad SB/NSB => reduced activity, low sensitivity

specific binding non-specific bmdinq

Scheme 1 Schematic view of problems and limitations m the context of

the immobilization of recognition molecules and detection of

target molecules

1.2.1 Adsorption

Antibodies and enzymes easily adsorb on hydrophobic surfaces byhydrophobic-hydrophobic interactions Ihe biochemical activity on such

coated surfaces ma) be reduced by non-optimum orientation of the

recognition molecules (Scheme 1) The most prominent surface modification

method is silamzation from the gas phase oi from silane solutions

Hydrophobic surfaces can be produced b\ self-assembly processes (self-assembled monolayer. SAM) or b\ the Langmuir-Blodgett technique (LB)[Maoz, 1984] SAMs have the advantage of producing well-organized and

reproducible surfaces, which is essential foi high-perlormancc sensor

surfaces

21

II Theory

1.2.2 Covalent Binding

Covalent binding to the sensor surface is a technique that offers the

chance to control the orientation and functionality of the recognitionmolecules [Will, 1994; Lösc,1998]. However, functionality and orientation

depend on the recognition element of interest and might show considerable

variation from protein to protein. The reaction of primary amino-groups of

the Ab with succinimidyl-terminated surface adlayers [Pate, 1997] or the

addition reaction of thiol groups to maleimidyl-terminated molecules are the

most successful methods [Shri,1997] to date. Both immobilization methods

are classical chemical reactions.

1.2.3 Binding through Biotin-Streptavidin-Link

Biotin binds to streptavidin with very fast adsorption kinetics and highaffinity (affinity constant Kaff in solution = 10bM"*). The big advantage of

this method is the specificity of the binding and the mild reaction conditions

(in comparison to covalent immobilization) under which the adsorption takes

place [Spin, 1993]. The reaction molecules can simply be dissolved in a

buffer (e.g. phosphate buffer saline, PBS) and brought into contact with the

fiinctionalized transducer surface. For this method one is free to choose an

Ab or an enzyme that is conjugated with streptavidin to link it to a

biotinylated surface or vice versa. Furthermore, the blotin-streptavidinbinding is stable over a wide range of pH-values and even in aqueous

solutions containing organic solvents.

Biotin is a small molecule (Mr = 244 g/mol) that can be found in eggsand in living cells.

Streptavidin is a protein with a molecular weight of 60 - 75 kDa

(kg/mol) and can be isolated from Streptomyces avidinii. One molecule

possesses four binding sites for biotin.

22

II Theory

1.3 Transducer Technique

The transducer transforms the binding of target molecules into a

measurable signal. For this transformation, different physical or chemical

principles can be used [Kric,1993]. Biosensors can be divided into different

classes based on transducing mechanisms such as optical, electronical,thermal or piezoelectrical methods.

Biosensors are prominently based on immobilized antibodies or

enzymes. In the case of immunoassay, analytes (antigens) bind to the

surface-immobilized antibodies (Ab) with high specificity to build antibody-antigen complexes. The stability of such complexes depends on the affinityconstant and varies from Ab to Ab and from antigen to antigen. The

quantification of target molecules is performed by direct measurement of

adsorbed antigens (e.g. detection of mass adsorption), of tracers on them or

by addition of secondary labeled antibodies(e.g. fluorescence detection).

The molecular recognition of enzymes is similar to that of the

immunoassay system. Adsorbed target molecules (substrate) react duringcomplexation with the enzymes (e.g redox reaction). The activity of the

enzymes depends on the adsorption-, reaction- and desorption-kinetics, and

may be considerably influenced by environmental parameters, such as

temperature, pH or immobilization (such as for antibody-antigen systems).The quantification of target molecules is often performed by electronic

determination (amperometry) of the redox reaction or by optical detection of

decomposed substrate or by further reaction thereofwith an indicator.

The combination of both the immunoassay and the enzymatic reaction

is realized in the case of enzyme linked immunosorbed assay (ELISA),where the tracer is chosen in the form of an enzyme-conjugated antibodythat binds to immobilized antibody-antigen complexes. The concentration of

bound enzyme-antibody conjugate is then determined after rinsing (removalof non-bound tracer) and subsequent application of substrate. The amount of

decomposed substrate is proportional to the analyte concentration and is

measured by means of photometry.

23

TT Theory

1.3.1 Detection Methods

Optical methods (absorption, luminescence and index of refraction),surface-plasmon resonance (SPR) [John, 1991; VanG,1991], electrochemical

methods and piezo-electrical methods (frequency of piezo crystals) are the

most prominent techniques for biosensor transducers.

1.4 Specific and Non-Specific Binding

A real problem in the biosensor technique (as well as in other analyticaltechniques) for the detection of proteins or nucleic acids is the non-specificadsorption of target molecules at the surface causing a loss of sensitivity of

the system. The non-specific binding of tracer (= fluorescent-labelled

analyte or marker) increases the background signal and reduces the

sensitivity and the detectability range of the system as well. To increase the

concentration limit and to decrease the detection limit at the same time, it is

necessary to design a densely packed layer of recognition molecules on a

protein-repulsive interface. In the case of hydrophobically adsorbed Abs, the

non-occupied space between the recognition molecules can be passivatedwith albumin (e.g. BSA), which blocks the area for non-specific adsorptionof target and/or tracer molecules.

In the case of covalently bound Abs or binding via biotin-streptavidin,the backbone can be chosen as a suitable polymer structure, which adsorbs a

lot of water (e.g. poly(ethylene glycol) copolymers). These water-like

surfaces have excellent protein repulsion effects and give enough flexibilityto the Abs to allow them to keep their original shape for a maximum

activity.

24

II Theory

2 Optical Sensors

Evanescent field waveguide optical sensors can be divided into two

groups: planar sensors and fibre-optic sensors [Abel, 1995; Duve,1995;

Nath,1997]. Both are based on detection using the evanescent field

technique (details see chapter II 2.2).

The simplest detection method with waveguide optical sensors is to

measure the light absorption by the analyte: Monochromatic light is sent

through the waveguide. The absorption of the light at the same wavelength is

measured at the end of the sensor and is proportional to the quantity of

adsorbed analyte.

If higher sensitivity is desired, the fluorescence technique can be used:

Again, monochromatic light is sent through the waveguiding layer and

fluorophores at the sensor surface are excited. The measured light intensity(fluorescence light wavelength) is proportional to the analyte concentration.

Usually the target molecules are not fluorescent and must be labelled for

detection. The monitoring of fluorescence extinction (quenching) upon

binding of the analyte is a further method that is used e.g. in the case of

oxygen sensors (e.g. oxygen/Ru-complex). The disadvantage of fluorescent

tracers is the bleaching of the fluorophore during the illumination, as well as

the binding characteristics of the tracer molecules, which might change due

to the labeling. This may cause a problem in lenghty investigations.

An already-established method for detecting electro-chemilumines-

cence is the use of ruthenium-bipyridin complexes as tracers for the

detection of nucleic acids.

The change of the refractive index, due to the adsorption of targetmolecules at the sensor surface, is a method for the investigation of small

molecules, macromolecules as well as living cells. The disadvantage when

detecting the change of the refractive index is that the adsorption of anysubstance or small variations in the solvent composition may influence the

result.

Optical sensors have the advantage that they are much less affected byelectromagnetic fields from the environment, e.g. in an emergency room, in

comparison to electrochemical methods. The investigation of living cells and

25

II Theory

the study of the biocompatibility of materials in real time is possible. The

fact that optical sensors are easily miniaturizable allows for the possibility of

on-line quantification of substances in human and veterinary medicine.

Optical sensors are not limited to conductive liquids; investigations in gas or

even in organic solvents are, in principle, possible. Variations in the bulk

solution composition do not influence the signal directly, except in the case

of systems that are based on the measurement of the refractive index.

2.1 Principle of the Optical Waveguide Technique

Snell's law describes the refraction of an incident light beam at the

interface of two transparent materials (equation 1 ). If the beam penetratesthe material interface from the optically denser material into the opticallyless dense material, the beam is refracted along a larger angle with respect to

the surface normal (Scheme 2). If the refraction angle (oci) corresponds to

the surface parallel (90°), the incident beam angle is called the critical angle.Above the critical angle, the light is totally reflected. If a high refractive

material is in between two low refractive materials, a light beam below the

critical angle is repetitively (or internally) totally reflected (Scheme 3). Such

materials are called waveguide layers. Light can be launched into such a

waveguide layer by means of an incoupling grating. If the thickness of such

a waveguide layer is smaller than the wavelength of the incoupled light, the

wave-guide behavior can not be described by the simple internal total

reflection model anymore. Nevertheless the light is still transported throughsuch high refractive interfaces (Scheme 4).

sin a, /?,.

.

1X= —^

(equation 1)sin a, n{

26

II Theory

b)

Scheme 2: Possibilities of behavior of a light beam penetrating the

interface of two materials with different optical densities,

a) refraction, b) critical angle and c) reflection.

Scheme 3: Internal total reflection

n2

Scheme 4; Total internal reflection (a) and waveguide-mode (b) by lightincoupling into a waveguide layer, (n2<ni>nO

li Theory

2.2 The Evanescent Field

In case of total internal reflection, the energy of the electromagneticfield is not zero in the material of lower refractive index, but penetrates into

this material for a certain distance. Inside the waveguiding layer, a standingwave forms and outside the waveguide (against the hulk solution) the energy

of the light decreases exponentially. This field of light energy in the low

refractive material is called the evanescentfield. (Scheme 5)

i

a

<

ô

I

dp

r

\ /\ 1 n2

^\ til

r

Scheme 5: The depth of the evanescent field (dp) depends on the ratio of

the refractive indices of the corresponding materials (nL and iii).the light beam angle (a) and the wavelength (k) of the lightsource (equation 2).

28

II Theory

d = =, (equation 2)2/r-«2 -a/w^ -snr a-1

with n,-111

rel*

1U

The energy of the evanescent field at a certain distance from the surface

can be calculated using equation 2 (equation 3),

E(8) = Es e" p(equation 3)

where 8 is the distance from the surface, E(8) is the amplitude of the

electromagnetic field at the distance 8 and Es is the amplitude at the surface.

The angle a depends on the grating periodicity and the light wavelengthX (equation 2). The intensity maximum of the evanescent field is reached at

the critical angle.

The evanescent field vector is parallel to the surface and has no

component parallel to the surface normal. For an ideal, flat waveguidesurface (no diffraction) there is no loss of energy and the evanescent field

thickness is finite (dp). Surface-adsorbed substances that are able to absorb

light with the corresponding energy are excited. The excitation leads to a

loss of energy (light attenuation) which is used as an analytical technique(attenuated total reflection, ATR). The relaxation (fluorescence) of such

molecules can be measured as well.

29

11 Theory

2.3 Analytical Methods with Optical Waveguide

2.3.1 Change of Refraction Index [Krög.1998]

The incoupling angle of a light beam in a vvaveguideing layer dependson the grating periodicity, the ratio of iVm and the wavelength of the light,as mentioned above. If the light wavelength is kept constant (e.g. using laser

light), the only parameter that varies is the refractive index at the surface

(n2). The adsorption of molecules or cells at the waveguide surface provokesa shift of the effective refractive index. This shift depends on the differential

molar refractive index (dn/dc) of the adsorbents. Using a grating coupler

system, the adsorption kinetics can be studied in real time. (Scheme 6)

3\\

•L_L_

*I I 1 L|

Scheme 6: Example of a grating-coupler apparatus; A laser beam (1) from

a TleNe laser source (2) is de focused by a cylindrical lens. The

waveguide chip (4) is irradiated b> the defocused beam that is

reflected at the chip surface except for the small part of the

beam having the exact mcoupling angle of the waveguide chip(5). Mass adsorption at the waveguide chip surface results in a

change of effective refraction index and a shift of the incouplingangle (or missing reflection). The reflected light intensity is

measured by a CCD chip (6) and visualized on a screen (7).

30

II Theory

2.3.2 Fluorescence Measurement

The evanescent field of a transparent optical waveguide sensor with an

extended waveguide mode can be used for the excitation of fluorescent-

labelled target molecules or tracers. The fluorescence wavelength (emission)is different (larger) in comparison to the incident wavelength (excitation).

Therefore, it can be measured using monochromatic filters. Two different

detection methods are presented in Scheme 7. A part of the fluorescence is

back-coupled into the wave-guide layer and can be measured after out-

coupling on a second grating. Another part is transmitted through the planarwaveguide (PWü) chip and can be measured by a detector, which is placedbelow the PWG-chip. (Scheme 7)

Scheme 7;

2)3)

4)

\

** V 1 *

C-~~~___ J^>

pen .:5j (a)

(b)

Fluorescence detection system of a planar waveguide, PWG:

Laser light (1) is coupled into the waveguide layer by a grating(2). The evanescent field excites adsorbed fluorescent-labelled

target molecules or tracers (3). The fluorescence is partly back-

coupled into the waveguide layer and out-coupled by a second

grating (4). The out-coupled light is filtered through a

monochromatic filter and quantified by a photo-detector (a).Another part of the fluorescence is transmitted through the

PWG-chip, focussed, filtered trough a monochromatic filter and

quantified by a photo-detector (b).

31

Ill Materials and Methods

III Materials and Methods

1 Substrate Material and Chemical Products

1.1 Substrate

The substrate for optical biosensor chips has to fulfil special demands:

A high refractive index of the waveguide layer is needed for a total internal

reflection. A high transparency, very low roughness of the surface and

amorphous (to semi-crystalline) state of the material is needed to get a

minimum of light energy loss within the waveguide layer.

Most of the surface modifications were studied on different metal oxide

films (tantalum pentoxide, niobium oxide and titanium oxide) as well as on

metal surfaces (titanium and aluminum). Particular investigations were

performed using additional types of metal oxide substrates.

1.1.1 Metal Oxides

Tantalum pentoxide (Ta2Os), niobium oxide (Nb205) and titanium oxide

(TiC>2), have a semi-crystalline to amorphous structure. The surfaces have a

roughness in the sub-nanometer range (Ta205 and ND2O5) or in the

nanometer range (TiC^), respectively [Xiao, 1999].

The low-temperature form of Ta20s (L- Ta20s, below 1360°C) is

characterized by chains built from octahedral (and partly bipyramidal)coordination groups [Well, 1991; Step, 1971]. These chains are linked to each

other by edge and vertex sharing to form the 3D network and satisfy the

overall stoichiometry Ta02.5. The coordination number in the bulk structure

of Ta205 is 6 and 7. The relative proportion of the different structural

elements, however, varies depending on the synthesis conditions; Ta205 (as

32


well as other M205 oxides such as Nb205) shows a pronounced tendency to

polymorphism orpolytypism [Cott,1988].

Niobium oxide shows the formation of blocks of different size and

different coordination between these blocks. In Nb25062 [Roth, 1965] (3 x 4)-blocks are formed with vortex-shared octahedral coordination of the metal

atoms within the blocks and edge-sharing between the blocks. In Nb20s

[Gate, 1964] blocks of (3 x 5) octahedra are joined at one level to form

infinite planar slabs, and these slabs are further linked by (3 x 4)-blocks.Tetrahedral holes, some of which are occupied by metal atoms, are built in

between these blocks. In Nb2Os 27 Nb atoms in the unit cell are in positionsof octahedra coordination and 1 Nb occupies a tetrahedral hole. The relative

proportion of the different structural elements, however, varies depending on

the synthesis conditions.

Ti02 forms monocrystals in an assembly of edge-shared octahedra

[Well, 1984] (Anatase) and in tetragonal structure (Rutile) consisting of

chains of Ti06 octahedra in which each octahedron shares a pair of oppositeedges, which are further linked by shared vertices to form 3D structure of

6:3 coordination [Well, 1984].

150 nm Ta205 or Nb205 coating was deposited on Corning 7059 glassby physical vapor deposition (sub-nanometer average roughness, from

Balzers AG, Liechtenstein). The Ti02 films (20 nm) were deposited by a

sputter-coating process on commercial glass (PSI, Villigen, CH). The

average Ti02 substrate roughness is in the nm range [Kurr,1997].

AF45 glass chips (8 x 12 mm) were coated by a 150-nm-thick mixed

titanium-silicon oxide layer (Tio.4Sio.602). For the studies on iron oxide

(Fe203), zirconium oxide (Zr02) and silicon oxide (Si02), an additional layerof-14 nm of the corresponding metal or semi-metal oxide was applied at the

Tio.4Sio602 layer using a magnetron sputtering unit. The samples were

purchased from Microvacuum, Ltd. (Budapest, Hungry).

33


1.1.2 Metals

Titanium metal and aluminum metal surfaces spontaneously form a

natural amorphous oxide film. Therefore, the adsorption mechanism takes

place at the metal oxide in both cases. The thickness of the oxide film, whichacts as protecting layer against corrosion, can be increased by anodizing the

bare metal.

Titanium metal (100 nm) films were deposited by a sputter coatingprocess on silicon wafers (PST, Villigen, CH). Aluminum was received in

electro-polished form (25 V anodization polish) and in high-brilliance-rolledform with a natural oxide layer (Algroup Alusuisse, Neuhausen, CH). The

average metal oxide surface roughness is in the nm range.

1.1.3 GoldAdlayer

Long-chain alkyl thiols, such as octadecyl thiol form highly ordered

self-assembled monolayers (SAMs) on gold. The high stability of these

SAMs is due to the complex binding between the soft ligand and the soft

metal, as well as the Van der Waals interaction between the alkyl chains.

Thiol-SAMs on gold surfaces are well known and reported [Ulma,1990;Poir,1996; Wood, 1996a]. The stable bonding to gold and different available

bifunctional amphiphiles make it an interesting system. Biosensor, based on

gold adlayers can not be applied on the optical waveguide setup, since the

gold layer is opaque. If it were possible to adsorb a transparent gold salt onto

metal oxide surfaces such as Ta205, gold ions could work as an interface

between the thiol group and wave guide layer.

This was the motivation to deposit gold chloride from an aqueoussolution onto the tantalum oxide chips described above. The coating protocoland the SAM-formation tests are described in chapter IV 2.2.1.

34


1.1.4 Planar Waveguide Chips

For the measurements of immunoassay performance, planar wave¬

guides were used, consisting of 1 a->(Vcoated Corning 7059 glass, as

mentioned above A grating for the mcouplmg of HeNe-laser light (633 nm)

having a grating periodicity of 320 nm was etched in the glass support The

glass was covered with the metal oxide, as described m Scheme 8

î20 nm

< »

a

SchemeS Drawing of a planar waveguide on a glass chip with a

submicron grating for lasei-light mcouplmg

35


1.2 Adsorbents

1.2.1 Organic Phosphoric and Phosphonic Acids

Different organic phosphoric and phosphonic acids were used to studyadsorption onto metal oxide surfaces. The monofunctional amphiphilesdodecyl phosphonate (DDPO3), dodecyl phosphate (DDPO4), tetradecylphosphate (TDPO4), hexadecyl phosphate (HDPO4), octadecyl phosphonate(ODPO3), octadecyl phosphate (ODP04) and biphenyl phosphate (BIPH)were compared to the bifunctional amphiphiles 12-hydroxydodecyl phos¬phate (OH-DDPO4) and 12-(N-ethylamino)dodecyl phosphonate (NEt-DDPO3).

The amphiphiles DDPO3, OH-DDP04, NEt-DDP03 and BIPH were

prepared according to the protocol reported by Maege [Maeg,1998]. The

long-chain monofunctional phosphates TDPO4, HDPO4 and ODPO4 were

synthesized by Novartis Pharma AG, according to the protocol by Okamoto

[Okam,1985] (Figure 1).

% P4010 Me3SiOSiMe3toluene

80°C, Ih

-o-

?-o

OSiMe3

n

n ROH

~80oC, 2h"

0

R(> -OH

OSiMe,

11,011 R( -OH

OH

Fig. 1: Alkyl phosphate synthesis according to the method reported byOkamoto. [Okam,1985]

36


ODPO4 is a stable, waxy solid, and was re-crystallized from hot n-

hexane.

]H-NMR spectra (CDC13): 0.83-0.93 (t, 3H, -CH3), 1.18-1.45 (m, 30H,

-(CH2)l5), 1.62-1.76 (quintet, 2H, P(O) CH2- CH2), 3.97-4.13 (quintet, 2H,

P(O) CH2).

Elemental analysis: Calculated: [C] 61.69%, [H] 11.22%, [P] 8.84%.

Analysis: [C] 61.72%, [H] 11.02%, [P] 8.82%

The atomic ratios of H/C, C/P and O/P of 2.13, 18.04 and 4.05 respectively,calculated from these elemental analysis data, are in good agreement with

the values expected for the formal stoichiometry of the compound (2.17,18.00 and 4.00).

4.5 g ODPO3 (waxy yellowish solid in a quality of 98%) was re-

crystallized by dissolving the amphiphile in a mixture of n-heptane/2-propanol (10:1) and slow self-evaporation of the 2-propanol at room

temperature. The precipitate was filtered and washed with n-heptane at room

temperature. 2.32 g of white waxy product could be isolated, correspondingto a yield of 52%.

'H-NMR spectra (CDC13) 6: 0.88 (t, 3H, -CH3), 1.07-1.45 (m, 30H,

-(CH2)]5), 1.56-1.78 (m, 4H, P-CH2-CH2).

A list of all alkyl phosph(on)ates is presented at the end of this chapter(Table 2) and their structures are shown in Figure 2.

37


Fig. 2: Structures of the amphiphiles with their corresponding abbrevia¬

tions and the length of the unfolded molecules: dodecylphosphonate (DDP03), tetradecyl phosphate (TDP04), liexadecylphosphate (HDPO4), octadecyl phosphate (ODPO4), 12-hydroxy-dodecyl phosphate (OH-DDPO4), 12-(N-ethylamino)dodecylphosphonate (NEt-DDPCh) and biphenyl phosphate (BIPH).

38


Ammonium salts of DDP04 (DDP04(NH4)2) and OH-DDPO4 (OH-DDP04(NH4)2) were produced according the protocol below:

DDP04(NH4)2: 2.00 g of DDPO4 were dissolved in 200 ml of 2-pro¬

panol (UVASOL, Merck), heated up and refluxed. 6 ml of ammonia (25% in

water, p.a., Merck) were added and the precipitated ammonium salt of

DDP04 was filtered after cooling the reaction mixture with ice water. The

product was washed with ice-cooled 2-propanol and dried at 60° and

10 mbar in a vacuum oven for 20 hours. 1.61 g of white powdery product

(mp: 225° C) could be isolated corresponding to a yield of 71%.

'H-NMR (DMSO or CD3OD): 0.88 ppm (t, 3H, -(CH2)nCH3), 1.28 ppm (m,

18H, -CH2CH2(CH2)9CH3), 1.54 ppm (m, 2H, -CH2CH2(CH2)9CH3), 3.72

ppm (q, 2H, -OCH2CH2(CH2)9CH3), 4.88 ppm (s, 8H, NH4).

The " P-NMR showing one single peak suggests that the product consists of

a uniform alkyl phosphate species.

Elemental analysis: Calculated: [CI 50.87%, [H] 10.67%, [N] 4.94%

[O] 22.59%, [P] 10.93%

Analysis: [C] 50.61%, [H] 10.94%, [N] 4.95%

[O] 22.75%, [P] 10.69%

The measured concentrations in the elemental analysis agree with the

calculated stoichiometry of the di-ammonium salt of DDP04.

OH-DDP04(NH4)2: Hydroxy dodecyl phosphate (OH-DDPO4) was

received from Universität Dresden, Germany (Institut für Makromol.

Chemie u. Textilchemie, Prof. H.-J. Adler). 275 mg of OH-DDPO4 were

dissolved in 20 ml of 2-propanol (UVASOL, Merck) and NH3-gas was

bubbled through the solution for 5 minutes. (NH3 was extracted from a hot

25% ammonia solution in water and transferred into the amphiphile solution

via a glass pipette fixed on a plastic tube.) The precipitated ammonium salt

of OH-DDP04 was separated from the solvent by centrifugation and

removal of the solvent. The product was dried in a flow of dry nitrogen at

room temperature using an evaporation system (EVAPOR). 234 mg of white

39

Til Materials and Methods

powdery product (mp: 165° C) could be isolated, corresponding to a yield of

76%.

lH-NMR (DMSO or CD.OD): 1.24 ppm (m, 16H, -CH2CH2(CH2)8CH2-CH2OH), 1.38 ppm (m, 2H, -CH2CH2(CH2)8CH2CH2OH), 1.44 ppm (m, 2H,

-CH2CH2(CH2)]0OH), 3.35 ppm (t, 2H, -(CH2)nCH2OH), 3.57 ppm (q, 2H,

-OCH2CH2(CH2)10OH), 5.2 ppm (s, 8H, NH4).

Elemental analysis: Calculated: [C] 45.6%, [H] 10.5%, [N] 8.9%

Analysis: [C] 44.1%, [H] 9.6%, [N] 5.7%

The 'H-NMR results of the obtained product suggest that the substance

is pure. The somewhat low nitrogen content found experimentally(elemental analysis) is likely to reflect a partial loss of ammonia and partialformation of the mono-ammonium salt. To be consistent with

DDP04(NH4)2, we abbreviate the ammonium salt of hydroxy-dodecylphosphate with OH-DDP04(NH4)2.

XPS spectra were performed from ODP04 as an example of the

alkylphosph(on)ate amphiphiles:

The detailed spectra of Ci„ Oh and P2p obtained on bulk ODP04powder (free acid) are shown in Figure 3. The Cis signal of the ODP04

powder is asymmetric, containing a large contribution at 285.0 eV and small

one at 286.8 ± 0.2 eV. The first component is assigned to the carbon of the

aliphatic chain, the second to the carbon covalently bound to one oxygen of

the phosphate group (C-O-P). Two Gaussian-Lorentzian curves have been

used in the curve fitting routine applied to the 0]s signal, which also has

been resolved into two signals: one at 532.1 ±0.1 eV and the other at 533.6 ±

0.1 eV. The assignments have been carried out taking only initial chemical

state effects into account and on the basis of literature data on sodium

phosphate glasses [Gres,1979],

40

Ill Materials and \lethods

"3

< t-

_,

Rindm« 1 net<n (e\ )

x.

\ M V

Rinclma l neun (eV)

_j

OK I •'p

C

Bindma \ notuv (o\ )

'j

Binding I nenn (cV)

tig 3 XPS survey spectrum and CV, Ok, P?p high-resolution spectraafter background subtraction and curve fitting of ODPC)4 powderpressed onto indium foil I he difference between the original and

curve-fitted spectra is also shown

The Pip signal is a doublet with P>pl/2 and P2P3/2 components their

theoretical energ> separation (0 9 eV) and the intensity ratio of 0 5 have

been fixed for the curve-fitting anal} ses of all 1% spectra I he bindingenergy of the P>p3/2 component was at 134 7 i 0 1 eV The results of

qualitative and quantitative anal) sis are summarized in Table 1

41


Table 1 : XPS binding energies (Eb ± 0.1 eV), Full Width at Half-Maximum

height (FWHM) and experimental results of quantitative analysisof bulk ODPO4 (free acid). Comparison to atomic concentrations

('calc.'), calculated from the formal stoichiometry of C18H37PO4.Photoelectron take-off angle 9 = 45°

Cls(l) Cis(2) 0,s(l) 0ls(2) P2p

Assignments CH2, CH3 C-O-P p=o P-OH,

P-OC

C-0-P(0)(OFI)2

EB (eV) 285.0 286.8 532.1 533.6 134.7

FWHM 1.4 1.4 1.8 1.9 1.7

Atomic ratio 17: 1 1 :3 -

At. % cale.l

78.3 17.4 4.3

At. % exper." 81.6 ± 1 14.5 ±1 3.9 ±0.5

Atomic ratio O/P exp. 3.7 ±0.6

At. ratio O/P (ODPO4)3 4.05

expected concentration based on formal stoichiometry

(a) based on the use of photoelectron ionization cross sections [Scof, 19761 an(l

attenuation length [Seah,l 979] values from the literature

based on elemental analysis of ODPO4 powder

The carbon content (81.6 % at.) determined from XPS, is slightlyhigher than expected from elemental analysis (C = 78.3% at). This is

probably due to a slight hydrocarbon surface contamination.

42


The interpretation and quantitative results of the XPS spectra obtained

from bulk ODPO4 are straightforward. The assignment of the curve-fitted

Cis at binding energies of 285.0 and 286.8 eV to hydrocarbon and C-O-P,respectively, and of Ois at 532.1 and 533.6 eV to P=0 (O type 1) and P-O-R

(O type 2), respectively, are in agreement with expectations based on

published reference data [Gres,1979; Moid, 1992]. The experimentallydetermined O/P atomic ratio of 3.7 is consistent with the stoichiometry of

the phosphate functional group (Table 1 ).

Octadecyl thiol was used for special investigations or tests that were not

relevant for the sensor application. Nevertheless, these studies will be

discussed in the context of a comparative understanding of processes,

material behavior and/or other applications.

43


Table 2: Amphiphiles with their abbreviation used in this thesis and their

suppliers

Amphiphile Abbreviation Supplier

Dodecyl phosphonate DDPO3 University ofDresden

(Germany)

Dodecyl phosphate DDPO4 Alfa Aesar

(Germany)

Dodecyl phosphateammonium salt

DDP04(NH4)2 Precipitation ofDDPO4 from

Alfa Aesar

Tetradecyl phosphate TDPO4 Novartis Pharma AG, Basel,Switzerland

Hexadecyl phosphate HDPO4 Novartis Pharma AG, Basel,Switzerland

Octadecyl phosphonate ODPO3 Re-crist, of ODPO3 from

Alfa Aesar (Germany)

Octadecyl phosphate ODPO4 University ofDresden

(Germany)

Octadecyl phosphate ODP04 Novartis Pharma AG, Basel,Switzerland

Biphenyl phosphate BIPH University ofDresden

(Germany)

12-hydroxydodecylphosphate

OH-DDP04 University ofDresden

(Germany)

12-hydroxydodecylphosphate ammonium salt

OH-

DDP04(NH4)2Precipitation of OH-DDPO4from University of Dresden

12-(N-ethylamino)dodecy1

phosphonateNEt-DDP03 University ofDresden

(Germany)

Octadecane thiol ODT Fluka

44


1.2.2 Poly-(L)-Lysine-Poly(EthyIene Glycol) (PLL-PEG)

Different derivatives of pofy-(L)-lysine-poly(ethylene glycol) (PLL-PEG) were synthesized. Non-functionalized PLL-PEGs were used for

protein-resistance studies (non-specific binding). Biotinylated poly(ethyleneglycol) (PEGbiotin) was used to synthesize functionalized PLL-PEG

derivatives (PLL-PEG/PEGbiotin). The specific binding of streptavidin-conjugated enzymes as well as the specific immobilization of biotin-

conjugated antibodies were studied on PLL-PEG/PEGbiotin coated surfaces.

The abbreviations for the non-functionalized copolymers describe the

ratio of grafted poly(ethylene glycol) chains (PEG) per poly-(L)-lysine(PLL). PLL-g[2]-PEG for example means that the ratio of lysine-units to

grafted PEG-chains is 2:1. In this case every second lysine-unit of the PLL-

backbone is coupled to a PEG-molecule. The molecular weight of each

copolymer component is described in parentheses. If every second lysine-unit of a 20 kilo-Dalton (kD) PLL is grafted with a 2 kD PEG-molecule,then the abbreviation for the derivative would be PLL(20)-g[2]-PEG(2).

For the description of functionalized PLL-PEG derivatives, the

presence of biotinylated PEG (PEGbiotin) is indicated by adding PEGbiotin

to the abbreviation of the corresponding non-functionalized PLL-PEG (e.g.PLL-PEG/PEGbiotin). The ratio of grafted PEG/PEGbiotin is described byadding the relative amount (%) of PEGbiotin (with respect to the total

amount of grafted ethylene glycol chains) at the end of the compoundabbreviation. If 10 % of the PEG-chains in the previously described non-

functionalized derivative are replaced by biotinylated PEG of a 3.4 kD

molecular weight, the abbreviation would be the following: PLL(20)-g[2]~PEG(2)/PEG-biotin(3.4) 10%.

The synthesized PLL-PEG derivatives and the amount of educts that

were used are listed in Table 3.

45


1.2.2.1 Synthesis of PLL-g-PEG Derivatives

The structural parameters of the copolymer had to be optimized in order

to yield polymers with efficient protein- [Glas, 1996; Elbe, 1996; Hard, 1998]and DNA- [Thie,1997] repellent characteristics. Among these parameters,the grafting density of the PEG chains to the PLL backbone is of particularimportance. The latter also depends on the molecular weight (i.e. length) of

the PEG chains used. Polymers that exhibit high protein resistance are often

in the so-called "brash regime"; the grafting density is high enough, so that

the PEG chains start to fully overlap and stretch away from the surface, thus

leading to a large repulsive forces and preventing proteins from penetratinginto the brush. Using a simplified approach in a good solvent, such a brush

regime is achieved when the average distance D between the attachment

moieties of the coil (in our case PEG grafted onto poly-L-lysine backbone)on the surface is smaller than the radius of gyration (Rg) of the coil

[Halp,1992]. Furthermore, a recent study by Sofia et al. [Sofi,1998] showed

that protein repulsion was maximized on silicon surfaces with the highestPEG grafting densities, for which the chain spacing was approximately equalto the typical chain dimension, Rg, of the PEG molecule.

A scheme of the functionahzed copolymer synthesized in this study is

shown in Figure 4. The polymer backbone consists of poly-L-Lysine MW 20

kD, which was chosen rather than the high MW 350 kD [Kena,2000] to

achieve higher packing density of the comb/brush polymer on the flat

sensing surfaces. The protein-repellent matrix is built with methoxyterminated PEG of MW 2000, and the biotin-functionalized PEG tether

consists of a PEG chain of higher molecular weight (MW 3400), to improveaccessibility to the functional sites.

46

1TI Materials and Methods

+

PI I

4

SPA-PI G

IUI'

ifo

N—-btotm

70

NHS-PLG-biotm

\

hi 01 it

1 m

0

\

_J 14

PI F -g-Pf G

Schematic dtawing ot the svnthesis of a PI 1 -g-PFG/Pl Gbiolm

copolymer derivative I he red pan of the molecule tepiesents the

poly-(I )-l\sme backbone, the green pait stands loi PI (j and the

blue part tor the PPGbiotm side chain

47

TTT Materials and Methods

The PEG as well as the PEGbiotin chains (educts) are functionalized

with hydroxy succinimidyl ester for reaction with the primary amine side

chain of the lysine unit. Since this amino side chain also serves as the

anchoring point when the PEE-g-PEG polymer is self-assembled onto a

negatively charged surface, an optimum number of the lysine side chains

must be grafted by PEG to insure both efficient protein resistance and

surface anchorage.

The protein resistance characteristics of the PEL-g-PEG polymeressentially depends on the grafting density of the PEG side chains

[Kena,2000]. It has been shown, that a lysine : PEG grafting ratio of 3.5 (1of 3.5 lysine units is derived by PEG) results in a highly protein-resistantand well-adsorbing copolymer (see chapter IV 2.E4). Three PLL-g-PEGderivatives having different grafting ratios without any biotinylated PEGchains and several derivatives with different PEG-biotin concentrations were

synthesized, keeping the lysine : PEG(total) grafting ratio of 3.5 constant.

Synthesis of PLL(20)-g-PEG(2):

Three non-functionalized PEL(20)-g-PEG(2) derivatives were

synthesized with different lysine unit to PEG chain ratios (Lys:PEG) such as

2.0, 3.5 and 5.0, in order to obtain grafted polymers for which the PEG

interchain spacing D is respectively smaller, equivalent to and greater than

the RMS radius of gyration of the PEG (Mw 2000). Synthesis was based on

the protocol described by Elbert et al. [Elbe, 1998]. Poly(L-lysine) (PLL) was

dissolved in 50 mM sodium tetraborate buffer, pH 8.5 at a concentration of

1 g of PEL per 12.5 ml of buffer and the PLL solution was sterilized byfiltration (0.22 |im Durapore Millex, Sigma-CH). The appropriate stoichio¬

metric amount of solid succinimidyl derivative of PEG propionic acid (SPA-PEG) (see Table 3) required to yield the desired grafting ratio of Lys:PEGwas added to the vigorously stirred PLL solution, and the reaction was

allowed to proceed for 6 hours at room temperature. The reaction productswere then dialyzed (Spectra/Por, MWCO 6-8000, Spectrum, Socochim, CH)against phosphate buffer saline (PBS) for 24 hrs and subsequently againstdeionized water for 24 hrs. The dialyzed solution was lyophilyzed for48 hours and stored under nitrogen at -25°C.

48


Synthesis of biotin-functionalized PLL-g-PEG derivatives:

Synthesis was performed as described above. Several biotin-functiona¬

lized PLL-g-PEG copolymers with a Lys:PEG ratio of 3.5 and different

amounts of biotin (i.e. 1%, 10% and 50% of the PEG chains are

biotinylated) were synthesized. The desired stoichiometric amounts of solid

N-hydroxy-succinimidyl PEG-biotin derivative (NHS-PEG-biotin) and

SPA-PEG (see Table 3) were added to the vigorously stirred PLL solution

and reaction was allowed to proceed for 6 hours at room temperature. In the

case of PLL-g-PEG/PEGbiotin 5%, 10%, 15% and 20%, the NHS-PEG-

biotin was added first and allowed to react for 1 hour before addition of the

SPA-PEG.

^-NMR spectra were measured from the PLL-g-PEG and PLL-g-PEG/PEGbiotin derivatives. The peaks were assigned to the correspondingprotons according to the drawing of the molecule in Figure 5. The generalinterpretation of the lH-NMR spectra of the PLL(20)-g[3.5]-PEG(2)/PEG-biotin(3.4) derivatives is shown below.

49


Table 3: PLL quantity and stoichiometric amounts of solid NHS-PEG-

biotin and SPA-PEG.

Polymer abbreviation PLL'HBr

(MW20'000)

SPA-PEG

(MW 1889)

NHS-PEG-

biotin

(MW 3445)

PLL(20)-g[2.0]-PEG(2) 27.2 mg

(1.36 umol)

123 mg

(65.1 (imol)

-

PLL(20)-g[3.5]-PEG(2) 83.6 mg

(4.18 umol)215.7mg(114 (imol)

-

PLL(20)-g[5.0]-PEG(2) 106.8 mg

(5.34 (imol)

193.2 mg

(102 umol)

-

PLL(20)-g[3.5]-

PEG(2)/PEGbiotin(3.4) 1%

41.6 mg

(2.08 (imol)

107 mg

(56.6 (imol)

1.96 mg

(0.57 umol)

PLL(20)-g[3.5]-

PEG(2)/PEGbiotin(3.4)5%

30.0 mg

(1.50 (imol)

77.8 mg

(41.2 (imol)

7.0 mg

(2.0 (imol)

PLL(20)-g[3.5]-

PEG(2)/PEGbiotin(3.4)10%39.5 mg

(1.98 |imol)

92 mg

(48.7 (imol)

18.6 mg

(5.4 (imol)

PLL(20)-g[3.5]-

PEG(2)/PEGbiotin(3.4) 15%102.4 mg

(5.12 mnol)

225.2 mg

(119.2 mmol)

72.6 mg

(2.2 (.imol)

PLL(20)-g[3.5]-


30.0 mg

(1.50 (imol)

65.5 mg

(34.7 (imol)

27.9 mg

(8.1 (imol)

PLL(20)-g[3.5]-


32.3 mg

(1.62 umol)

42 mg

(22.2 mnol)

76 mg

(22.1 (imol)

50


_/J H il H II H _\__t_c_(p—NH—c—r—N—c_c—N_i_

(ÇH2)3

b CH2

NH2

(ÇH2)3

h

9

VCH-NH2 e H

Fig. 5: Schematic drawing of the PLL-g-PEG/PEGbiotin molecule. The

letters a - k are assigned to the different 'H-NMR peaks.

51


The peak shifts were compared to NMR spectra of the different

fragments and functional groups from the literature and assigned to the

corresponding protons (indicated in brackets):

'H-NMR (D20): 1.05-1.45 and 1.45-1.85 ppm (m, -CH2-, lysine side chain

[a] and -CH2-, biotin butyl chain [i]), 2.13 ppm (t, -CH2C(0)O, biotin [k]),2.40 ppm (t, -CH2OC(0)NH-, PEG [d']), 2.63 ppm (d, -CH(S)(CH<), biotin

[f]), 2.86 ppm (t, NH2CH2-, free lysine side chain [b]), 3.05 ppm

(t, -NHCH2-, derivatized lysine side chain [b']), 2.95 and 3.19 ppm (m and

q, -SC(HaHb)CH<, biotin [g]), 3.24 ppm (s, CH30-, PEG [h]), 3.55 ppm

(m, -OCH2CH2-, PEG [d]), 4.06 ppm (t, -CH2NC(0)-, PEGbiotin [b"])54.15 ppm (t, >CHC(0)NH-, lysine [c]), 4.28 ppm and 4.47 ppm (2 •

q,

>CHCH<, biotin [e]).

52


1.3 Biochemical Substances

The biochemical products with their abbreviations and applications are

listed in Table 4. In the following only the abbreviations will be used.

Table 4: List of biochemical substances.

Product Abbrevi

-ation

Supplier Application

Albumin (= bovine serum albumin) BSA Sigma Additive in immunoassay-buffer

Albumin-biotin (= bovine-biotin-

amido-caproyl labeled)

(llbiotin/BSA)

BSA-

biotin

Sigma Immobilization process at

sensor surfaces

Biotin conjugate goat anti-Rabbit

IgG antibody

a-RlgG-

biotin

Sigma Recognition molecule for

biosensor

Streptavidin from streptomyces

avidinii

Strepta¬

vidin

Sigma Immobilization process at

sensor surfaces

Streptavidin-horseradish peroxidase streptav-

HRP

Amersham

Pharmacia

Tests with immobilized

enzymes

Rabbit IgG RIgG Sigma Target molecule for

immunoassay

Cy5 labeled goat anti-Rabbit IgG

antibody

Cy5-a-

RlgG

Amersham

Pharmacia

Tracer for immunoassay

Serum, bovine newborn serum Anawa

M , ,

Additive in immunoassaybuffer

53


1.4 Solvents

The most important solvents and their applications are listed in Table 5.

The quality of these substances is very often essential for the performance of

the resulting products.

Table 5: List of solvents and buffers.

Product Quality Supplier application

2-propanol UVASOL Merck Amphiphile solvent

and cleaning process

Citrate buffer

(=0.1 M sodium acetate in water, pH

adjusted to 6.0 with solid citric acid)

Enzymatic tests with

horseradish peroxidase

ethanol p.a.' Fluka Amphiphile solvent

ethyl acetate LiChrosolv Merck Amphiphile solvent

HEPES buffer

(= 10 mM 4-(2-hydroxyethyl)-

piperazine-1 -ethanesulfonic acid.

adjusted to pH 7.4 with 1 M NaOII)

Fluka Enzymatic tests with


methanol p.a.' Fluka Amphiphile solvent

n-heptane UVASOL Merck Amphiphile solvent

PBS (Phosphate Buffered Saline)

(= 0.001M KIÏ2PO4, 0.01 M Na2HP04.

0.137 M NaCL 0.0027 M KCl)

Roche

Diagnostics

Immunoassay

water (osmosis) 18.3 MO Milliporc Amphiphile solvent

and cleaning process

pro analysi

54


1.5 Further Substances

Table 6: List of further substances.

Product Quality Supplier application

Ammonia solution 25% in water p.a. Merck Alkyl phosphates precipitation

FluoroLink Cy5 monofunctional dye Amersham

Pharmacia

Fluorescence labeling process

Gold chloride hydrate p.a. Fluka Coating of TJbOs for adsorption

of thiols

Hydrogen peroxide 30% p.a. Fluka Cleaning procedures and

enzymatic tests

PD-10 columns, Sephadex G-25M Pharmacia

Biotech

Fluorescence labeling process

PEG (=N-Hydroxysuccinimidyl ester of

methoxy-poly(ethylene glycol)

proprionic acid, Mw 2000)

ca. 95% Shearwater

Polymers,

Inc.

PLL-PEG synthesis

PEGbiotin (= oc-Biotin-co-

hydroxysuccinimidyl ester of poly-

(ethylene glycol)-carbonate, Mw 3400)

ca. 95% Shearwater

Polymers,

Inc.

PLL-PEG synthesis

PLL (= Poly(L-lysine) hydrobromide,

Mw 20'000)

Sigma PLL-PEG synthesis

Poly-oxyethylene sorbitan monolaurate

(Tween 20)

Sigma-

Ultra

Sigma Additive in assay buffer

Sodium carbonate p.a. Fluka Fluorescence labeling process

Sodium hydrogen carbonate p.a. Fluka Fluorescence labeling process

Sulfuric acid p.a. Fluka Cleaning procedures and

enzymatic tests

TMB

(== 3,3',5,5'-tetramethylbenzidine)

p.a. Merck Enzymatic tests with


55


2 Apparatus

^-NMR spectra were recorded for the synthesized PLL-PEG and PLL-

PEG/BEGbiotin derivatives as well as for some phosphonates using a DRX

400 MHz instrument (Bruker).

The following items of apparatus were used for specific parts of the

work and have therefore been separated into different subchapters.

2.1 Apparatus for Cleaning and Coating of Substrates

The cleaning procedure is, depending on the application of the product,a more or less important step. In the case of optical biosensors it is essential

to start with a very clean substrate. Impurities at the surface would result in

inhomogeneous immobilization of recognition molecules and insufficient

reproducibility. Patterning capture molecules on one chip to get multi-

sample sensors or multi-analyte chips would lead to a much lower

quantitative performance if they would be immobilized onto dirty or

otherwise ill-defined surfaces.

Ultrasonic bath: Ultrasonic treatment is a straightforward technique for

removing particles from the substrate surfaces, for supporting amphiphiledissolution and for periodic cleaning of containers and glassware. Sonication

was in general performed in a Ultrasonic bath (BRANSON 3200) for 15

minutes.

UV-cleaner: An easy method for cleaning substrate surfaces is the

treatment in a UV-cleaner [Saka, 1998] (model 135500, Boekel Ind. Inc., PA,

USA) for 30 minutes. The mercury vapor lamp that is placed close to the

substrate surface transforms ambient air oxygen into ozone. Singlet oxygen,

ozone and/or high energy UV-light irradiation oxidize the traces of organic

56


impurities. However, we believe that significant higher amounts of

impurities at the surface cannot be burned away with this method.

Oxygen plasma cleaner: A similar method is oxygen plasma cleaning.Typical treatment times are a few minutes, It has been shown that repetitivecleaning cycles or cleaning for a longer time is not useful to remove dirt

particles [Bier, 1991] and we believe that an excessively long irradiation mayresult in surface damaging. The apparatus we used for several investigationsis a Harrick Plasma Cleaner/Sterilizer PDC-32G instrument (Ossining,NY,USA).

Coating chambers: For cleaning and/or coating of large numbers of

metal oxide samples we used PTFE containers and for smaller numbers

15-ml pill glasses (space for one or two chips per glass). Containers and

glasses were pre-cleaned in piranha acid for 2 hours (piranha acid is a 1:3

mixture of 30% H2O2 and 98% H2S04), followed by extensive rinsing in

ultra-pure water prior to first use. Further rinsing with portions of 2-propanolwere carried out to remove the last traces of water. During the followinginvestigations the containers were cleaned with 2-propanol (or the solvent of

the applied adsorbent solution) in the ultrasonic bath for 15 minutes before

and after every step.

The standard cleaning method for the cleaning of metal oxide substrate

chips was, in general, sonication in 2-propanol for 15 minutes, blow-dryingwith nitrogen, followed by UV-cleaning for 30 minutes.

57


2.2 Surface Characterization

In this chapter, methods for quantitative surface analysis with an

information depth of a few nanometers are presented. These methods were

used for surface characterization before and after coating with adsorbents

such as phosph(on)ate SAMs.

2.2.1 X-ray Photoelectron Spectroscopy (XPS)

XPS analyses were performed using a PHI 5700 spectrophotometerequipped with a concentric hemispherical analyzer in the standard

configuration (Physical Electronics, Eden Prairie, MN, USA). Spectra were

acquired at a base pressure of 10"9 mbar using a non-monochromatic Al-Kasource operating at 200 W and positioned -13 mm away from the sample.The instrument was run in the minimum-area mode using an aperture of

0.8 mm diameter. The concentric hemispherical analyzer (CHA) was used in

the fixed analyzer transmission mode. Pass energies used for survey scans

and for detailed scans of tantalum Ta4f, carbon Cis, oxygen Oi, and

phosphorus P?p were 187.85 eV and 23.5 eV, respectively. Under these

conditions, the energy resolution (full with at half-maximum height,FWHM) measured on silver Ag]tl5/2 is 2.7 eV and 1.1 eV, respectively.Acquisition times were approximately 5 minutes for survey scans and 9

minutes (total) for high-energy resolution elemental scans. These

experimental conditions were chosen in order to have an adequate signal-to-noise ratio in a minimum time and to limit beam-induced damage. Under

these conditions, sample damage was negligible, even after 90 minutes

X-ray exposure, and reproducible analyzing conditions were obtained on all

samples. In addition, only one sample was introduced into the analyzingchamber at a time.

Angle-resolved XPS (AR-XPS) measurements were conducted at different

take-off angles (detection angles), to obtain depth-dependent information

58


and to determine the thickness of the self-assembled monolayer deposited on

the metal oxide substrates.

Spectra were referenced to the aliphatic hydrocarbon C]s signal at 285.0 eV.

Data were analyzed using a least-squares fit routine following Shirleyiterative background subtraction. Atomic concentrations were calculated

using published ionization cross sections [Scof,1976] and calculated

attenuation length values [Seah,1979], Intensities were also corrected for the

energy dependence of the transmission function. Spectra were fitted usingGaussian-Lorentzian functions.

As a reference, octadecyl phosphoric acid bulk powder pressed onto an

indium foil was analyzed. Both as-received and sputter-cleaned bare

tantalum pentoxide substrates were analyzed as reference substrates.

2.2.2 Time-of-FIight, Secondary-Ion Mass Spectrometry (ToF-SIMS)

Secondary-ion mass spectra were recorded on a PHI 7200 time-of-flightsecondary ion mass spectrometer in the mass range 0 - 1000 m/z. The total

ion dose of the 8 kV Cs+ primary ion beam (diameter: 200 pm) was typically9-1011 ions/cm2, corresponding to a value below the static limit. Time perdata point was 1.25 ns. Due to the low conductivity of the metal oxide/glasssubstrates, intermittent, pulsed electron-beam neutralization had to be used

during measurement of both positive and negative SIMS spectra. Mass

resolution M/AM was typically 5500 in the positive and 2000 in the negativemode (masses 43 and 17, respectively). To calibrate the mass scale, the

whole mass range was first calibrated using a single standard set of low ion

masses followed by the assignment of species in the whole mass range usingthe PHI software TOFPAK. To improve the quality of the calibration at

higher masses, different sets of ion species were used due to the large mass

range analyzed and the different natures of the observed secondary-ionspecies (likely to leave the surface with varying velocities):

59


a) low mass range (< ca. 200):CnHm species (positive ions) and OH", P02\ P03" (negative ions)

b) medium mass range:

MaOb* and MPaObHc* species (both types of ions)

c) high mass range (> ca. 400):MPaObCcHd+ species (positive ions) and M2OaHb" species (negative ions)

The mean calibration deviations from the exact mass of the assigned specieswere always below 50 ppm, in most cases below 20 ppm.

2.23 Contact Angle [Mitt, 1993]

Surface wettability was investigated by measuring the advancing water-

contact-angle (Contact-Angle Measuring System G2/G40 2.05-D, Krüss

GmbH, Hamburg, Germany). The contact-angle measurement (20 volume

pulses of 0.22 pi each with a pulse frequency of 1 second) was performed on

three-to-five different places on each chip, and the average contact-anglevalue was determined. A systematic drawing of the contact-angle-measure¬ment principle is presented in Scheme 9. Surface wettability has been shown

to be a useful marker for estimating the monolayer quality of methyl-terminated adsorbents [Folk, 1995].

60


Scheme 9; Systematic drawing of the contact-angle measurement system;A chip (1) is placed on a sample holder (2). The water drop (3)is brought in contact with the chip surface through the capillary(4). The drop grown by pulsed addition of water and data are

collected in the short pause between two growth steps.

2.2,4 Microdroplet Density

Condensation figures (CTs) enable the imaging of surfaces with local

information about regions having different wettabilities (different interfacial

free energies). The use of CFs as a method for delecting contamination on

homogeneous surfaces was first described by Lord Rayleigh [RaylJ^llJand subsequently used by Mérigoux [Méri, 1937) Water from a humid

atmosphere condenses more readily on contaminated, polar |Fris,1994;Aize, 1998] or angular (rough) regions with a high wettability [Blan.,19%] as

compared to clean, apolar and flat surfaces showing reduced wettability. If a

surface is inhomogeneous in its physico-chemical properties, the dropletpatterns reflect this heterogeneity J Lope, 1993].

61


The use of the condensation technique on TaaOs surfaces, modified byself-assembled monolayers (SAM) and other surface-chemical treatments,have been described in the literature [Hofe,2000; Sigg,2000]. The imageswere used to estimate the quality of these surfaces in terms of the large-scaleand local homogeneity/heterogeneity. The quantification of the dropletdensity allows for the definition of a new surface property called

microdroplet density (|idd), consisting of the number of droplets condensing

per mm .The (idd corresponds to a complex surface property involving

purity, wettability, roughness, homogeneity and charging aspects of a

surface. A correlation between the water contact angle and the [idd could be

found within a particular system. Using the microdroplet density techniquethe wettability of a large surface area can be investigated in a few seconds

by measuring a number of small areas at high magnification. Local

variations can be readily observed.

The (idd apparatus (Scheme 10) consists of a metal sample table placedin a transparent humidity chamber. The metal table can be cooled from

outside, e.g. by using ice water. The analyzed surfaces were imaged by a

CCD-camera (Panasonic, model WV-BP 310/6, Matsushita Communication

Deutschland GmbH, Germany) coupled to a microscope (Zeiss, Carl Zeiss

(Schweiz) AG, Switzerland) having 2.5x and 6.3x objectives. The

quantification of the (idd was performed using video-capture software

(Ulead VideoStudio, Video Capture, Version 2.0, Ulead Systems, Inc.).

62


5

•.y.y.y.y."*]

."./"."".I.Ï'.Ï'.V ".**]

"ï'-ï'-ï'-ï'-'ï'!

2

Scheme 10: Schematic drawing of the microdroplet density apparatus: The

object (1) is placed on a temperature-controlled table (2) within

a transparent box (3). An atmosphere saturated with water vapor

is achieved by the presence of water on the bottom of the box.

Ice water is pumped from outside the box through the sampletable. The growth of condensed water droplets is monitored via

a CCD-camera system (4, 5).

On hydrophobic surfaces it is not essential to take pictures after an

exactly defined time of condensation. Once water begins to condense, the

surface of the formed droplet is of high polarity, and water vapor will readilycondense on the droplets. The condensed droplets continue to grow, but the

number of droplets per unit area remains constant.

The (idd is not only a tool to qualitatively image hydrophobic surfaces;it is also possible to get quantitative or semiquantitative results for a giventype of surface. A large area is measurable in a shorter time, with detailed

information on local (macroscopic to microscopic) properties.

It should be noted that quantification of the surface quality with

contact-angle and u,dd measurements is only possible within a well-defined

system. The techniques cannot be used to compare surfaces with different

roughness and/or different chemistry.

63


2.2.5 Atomic Force Microscopy (AFM)

AFM measurements were performed with a commercial scanning probemicroscope (Nanoscope E, Digital Instruments, Santa Barbara, CA).Measurements of surface topography and lateral force were made

simultaneously by operating the instrument in the contact mode while

scanning the cantilever laterally. As AFM probes, we used sharpened Si3N4Microlevers (Park Scientific, Sunnyvale, CA) with a nominal probe tipradius of 20 nm and a force constant of 0.03 N/m. Only those probes that

provided good quality in the high-resolution imaging of freshly cleaved mica

were employed for the imaging of the samples. The applied load duringscanning was in all cases below 0.5 nN and generally a slight 'pulling' of the

tip was necessary, i.e. applying negative loads, in order to get best

resolution. All measurements were performed in ambient air.

64


2.3 Determination of Adsorption

The determination of adsorbed material can be performed by different

methods. If the system is well known, such as ODPO4 onto TaaOs, one can

control and estimate the adsorption reaction by a simple measurement of the

contact angle (see chapter III 2.2.3). This is only valid because the reaction

starts with a very hydrophilic clean substrate and ends with a very hydro¬phobic monolayer. Another method is the determination of fluorescence of a

tracer molecule. This method will be discussed in chapter III 2.4.2.

For our investigations of adsorption kinetics we had two methods at our

disposal. Both the grating coupler system and the OWLS system are based

on the waveguide technique and the dependence of the effective refractive

index, which is related to the mass adsorption.

2.3.1 Grating Coupler

A schematic drawing of the grating coupler apparatus (GKR 401,Frauenhofer-Institut für Physikalische Messtechnik, Freiburg, Germany) is

shown in chapter II 2.3.1. The method uses the reflection of a defocused

laser light beam that is directed on the grating of a waveguide chip. All the

light is reflected except the part of the beam having the incoupling angle of

the grating. As the incoupling angle depends on the mass adsorption at the

surface, this missing reflection shifts during the adsorption process. The

important advantage of the system is the real-time determination of

adsorption down to the millisecond scale.

65


2.3.2 Optical Waveguide Lightmode Spectroscopy (OWLS)

The technique involves the incoupling of He-Ne laser light into a planarwaveguide (PWG) that allows for the direct online monitoring of

macromoleculc adsorption such as proteins, lipids or polymers [Rams, 1993].The incident laser beam angle (a) is varied by continuous periodic waggingof the PWG-chip placed in a position where the laser beam permanentlyilluminates the PWG-grating. Light intensity is measured at the edge of the

PWG-chip and exhibits a maximum at a wagging angle corresponding to the

laser-beam incoupling angle. As described in the theory section, the

incoupling angle varies with the mass adsorbed at the PWG-surface. Anymass adsorption results in a shift of the measured light intensity peak.(Scheme 11). Shift of the incoupling angle can therefore be correlated to the

mass adsorption and used for adsorption kinetics and surface coveragestudies. The method is highly sensitive (i.e. ~1 ng/cmf) up to a distance of

100 nm (corresponding to the evanescent field depth) above the surface of

the waveguide. Furthermore, a measurement time resolution of 3 seconds

allows for the in situ, real time study of adsorption kinetics.

Areal adsorbed mass density data were calculated from the thickness

and refractive index values derived from the mode equations [Rams, 1993]according to Feijter's formula. A refractive index increment (dn/dc) value of

0.182 cm /g was used for the protein adsorption calculations [Rams, 1993;Rams, 1995], and a value of 0.202 cm7g as determined in a Raleighinterferometer was used for the PLL-g-PEG adsorption calculations. All

OWLS experiments were conducted in a BlOS-t OWLS instrument (ASIAG, Switzerland) using a KalrezR (Dupont, USA) flow-through cell

[Kurr,1997; Kurr,1998]. The flow-though cell was used for studying both

PLL-g-PEG adsorption and protein adsorption. The flow rate and the wall-

shear rate were 1 ml/hr and 0.83 s"1, respectively.

66

MI Materials and Methods

• ; i i

a

/

Scheme 11: Schematic drawing of the optical waveguide lightmodc spectro-

seopv (OWLS) apparatus: A laser beam (1) is directed onto the

grating (2) of a planar wavegide chip (3). The chip is moved in

periodic wagging mode and the permanently measured lightintensity at the edge (4) exhibits maxima at the mcouplmgangle a ( plain line - to, dashed line = tj, striped line ~= t2) (5),The maxima of the light intensity data 1 = f(a) are converted

into a = f(t) corresponding to mass m = f(t) (6),

67


2.4 Determination of Bioactivity

For the performance studies of the different interfaces (phosph(on)atesand PLL-PEG/PEGbiotin derivatives, respectively) between the metal oxide

substrate and the recognition molecule, two different analytical systems were

used: a) determination of enzyme activity and b) fluorescence measurements

of an immunoassay.

2.4.1 Determination of Enzymatic Activity

Streptavidin horseradish peroxidase is a well-known and widely used

model system for determining enzymatic activities.

Quantification of the enzymic activity was performed by measuring the

absorption of the transformed indicator, TMB (3,3'-5,5'-tetramethyl-benzidine), at 450 nm using a photometer (JASCO 7800 UV/Vis Spectro¬photometer).

2.4.2 Fast Optical Biospecific Interaction Analysis (FOB1A)

Fast optical biospecific interaction analysis (FOBIA) is a system that has

been developed at Novartis Pharma AG, Basel, and its performance was

improved at Zeptosens AG, Witterswil (Scheme 12) [Duve,1996].

This combination of evanescent-field excitation and fluorescence quanti¬fication is a very surface-selective (ca. 200 nm information depth towards

bulk solution) method that provide results with low bulk signal and hightracer selectivity.

68


Scheme 12: Schematic drawing of the FOBIA apparatus: The light beam of a

He/Ne laser is directed onto the grating (1) of a sensor-chip. The

light incoupled into the waveguide layer forms an evanescent

field (indicated with the red zone on top of the sensor surface)whose energy decreases exponentially with increasing distance

from the surface. The emission (fluorescence, indicated with

yellow arrows) that arises from adsorbed tracer (yellow stars) and

the excitation that originates from dispersed laser light (pinkarrows) are collected by a lens-set (2). The emission intensity is

measured by a photo-multiplier after an emission filter (3). Asmall fraction of the connected light is reflected by a semi-

transparent mirror (4) and measured by a second photo-multiplierafter passing through an excitation filter (5).

69


The apparatus consists of a He/Ne-laser, emitting monochromatic light of

633 nm wavelength. The laser beam is directed onto the back side (oppositeside of the bioactive waveguide layer surface) of the sensor chip, which is

positioned on a chip holder. The laser light is incoupled in the waveguidelayer as mentioned in chapter II 2.2. The chip holder is connected to highprecision electric motors allowing for the control of the x-, y- and z-position,and for adjusting the incident angle of the laser beam (incoupling angle) to

provide for the maximum incoupling intensity. The fluorescence is measured

by a photo-multiplier after a confocal lens set and an emission filter, which

filters out the excitation wavelength of the laser. A small fraction (typicallyabout 1%) of the collected light of the lens set is reflected by a semi-

transparent mirror and measured by a second photo-multiplier after an

excitation filter, which filters off the fluorescence wavelength. The opticalsetup of the FOBIA apparatus allows one to quantify the fluorescence

(emission) of a tracer and to control the incoupling intensity (excitation) in

parallel. A schematic drawing of the apparatus is shown in Scheme 12.

A flow-cell has been designed and constructed from poly-dimethylsiloxane (PDMS). The so-called counter-flow cell has a fluidic system which

starts from the two opposite sides of the flow cell (inlets) and makes the

solutions leave in a common outlet tube which is placed between the two inlet

tubes (Scheme 13). The advantage of this fluidic system is the permanentrinsing of the grating area of the sensor surface. Since the incoupling angledepends on the effective refraction index, adsorption of biomolecules could

result in a decrease of the light intensity within the waveguide layer due to a

shift of the initial incoupling angle. The area of detection is located between

the right inlet and the outlet.

70


Scheme 13: Cross-section of a counter-flow cell: The arrows indicate the

flow of the solutions. Sample solutions, tracer solution, rinse

solution, etc., are injected at the right (1). At the same time a

constant flow of pure assay buffer is injected at the left (2) inlet

of the fluidic cell. The solutions and the buffer leave the cell

through the common outlet (3). This design prevents the grating

(4) from adsorbing material and from shifting the incoupling

angle.

71

IV Results and Discussion


1 Cleaning of the Substrate Surfaces

The requirements of a waveguide layer for application in optical

sensing are high transparency and high refractive index. The formation of

amorphous (or polycrystalline) films by a physical vapor deposition process

provides highly transparent layers and makes metal oxides such as tantalum

oxide (Ta205) and niobium oxide (Nb2Oi) interesting base materials.

Diffuse diffraction of incoupled laser light leads to the loss of

evanescent field energy, an increase in the noise signal in the sensor

application and a decrease in the reproducibility of the results. The reasons

for diffuse diffraction in such waveguide layers include the formation of

metal oxide crystal, the roughness of the surface and the presence of

impurities within and/or on top of the surfaces. The cleaning procedure of

waveguide devices immediately prior to the functionalization process is

therefore essential. It is important to use an efficient but mild cleaningmethod, which does not increase the surface roughness. We successfullyused ultrasonic treatment in 2-propanol, UV-cleaning or oxygen plasma

cleaning as well as combinations thereof.

1.1 Cleaning Procedures

All glass or PTFE containers were cleaned with piranha acid (cone.H2SO4 : H2O2 - 3:1 mixture) for 1 hour before use, followed by extensivelyrinsing with ultrapure water until the water was pH-neutral. The containers

were then washed several times with 2-propanol. Before and after every

batch of sample cleaning and/or coating process these containers were

treated in an ultrasonic bath with 2-propanol for 15 minutes.

72


All substrates were placed either in a highly clean glass container or in

a highly clean PTFE container and treated in an ultrasonic bath with

2-propanol for 15 minutes. It is very important to place the samples in a

vertical position to remove microscopically small particles, such as glasspowder, during the somcation. Such small particles cannot be removed bysimply blowing away, not even with a strong gas stream. The samples were

removed from 2-propanol and blow-dried in a nitrogen stream followed byan additional cleaning in a UV-ozone cleaner for 30 minutes or in an

oxygen-plasma cleaner for 2 minutes.

Physical vapor deposited Ta205-chips (ion plated or sputter coated) and

NbiCVchips (sputter coated) were investigated before and after different

cleaning procedures by dynamic water-contact-angle measurements and byXPS at two different take-off angles. The 75° take-off angle (with respect to

the surface plane), with an information depth of about 9 nm, is more

substrate sensitive, providing data for determination of metal oxide

composition. The take-off angle of 15° (depth information of about 2.5 nm)is more surface-sensitive, and contamination is more easily detected. The

combination of data from the two take-off angles provides importantinformation about the contamination (e.g. adsorbed contaminants at the

surface or impurities within the metal oxide layer).

The cleaning procedures were: a) ultrasonic treatment in 2-propanol for

15 minutes, b) oxygen-plasma cleaning for 2 minutes and c) combination of

both.

1.2 Results

As received, the substrates were contaminated with small white

particles that were easily visible by eye. These particles could not

completely be removed by a nitrogen gas stream and were still present after

the direct oxygen plasma cleaning. After sonication in 2-propanol no more

visible contamination could be detected.

5


Advancing water-contact angles of the waveguide chips were

measured at three different regions on each sample. The samples were

measured as received and 48 hours after the different cleaning procedures(Table 7).

Table 7: Advancing water-contact-angle data from Ta205- and Nb2C>5-coated glass samples after different cleaning procedures.

Substrate Deposition Method Cleaning procedure Contact Angle[°]

Ta205 ion plated non-cleaned 79

sonication 76

oxygen plasma 53

both 46

Ta2Os sputter coated non-cleaned 80

sonication 57

oxygen plasma 50

both 50

Nb205 sputter coated non-cleaned 82

sonication 57

oxygen plasma 57

both 31

The water-contact angle of smooth metal oxide surfaces, which are

measured immediately after the oxygen plasma cleaning, are in generalbelow 5°. Due to the high activity of freshly cleaned surfaces, it is importantto carry out further modifications in a clean atmosphere and immediatelyafter the cleaning process. The data presented in Table 7 loosely reflect the

performance of the different cleaning methods.

74


The chips were stored after cleaning in closed glass boxes overnightuntil the XPS investigation was performed. The atomic concentrations of

metal, oxygen and carbon are listed in Tables 8 a) and b) and in Figure 6.

Table 8 a): XPS data from 75° take-off angles: Atomic concentrations of

metal, oxygen and carbon.

Substrate Deposition method Cleaning

procedure

Atomic concentration

[%]

Nb;Ta 0 C

Ta205 ion plated non-cleaned 21.3 63.5 15.2

sonication 20.1 61.5 18.4

oxygen plasma 22.7 72.1 5.2

both 22.4 72.2 5.4

Ta205 sputter coated non-cleaned 17.0 54.3 28.7



both 21.9 72.1 6.0

Nb2Os sputter coated non-cleaned 19.1 54.2 26.7



both 23.6 70.9 5.5

75


Table 8 b): XPS data from 15° take-off angles: Atomic concentrations of

metal, oxygen and carbon.

Substrate Deposition method Cleaningprocedure

Atomic concentration

[%]

Nb;Ta 0 C

Ta205 ion plated non-cleaned 12.3 42.7 45.0



both 15.1 64.7 20.2

Ta205 sputter coated non-cleaned 6.2 25.1 68.7



both 15.6 66.3 18.1

Nb2Os sputter coated non-cleaned 10.9 37.7 51.4



both 18.6 65.7 15.7

76


o

4)oft

oo

o

o

Cd

Fig 6

80 0

70 0

60 0

50 0

40 0

30 0

20 0

10 0

0 0

"^

^\

^—

;

\ \

T1^

\— -

\

__ -

v^\\

\N

r a205 (i p ) Ta205 ( s c ) Nb20S (sc)

E3 non-cleaned

D sonicated

02- plasma

D both methods

XPS carbon signal from ion-plated T'àjO^ (i p ). sputter-coatedTaiOs (s c ) and sputter-coated Nb2Üs (sc) after different cleaningprocedures Take-off angle 15°

The comparison of the atomic concentration data (XPS) shows that the

carbon signal originates from the surface and not from the metal oxide bulk

material (higher carbon signal at the take-off angle of 15°) The oxygen-to-metal ratio is higher than the theoretical value of 2 5 I his can be explainedby the adsorption of oxygen, C02 or water lhe O/M ratio of the XPS data

measured at a take-off angle of 75° approaches the theoretical value I his

observation agrees with the hypothesis of adsorbed oxygen, CO; and/or

water

Drastic reduction of the carbon concentration aftei treatment in the oxygen-

plasma was detected, reflecting the cleaning performance of the method

jtmt


For the reduction of adsorbed hydrocarbons, the ultrasonic treatment

was not very successful. Nevertheless, it is an important precleaning step for

the removal of particles and for highly contaminated surfaces. However,

oxygen-plasma cleaning is a high-performance cleaning method, providingmetal oxide surfaces with a very low amount of hydrocarbons.Consequently, the combination of both cleaning methods seems to be an

ideal pretreatment for further biosensor applications.

Except for the carbon signal, no other elements beside the expectedmetal and oxygen peaks could be detected by XPS.

UV cleaning for 30 minutes instead of 02-plasma treatment for 2

minutes resulted in contact-angle values of < 10° if the samples were

measured immediately after the cleaning process. XPS results show low

carbon concentrations after the cleaning with similar recontamination as for

the oxygen-plasma-treated samples.

1.3 Influence on Further Surface-Modification Steps

A cleaning procedure has been shown to be necessary to remove

macroscopic-to-microscopic particles and even molecular surface

contamination. Whether small amounts of contaminants can be replaced byphosph(on)ates during the self-assembly process is discussed below:

Ta2Û5 chips were cleaned using three different procedures: a) one

cleaning cycle of ultrasonic treatment in 2-propanol for 15 minutes, b) two

cycles of ultrasonic treatment in 2-propanol for 15 minutes and c) one cycleof ultrasonic treatment in 2-propanol for 15 minutes followed by UV-

cleaning for 30 minutes.

All chips were then immersed in a 0.05 mM ODPO4 solution (in n-heptane +0.4% 2-propanol). The samples were removed after 48 hours, rinsed with

78


2-propanol and blow-dried with nitrogen (for more details see chapter IV

2.1.1).

The wettability of the resulting SAM surfaces is a good parameter to

determine the SAM quality. The cleaning by ultrasonic treatment followed

by self-assembly resulted in a contact angle of 108° ± 2° independent of

whether one or two cycles were effected. The additional UV-cleaning priorto coating resulted in a contact angle of 110° ± 2°. This does not coiTespondto a significant difference.

However, the microdroplet density ((idd) images show a much higherhomogeneity for the chips that had been pretreated with an additional UV-

cleaning. Figure 7 shows condensation patterns from samples that have been

cleaned by simply sonication in 2-propanol and by additional UV-cleaning,followed by immersion in a ODPO4 solution in both cases.

Fig. 7: judd condensation pattern following self-assembly of ODPO4 on

Ta205: Ta205-chip immersed in a 0.05 mM ODP04-solution for 48

hours after different cleaning procedures: sonication for 15

minutes in 2-propanol (left) and sonication in 2-propanol followed

by UV-cleaning for 30 minutes (right).

The droplet pattern indicates inhomogeneous SAM formation on lÔssurfaces that were only cleaned by sonication with 2-propanol, and

79


homogeneous monolayer formation after the combination of sonication and

UV-cleaning.

According to the results of the previous chapters we conclude that a

combination of both sonication in 2-propanol for 15 minutes followed byUV-cleaning for 30 minutes is needed for good surface coating performance.

1.4 Influence on the Waveguide Properties

As shown in the previous sections, the cleanliness of the metal oxide

surfaces is an important factor. On the other hand, it is also essential that the

transparency of the waveguide layer remain as high as possible and that the

light mode intensity attenuation is small.

Planar waveguide chips consisted of an incoupling and an outcoupiinggrating (Scheme 14) were treated by different cleaning procedures and the

intensity of the light mode was determined by measuring the intensity of the

outcoupled light.

State of the chip treatment:

a) chip measured as received without any treatment

b) ultrasonic treatment in 2-propanol for 15 minutes

c) UV-cleaning for 30 minutes after treatment b)

d) chip stored at lab conditions for 10 minutes after treatment c)

e) chip stored at lab conditions for 60 minutes after treatment c)

f) chip stored at lab conditions for 2 hours after treatment c)

g) chip stored at lab conditions for 24 hours after treatment c)

h) additional ultrasonic treatment in 2-propanol for 15 minutes

i) chip stored at lab conditions for 5 more days

80


The intensity of the non-incoupled laser light was measured in parallel(measuring the 0, order refraction beam) and did not vary during the

different treatments (results not shown). The outcoupling light intensities are

listed in Table 9.

2) 3) 4)

„^V. HMttWM

1)

Scheme 14: Schematic drawing of the setup for the detennination of

lightmode propagation. Laser light (1) is in-coupled by a first

grating (2). Attenuation of the light intensity occurs within the

waveguide layer (3) and the intensity of the light (5) that is out-

coupled by the second grating (4) is measured by a photo-diode(6).

81


Table 9: Outcoupling intensity of laser light measured after different

cleaning procedures (a - i) of a Ta205 PWG chip, as described in

the text. The values are relative to the intensity of the raw, non-

treated chip.

Treatment Outcoupling intensity

(relative value)

a) 1.00

b) 1.61

c) 0.05

d) 0.28

e) 0.44

f) 0.50

g) 0.40

h) 0.50

i) 0.48

The results show an extreme attenuation of the light mode. The partial

reversibility of this attenuation may be due to the energy that is brought into

the material through the laser light irradiation during the measurement, in

the form of electromagnetic waves, or by a local increase of the metal oxide

temperature.

A second series of chips were treated by 15 minutes ultrasonic

treatment in 2-propanol followed by a UV-cleaning for 30 minutes. The

chips were then left in the apparatus for the determination of outcoupling

light intensity and stored under permanent irradiation with laser light

(Figure 8).

82


Fig. 8: Light mode intensity measurement of a Ta205 planar waveguidechip after UV cleaning and during laser light irradiation.

After an irradiation time of 2 hours the initial attenuation (or wave

propagation) of the metal oxide layer was restored (Figure 8).

After this period the chip was treated in the UV-cleaner for a second

time. Subsequently, the chip was mounted in the apparatus and laser lightwas permanently incoupled. Less then 30 % of transparency could be

recovered within 2 hours of irradiation (Figure 8).

Another chip that was cleaned in exactly the same way (ultrasonictreatment and one UV cleaning) was removed after an initial laser irradiation

time of 30 minutes. The transparency recovery was about 50 % after this

period. The chip was then transferred to an oven and stored at 180°C for 60

minutes. The outcoupling laser light intensity that was measured after the

thermal treatment was still as low as about 50 % of the initial value.

Oxygen-plasma cleaning for 2 minutes gave similar results but with a

slower recovery of the light mode intensity (recovery of 60 % of the initial

energy after 20 hours of irradiation with a HeNe-laser).

83


The influence on the optical biosensor performance of both the UV

cleaning and the plasma cleaning was tested with a simple rabbit IgGimmunoassay (for details see section IV 3.2). The chips of both cleaningprocedures gave good and sensitive detection by measuring the fluorescence

of immobilized tracer. The apparatus measures a region on the chip surface

that is illuminated by the evanescent field only. A strong attenuation of the

light mode intensity or even a complete absorption of the laser light wouldresult in a strong signal reduction of the immunoassay.

We conclude that UV cleaning and oxygen-plasma treatment changethe oxidation state of the metal ions and/or the structure [Lee,2000;Gabr,1999] at the chip surface. Photo-reduction of titanium oxide caused byUV-cleaning has been reported [Wang, 1998]. The metal oxide seems to be

in a metastable state similar to a transition state. The return to the initial

stable oxidation state may be promoted by low energy such as laser lightand/or ultrasonic treatment [Heng,1987]. The temperature of 180°C seems to

be too low in energy to have any detectable effect.

The second UV cleaning for 30 minutes may result in a more profoundphoto-reduction. This may explain the lower recovery of the wave

propagation after a second UV irradiation due to difficulties in reoxidatingthe metal ions.

Even after a treatment of 90 minutes in the UV cleaner no difference

could be measured by UV/VIS spectroscopy (300 - 700 nm ) between a

cleaned and a non-cleaned chip.

An overall conclusion about the cleaning of biosensor surfaces is that a

dedicated pretreatment step is essential for high reproducibility and

homogeneity of the following surface modification steps and the final

biosensor performance. Due to the less pronounced effect on the waveguideproperties we decided to treat the biosensor surfaces with UV cleaning and

not with oxygen-plasma cleaning.

The standard cleaning protocol is, therefore, the following: sonication in

2-propanol for 15 minutes followed by UV cleaning for 30 minutes.

84


2 Coating of substrate surfaces

2.1 Self-Assembled Monolayers (SAMs)

Self-assembled monolayers (SAMs) on metal [Folk, 1995] and metal oxide

surfaces [Gao,1997; Bain,1989; Nuzz,1983; Ulma,1991] represent a powerful and

highly flexible approach for the creation of surfaces and interfaces with a highdensity of functional groups (concentrated planes of surface functionality). This

methodology has the potential for applications in many areas, such as bio¬

sensors [Nyqu,2000; Hick,1991; Swal,l987], protein-adsorption investigations[Shum,2000; Hard, 1998], cell-behavior study [King, 1999], corrosion-resistant

systems [Alst,1999; Bram,1997], adhesion promotion [Maoz,1988], etc. The

principal classes of SAMs investigated and applied until now have been based

either on the interaction of chlorosilanes [Ulma,1990] with OH-terminated oxide

surfaces, or on the adsorption of thiols on gold [Bain, 1989]. A smaller number of

publications has appeared where alternative chemistries have been employed to

coat oxide surfaces with SAMs. These have included hydroxamic [Folk, 1995] and

carboxylic acid [Aron,1997; Laib, 1989] systems.

To date, the best-studied system has been that of self-assembled alkyl thiolates

on gold (or in certain cases silver [Port, 1987]), where thiols adsorb onto the metal

surface, first in a "lying-down" phase, followed by a rearrangement into a

"standing-up" phase, which completes the monolayer and results in a highlyordered two-dimensional structure [Folk, 1995; Chid,1991]. However, the use of

gold and silver is a problem where optical transmission is required, as in our case.

Alkyl phosphates and phosphonates constitute two systems that have been shown to

form ordered SAMs on metal oxide surfaces [Wood, 1996a; Maeg,1997; Folk, 1995;Gao,1996; Brov,1999]. Transition metal oxides, such as tantalum oxide (Ta205),niobium oxide (Nb2Os) and titanium oxide (Ti02), in particular, are known to

interact strongly with phosph(on)ates and to form stable interfacial bonds. Both

systems, thiol-gold and phosph(on)ate-metal oxide, form stable monolayers with a

"tail-up" orientation and a tilt angle of the hydrocarbon chains of about 30° with

respect to the surface normal [Text,2000]. Such amphiphile adlayers are generallyproduced from solutions in organic solvents.

85

TV Results and Discussion

2.1.1 Octadecyl Phosphate (ODPO4)

Octadecyl phosphate (ODP04), self-assembled onto Ta205 was the most

thoroughly investigated amphiphile - metal oxide combination in this work for two

reasons: a) Tantalum oxide is the standard coating material for the biosensor chipswhich we used for the assay performance, b) Since ODP04 possesses a longhydrocarbon chain (which is important for SAM stabilization due to Van der Waals

interactions between the alkyl chains), we assumed that its SAMs are best suited to

produce reproducible hydrophobic surfaces. This is the reason why this system is

treated in a separate chapter.

The long-chain monofunctional alkyl phosphate ODPO4 was used in its non-

charged acidic form, which is soluble in organic solvents and insoluble in water.

2.1.1.1 Influence of Different Solvents [Folk, 1995 ; Dann, 1998]

To study whether the polarity of the solvent has any influence on the SAM

quality we prepared ODPCVsolutions in different solvents or solvent mixtures. The

target concentration was as low as 0.5 mM. Nevertheless, in some solvents the

solubility limit was even lower. In these cases, we used saturated solutions.

Ta205 chips were cleaned according to the standard cleaning procedure (15minutes ultrasonic treatment in 2-propanol followed by 30 minutes UV cleaning)and immersed in the ODPO4 solutions. The chips were removed after 48 hours,rinsed with 2-propanol and blow-dried with nitrogen.

The hydrophobicity and the homogeneity of the SAMs were investigated usingcontact angle and (idd data, respectively. The results are listed in Table 10.

86


Table 10: Contact-angle and judd data for ODP04-SAMs on Ta205 from 0.5 mM

solutions in different solvents (* = solution near the critical micelle

constant [cmcj; ** = the solution is saturated with a concentration below

0.5 mM).

Solvent Dielectric

constant

Dipole¬moment

Contact

angle [°]

fidd1(± 20%)

isooctane** 1.9 0.0 113 ±0.9 1150

ether* 4.3 1.3 108 ±1.5 900

ethyl acetate* 6.0 1.8 113 ±3.5 860

tetrahydrofuran 7.4 1.6 100 ± 1.1 >3000

1-octanol 10.3 1.7 88 ±1.3 2800

isobutanol 17.7 1.8 91 ±0.8 >3000

2-propanol 18.3 1.9 93 ±1.8 1800

1-propanol 20.1 1.7 89 ±1.4 1450

acetone** 20.7 2.7 111 ± 0.7 2500

ethanol 24.3 1.7 80 ±1.2 2000

acetonitrile** 37.5 3.4 110± 1.2 2200

n-heptane + 0.4% 2-propanol* 112 dr 1.6 220

ethanol + 5% dimethylformamide 79 ±1.7 2700

droplets per mm'

All solutions near the critical micelle concentration (cmc) or near the

saturation point of the amphiphile produced significantly higher contact-anglescompared to solutions made with solvents that dissolve the amphiphile easily.

Ether, ethyl acetate and the n-heptane/2-propanol mixture are the only solvents

leading to perfect solutions and resulting in good to acceptable contact-angle data

for high SAM quality.

87


The microdroplet density (|idd) results suggest low homogeneity of the SAMs

with low hydrophobicity, an observation attributed to incompleteness of monolayerformation and a high degree of disorder,

Adlayers from saturated ODPO4 solutions (isooctane, acetone, acetonitrile) are

highly hydrophobic and slightly opaque. Loss of transparency of the samplessuggests partial multi-layer formation and/or micelle adsorption, resulting in higherroughness that may increase the contact angle due to the so-called lotus effect

[Bart,2000]. The high (idd is also an indication of the in-homogeneity of the

adlayers.

Considering only the solvents that were able to dissolve ODPO4 to a 0.5 mM

concentration, one can see that high polarity may be a stipulation for good SAM

quality. Another fact is that 0.5 mM ODPO4 solutions in ether, ethyl acetate and

n-heptane/2-propanol are near to the cmc.

With the knowledge of these first studies, we decided to use n-heptane/2-propanol as a standard solvent for the following experiments with ODPO4.

2.1.1.2 Adsorption Kinetics and SAM Stability

The self-assembly processes of alkyl phosphates onto metal oxides and alkylthiols onto gold are known to be very fast. Thiols adsorb at the gold surface and are

able to diffuse laterally. The lateral déplacement allows for dense packing of the

final monolayer. Phosph(on)ate SAMs are believed to grow in the form of islands

[Wood, 1996b]. The monolayers are then completed by coalescence of these

islands. The adsorption of the first 90 % of the phosph(on)ate and the thiol

monolayer, respectively, takes place in a few minutes or even seconds while

completion of 99 % or more of the SAM may take hours.

Self-assembled monolayers are often compared to 2-D crystals. If the growthmechanism of SAMs were similar to that of crystals one would expect more

homogeneous monolayers, if the metal oxide surfaces were immersed in solutions

at lower amphiphile concentrations and monolayer completion performed as slowlyas possible.

88


For the kinetic studies of ODPO4 adsorption onto tantalum oxide, a Ta_)Os

waveguide chip was cleaned by sonication in 2-propanol for 15 minutes and

subsequent UV cleaning for 30 minutes. The chip was then placed in the grating-coupler apparatus (see chapter TU 2.3.1) and a solvent mixture of n-heptane +- 0.4%

2-propanol was pumped through the flow cell to get a stable base line signal. The

optical signal baseline was set to zero and a 0.05 niM solution of ODP04 in the

same solvent mixture was pumped through the flow cell (How rate: 1.3 ml/min).This setup allows for real-time investigation of the adsorption kinetics with a time

resolution of milliseconds. In parallel we immersed pre-cleaned tantalum oxide

chips (same cleaning procedure) in ODP04 solution with different amphiphilcconcentrations, namely 0.5 mM, 0.05 mM. 5 jaM and 0.5w,M. After incubation

times of 5, 10, 30 and 60 minutes, the chips were removed and rinsed with 5 ml of

2-propanol each, and the corresponding contact angles were measured. The results

of both the contact-angle and grating coupler data are shown in Figure 9a.

VO

è

1

a

g

600

500

400

300

200

100

0

*rÔ*/Sl*>!gj*«»^^

w

10 20

JSk

—GC data (0 05 ra MOD?)

u C\data(0 5mMODP)

A CA data (0 05 mM ODD

\ 0 \ data (0 005 mVIODP)

X 04. data (0 0005 mMODP)

120

110

100

90

80

70

60

50

40

30

20

30 40

time (min)

50 60 70

asw

u

Fig. 9a: Grating coupler (GC. line) and contact-angle (CA, data points) data as a

function of self-assembly time of ODP04 (solutions m n-heptane + 0.4%

2-propanol) on TaÔs for different concentrations.

89


The 0.5 mM and 0.05 niM solutions of OüP04 show similar adsorptionkinetics onto tantalum oxide. The 5 uM solution gives a slightly slower coverage of

the metal oxide, but after an immersion time of 60 minutes the wettability is almost

as high as for the SAMs formed from higher amphiphile concentrations. The 0,5

uM solution is definitely too low 111 concentration for formation of a complete

monolayer. In this case, the amphiphile solution volume to metal oxide ratio was

chosen in a way that the amount of molecules would cover 60 to 80 % of the

sample surface.

To verify that the contact-angle data can be correlated with the grating couplermeasurements we repeated the experiment with the 0.05 mM and the 5 u.M

concentrations as described in the previous series, fhis time we removed samplesfrom the ODPO4 solutions after 0.5, 1, 5, 10, 30 and 60 minutes for wettabilitydetermination. In parallel, the same solutions were pumped through the flow cell of

the grating coupler with a How rate of 1.3 ml/mm at room temperature, and the

amphiphile adsorption was measured in real time (figure %).

600 12^

105

- XI^

(^

45

->«;

0 n —

0 20 30 40

time (min)

<

Fig. 9b: Coverage variation with assembly time derived from contact-angle (CA)and grating-coupler (GO results for two different concentrations of

ODPO4 in n-heptane + 0.4% 2-propanol.

90


Comparison of the results from the two methods demonstrates that contact-

angle data can be used as a sensitive marker for the SAM formation of ODPO4.

Grating-coupler data show that lower concentrations do not automatically result in

a higher mass adsorption, that would suggest for higher order in the monolayer.From these data we conclude that self-assembled monolayers may be compared to

2-D crystals, but their formation mechanism must be somehow different.

The samples used for the ODPO4-SAM-formation study by contact-angle were

also investigated by the |idd method. The distribution of condensed droplets was

already homogeneous (on the 10-|im scale) after an immersion time of 10 minutes.

However, the high density (> 2000 droplets/mm") showed that the SAM was not

yet densely packed at this time. The value dropped to about 1000 droplets/mmafter 60 minutes and about 300 droplets/mm*" after 180 minutes of immersion in the

ODPO4 solution. After 48 hours, the jidd reached 220 droplets/mm2.

For biosensor applications, one of the most important solutions is the PBS

buffer. This is the reason why SAM stability has been extensively studied in this

medium:

ODPO4 SAMs on Ta2Os were prepared in a standard manner (substratecleaning by ultrasonic treatment in 2-propanol for 15 minutes followed by UV

cleaning for 30 minutes, subsequent immersion in a 0.5 mM amphiphile solution in

n-heptane + 0.4% 2-propanol for 48 hours, removal of the substrates, and

immediate rinsing with 2-propanol, and blow-drying in nitrogen gas stream). The

samples were then transferred into a PTFE container and PBS buffer was added.

Chips were stored in the permanently stirred solution at room temperature for a few

minutes up to 16 days. Desorption of ODPO4 was determined by measuring the

contact angle (Figure 10) as well as XPS spectra at a grazing take-off angle (15°).The intensities of the carbon, oxygen, phosphorus and tantalum signals are plottedas a function of immersion time (Figure 11). The Ois signal was fitted and

deconvolved into two different types of oxygen: the phosphate oxygen (binding

energies: 531.8 and 533.1 eV) and the substrate oxygen (binding energy:

530.7 eV).

91


The monolayer thickness was calculated using a simplified 2-layer model

allowing for easy thickness comparison within a series such as the SAM stabilitystudies as well as for SAM formation from different amphiphiles on different

substrates (see later).

The 2-layer model only considers the substrate signal attenuation as a function

of the applied take-off angle due to the mean free path (or "attenuation lengthvalue") A,(Ekin). The equation used to calculate A.(Eun) is:

X = B •.y h v

- H- - EB equation 4

where B is the attenuation factor (B = 0.087 for organic overlayers), hv is the X-

ray energy (hv = 1486.6 eV for Al Ka), w is the work function (w = 3.95 eV) and

Eb is the binding energy of the substrate element electron (Eb = 27 eV for Ta4f).A-(Ekin) for Ta4f has been calculated to be 3.0 nm.

The calculation of the monolayer thickness d is shown in equation 5.

X»

d =

A{15°)

A(75°)1 1

equation 5

sin 75° sin 15°

where A is the integrated peak intensity (area) for the substrate (e.g. Ta4f) at the

corresponding take-off angle (15° or 75°).

92


0 50 100 150 200 250 300 350 400

time [hj

Fig. 10: Contact-angle data for an ODPO4-SAM on Ta205 assembled from a

0.5 mM alkyl phosphate solution. Chips were stored in permanentlystirred PBS solution and removed after different immersion times. The

results are compared to the data of non-immersed ODPCVcoated chips(t = 0) and are plotted as a function of immersion time.

The linear decrease of the wettability suggests a lower amount of CH3 groups

at the top surface and an increase in the CH2 group concentration. This can only be

achieved by desorption of amphiphiles (possibly due to an exchange process with

inorganic phosphate anions) causing lower order and a higher average tilt angle of

the alkyl chain of the self-assembled molecules. The udd increased from 160 ± 20

before storage in PBS to 330 ± 30 after 50 hours and to 1500 ± 100 after 16 days of

immersion in PBS suggesting an increase in roughness and a decrease in

homogeneity and hydrophobicity. Nevertheless, the patterns of condensed

microdroplets, even at high uxld, suggest homogeneous surfaces on a \xm scale.

93


To get more insight. XPS was used, (Figure 11)

o

o

o

70

5 50

40

30

20

10

0(-11——Inr-

-o

100 200

time (h)

300 400

C

O from PCX

o O from Ta205

HI IB.

v

—— Linear (C)

Linear (() from PCX)

Linear (O from Fa205)

—Linear (Ta)

Linear (?)

Fig. 11: XPS data from ODP04 SAMs on Ta2Os stored in a permanently stirred

PBS buffer solution at room temperature for different times. The Ols

signal was deconvolved into two different types of oxygen. A take-off

angle of 15° with respect to the surface plane was used for all

measurements.

The linear decrease of the carbon signal reflects the desorption of ÜDPÜ4,

suggesting that self-assembled organic phosphate from ODP04 undergoes a

substitution reaction with the inorganic phosphate from the PBS. The increase of

the oxygen signal from the phosphate is significantly higher than the increase of the

oxygen from the metal oxide. This suggests that the SAM thickness decreases (lesssignal attenuation for the substrate and interface elements) and that inorganicphosphate adsorbs at the surface. Fhe calculated monolayer thickness from XPS

94


data of the ODP04-covered samples showed a reduction down to 70 % of the initial

SAM thickness after an immersion time of 384 hours (16 days) in PBS buffer

(calculation using the 2-layer model).

Again, the results were confirmed using the grating coupler method. For this

experiment, a waveguide chip was cleaned by sonication in 2-propanol for 15

minutes followed by UV-cleaning for 30 minutes. The chip was fixed in the gratingcoupler apparatus and the solvent mixture (n-heptane + 0.4% 2-propanol) was

pumped through the flow cell (flow rate: 1.3 ml/min). The baseline was set to zero,

and a 0.5 mM ODP04 solution in the same solvent mixture was pumped throughthe flow cell. The amphiphile adsorption was recorded for 60 minutes.

Subsequently the flow was again switched to the pure solvent mixture, and the

SAM stability was monitored in the amphiphile solvent for 2.5 hours under a

constant flow.

The chip was removed from the apparatus, rinsed and immersed in the 0.5 mM

ODPO4 solution for 23 hours to complete the SAM and to get closer to the standard

coating procedure. The chip was then rinsed with 2-propanol and immediatelymounted in the grating coupler apparatus. PBS buffer solution was pumped throughthe flow cell, and mass desorption measured for 2 hours (Figure 12).

Grating coupler data suggest the absolute stability of the ODPO4-SAM in the

n-heptane/2-propanol solvent mixture, even after only 60 minutes of self-assembly.Long-term investigations of SAM stability in PBS show a linear rate of amphiphiledesorption due to exchange reaction (substitution) between organic phosphates and

inorganic phosphate molecules (Figures 10 and 11). Results from short-time

measurements performed with the grating coupler method show that desorptionof ODPO4 takes place in an exponential manner over the first 30 to 60 minutes

(Figure 12).

95


1200

600

500

i 400

t "00

3 200 J (I)

100'

1

0'

i 260 B20 1 ^80 1140 1500 : 560

*.* ^ \ % * (*> \ »^ 0)

\

(2)

n-lieptane/2-ptopanol

-10 50 170 ^0

time (mm)

PBS bulfei

290

Fig 12 Grating coupler output as a function of immersion time

from a 0 5 mM ODPO4. solution in n-heptane -+ 0 4%

After 60 minutes the flow was switched to pure solvent

stability of the SAM m pute solvent is recorded (green 1

axis) (3) fhe chip was removed for SAM completion, r

apparatus and mass desorption recorded for 2 houts (redaxis (upper axis) reflects the total chip immersion time

formation

350

(1) Adsotption2-propanol (2)(arrow) and the

me and green \-

emounted in the

line) fhe red x~

includmg SAM

Pure water seems to be less corrosnc for ODP04 SAMs than phosphate buffer

A short rmse with water resulted in a decrease of udd (likely to reflect higher order

due to removal of non-oriented adsorbed molecules) fhe increase 111 wettability(contact angle drop of 2 - 3 degrees) and decrease in the calculated monolayerthickness (by about 0 1 nm, corresponding to 5 %) alter 10 minutes of immersion

in water aie consistent w ith suggestion of mass desorption

96


Results concerning the stability of other alkyl phosph(on)ates in contact with

water are listed in the appendices (App. 18 and 19).

To test the thermal stability of alkyl phosphate SAMs, ODPCVcoated Ta205

chips were stored in air at 75°C and 130°C for two hours. Contact-angle and XPS

data were recorded before and after the treatment. Thermal treatment of these

SAMs did not have any influence on the wettability (contact-angle: 112 ± 1°) of the

surface. The thicknesses calculated from XPS data (recorded at 75° and 15° take¬

off angles with respect to the surface plane) and the composition of the SAMs

(determined from XPS survey spectra) did not vary between samples measured

before and after the heating process.

This is a very important result because such thermal stability provides the

possibility of sterilizing ODPCVcoated substrates for further biochemical and/or

biomedical applications (implants or sensors with living cells).

2.1.1.3 Characterization of ODP04 SAMs on Ta205

Qualitative XPS:

The self-assembled monolayers deposited from 0.5 mM ODPO4 solution in

n-heptane + 0.4 vol. % 2-propanol were analyzed at four different take-off angles:15°, 45°, 75° and 90°. Comparison of the survey spectra from these take-off angles(Figure 13) shows a strong attenuation of the tantalum and oxygen signals and an

increase of the carbon signal with decreasing take-off angle (90° - 15°).

97


3

C

1000 800 600 400

Binding Energy (eV)

200

Fig. 13: XPS survey spectra of self-assembled ODP04 monolayer on tantalum

pentoxide obtained at different take-off angles (0=15, 45, 75 and 90°).

The detailed spectra of C|S, Ojs, ?2P and Ta4f taken at several different take-off

angles were resolved into their components by the curve-fitting proceduredescribed in chapter III 2.2.1. Some of the data are presented in Table 11.

98


Table 11: XPS binding energy (EB) values (± 0.1 eV) of a self-assembled ODP04

monolayer on Ta20j and chemical shifts (ÀE) referred to the ODPO4

powder and tantalum oxide substrate energy levels. Averages of data

taken at different take-off angles.

Photoelectron emission peaks

Cls(l) C.s(2) 0ls(l) 0|S(2) P2P 0)s(3) Ta4f7/2

Assignment aliphatic C-O-P R-OPO3 Ta205

EB (eV) 285.0 286.2 531.8 533.1 134.2 530.7 26.8

AE versus ref. (eV) _ -0.6 -0.3 -0.5 -0.5 + 0.1 + 0.4

Intensity ratio

0ls(l)/0ls(2)

1.6 ±0.2

The C]S signal was fitted with contributions from the aliphatic chain

(285.0 eV) and the carbon bonded to the phosphate group (286.2 ± 0.2 eV). The 0]s

signal is of particular interest for assessing of the type of bond between ODPO4 and

the substrate. This signal was fitted with three contributions (Figure 14): the

oxygen from the tantalum oxide was found at 530.7 ±0.1 eV (Ois(3)) and the two

oxygen components from the ODPO4 molecule were found at 531.8 ± 0.1 eV

(Ois(l)) and at 533.1 ±0.1 eV (0]s(2)). The energy separation between 0ls(2) and

01s(3) was 2.4 ±0.1 eV for all data, independent of the emission angle. The energy

separation Ois(2) - Ois(l) was found to be 1.35 ±0.1 eV at the 15° take-off angle.The intensity of 0]s(3) from the Ta205 substrate is a strong function of the chosen

detection angle. The intensity decreases as the take-off angle is reduced from 75° to

15° (grazing exit). The intensity ratio Oi,(l) : 0]s(2) was determined from

independent XPS measurements recorded at two take-off angles; at 15°, the mean

value was found to be 1.2 ± 0.2. At emission angles of 75° the intensity ratio was

1.9 ± 0.2. This difference may be attributed either to the influence in the curve-

fitting of the strong oxygen signal at 530.7 eV due to the substrate or to the fact that

99


at this angle the contribution of the oxygen which faces the tantalum pentoxide is

higher. Therefore, it has been decided to take the average over all 22

measurements, equal to 1.6 ± 0.2.

—, , , , , , , , 1 , 1 , , , , 1 1—-, , , , , , , , , , , , , , ,_

535 534 531 532 531 530 529 528 ->T> 514 533 532 511 530 529 528

Binding Energy (eV) Binding Energy (eV)

Fig. 14: Curve fitting of the XPS 0,s signal of an ODP04 SAM on Ta205 at 15°

and 75° detection (take-off) angle.

The P2p signal showed a comparatively poor signal-to-noise ratio and was

fitted with a single 2p doublet. The binding energy of the P2p3/2 component was

134.2 ± 0.1 eV, independent of take-off angle. The binding energy of the Ta4f was

found to be 26.8 ±0.1 eV.

The binding energy data in Table 11 show a shift of + 0.4 eV for Ta4i- and

ca. - 0.5 eV in Cis(2), 0!s(l), Ois(2) and P2p, relative to the Ta205 substrate and the

free acid (ODPO4 powder; results see chapter III 1.2.1), respectively. The shift to

lower binding energies of the ODP04-related signals is likely due to a (full or

partial) deprotonation of the phosphoric acid head group upon coordination to the

surface and the corresponding formation of a negative charge on the phosphatehead group. The positive shift of the Ta205 substrate suggests a charge transfer

from the substrate to the ODPO4.

100


Quantitative XPS:

The results of the quantitative analysis of four independent angle-resolvedXPS (AR-XPS) measurements are summari/ed in Figure 15. As the take-off angleis reduced to 15°, the amount of the t\ assigned to the aliphatic chain (285.0 eV)

increases and the contributions from the phosphorus, oxygen, and carbon

components of the C-O-P head of the molecule decrease. This indicates that the

polar head of the ODPO4 molecule is located m the inner part of the ODP04 SAM.

0 9

08S

a0

1 0 8

0 7

0 2^

A '< I)

1 ',1 (286 2) '1 c

( )

cc (

0 2

oO IS

0 1

.2 0 (Hss

20 40 f>0

angle (dec)80 100

Fig. 15: Angle-resolved XPS measurements of the elements C, 0 and P in the

ODPO4 SAM on Ta:(X The atomic fractions were calculated from the

intensities of the curve fitting procedure corrected for the sensitivityfactors given by Scofield |Seof,1976J. Results of two independentmeasurements (full and open symbols).

101


Tn the evaluation of the thickness and composition of the ODPO4 monolayeron Ta205-, the layered structure of the system under investigation has to be

considered; it is, therefore, not particularly useful to calculate an overall percentage

composition that does not take the different depths of the various elements and

groups into account. Instead, the depth origin of the different signals has to be

borne in mind and the composition and thickness of each layer calculated within a

multiple layer model as follows:

a) Firstly, all signals from Ta4f and the 0(s(3) oxygen at 530.7 eV originateexclusively from the Ta205 substrate and are thus attenuated by the

ODPO4-SAM.

b) Secondly, the results of the angle-resolved XPS measurements (see Figure 15)indicate that the P2p signals originate from the inner part of the ODPO4 layer.In other words, the P2P and the 0|, at 531.9 eV and at 533.2 eV are located on

top of the Ta205. The phosphorus and the oxygen components of the polarhead of the molecule can thus be considered as a very thin film containingonly P and O atoms. Their XPS signal intensities are attenuated by the carbon

chain of the ODP04 located at the top.

c) Finally, the C\s contribution of the C18H37 hydrocarbon chain originates from

the outermost part of the layer system.

Based on these observations, a "three-layer model" has been applied to the

ODPO4 SAM on tantalum oxide. This model is based on the composition of a) the

substrate, b) the PO4 interfacial layer and c) the hydrocarbon top layer

(Scheme 15). The SAM thickness can now be calculated based on the intensities

and the origin of the individual components of the elements as defined above. The

density of the Ta205 substrate was taken as 8.6 g/cm3, that of the hydrocarbonchains in the film as 1.1 g/cm"1 and that of the thin P04 polar head layer as

2.0 g/cm". Cross sections used for the calculations were taken from the literature

[Scof,l976]. The attenuation length values X(Ekin) of the photoelectrons were

calculated as X = BVEkin (with B = 0.096 for the inorganic compounds and

B = 0.087 for the organic layer [Seah,l979]). The parameters utilized in the three-

layer model are listed in Table 12.

102


a) substrate (semi-infinite)

Scheme 15: Drawing of the three-layer model of ODPO4 self-assembled on Ta2C>5:

Ta2Oi substrate (semi-infinite layer), PO4 head group (monomolecularinterfacial layer) and CigH^ (organic hydrocarbon chain layer).

Table 12: Photoelectron ionization cross-sections, 0 ,and electron attenuation

lengths, X, of self-assembled ODPOj monolayer on Ta2C>5, from data

given in references [Scof,1976] and [Seah, 1979].

Photoelectron emission peaks

C,s(D C,,(2) 0,s(D 0,s(2) P2„ 0ls(3) Talf7/2

Assignment aliphatic C-O-P R-OPCh Ta205

a (Scofield) 1 1 2.93 2.93 1.192 2.93 8.62

X (nm) inorg. - _ _ _ 3.53 2.97 3.67

X (nm) org. 3.32 3.32 2.69 2.69 3.20 2.69 3.32

The composition and thickness of the Ta205 substrate and the ODP04 self-

assembled monolayer, calculated with the three-layer model, are summarized in

Table 13. The results are averaged over all experimental measurements, and no

significant differences were found when analyzing data taken at different emission

103


angles. For both the ODPO4 layer and the substrate, good agreement is found

between the expected composition and that calculated from the three-layer model.

The thickness of the ODP04 layer (including both the polar head group and the

hydrocarbon chain) was found to be 2.2 ± 0.2 nm at all emission angles

investigated.

Table 13: Thickness and composition of self-assembled ODPO4 monolayer on

Ta2Q5, based on evaluation of the XPS data within a 'three-layer model'.

Physicalparameters

Elements

carbon oxygen phosphorus oxygen tantalum

Layer 1-3 C18H37 chain polar head P04 substrate Ta20s

Thickness (nm) 2.2 ±0.2 semi-infinite

Composition: atom%

Theoretical _ 80 20 71.4 28.6

Experimental - 77 ±4 23 ±4 70 ±2 30 ±2

ToF-SIMS:

The most prominent secondary ion masses observed in the static SIMS spectra

are listed in Table 14 (negative ions) and Table 15 (positive ions), together with

their intensities relative to the most intense peak within each class of fragment

species. SIMS spectra of the Ta205-ODP04 surface are only selectively shown in

Figures 16a and b for the most informative mass range m/z = 200-400 (negative

secondary ions), since all data are presented in Tables 14 and 15. However, spectra

across the full mass range are available as supplementary material.

104


Table 14: ToF-SIMS: List of the most prominent negative secondary ion masses

(m/z) and assignment to molecular fragment species. M = molecular

mass C18H37OPO3H2.

Secondary Ion

Mass [m/e]

(observed)

SpeciesCharge: -1

Formal

Oxidation State

of

Ta P

Deviation of

observed

mass1 [ppm]

Relative

Intensity2[%]

Class I: CaHbPOc

62.963 PO> +nt +4 23

78.958 PO< +v +4 100

95.957 PO4H +v +43 4

96.968 PO4H. +v +13 7

109.97 CtUOPCh +v +47 3

122.99 CiHiOPCMI +v -19 2

165.07 CfiHnOPO,H +III -13 0.7

181.06 CfiHnOPO,H +v +14 3

349.25 CmH,7OPO,H (M-H) +v -1 0.6

Class 11: TaaObHc

197.95 TaOH 0 +9 0.5

198.98 TaOH, -I +95 0.8

212.94 TaO, +111 +11 1.3

213.95 TaOOH +11 +31 1.3

228.94 TaCh +v + 13 2

229.95 TaCbOH +IV +33 2

230.95 TaO(OH), +111 0 0.8

246.96 TaCbCOl-Tb +v +51 1.6

266.97 TafOHVOH-.") +111 0 0.4

410.91 Ta.OOH +ir -54 0.5

411.87 TaÔ(OH), +1/11 +3 0.1

450.94 Ta,OH<OFMi 0 +3 0.04

458.87 Ta.OsOH +v +3 0.04

105


Class III: TaaPbOcHd

275.90 TaO(PO,) +1V +1IÏ +9 0.5

276.91 TaO(PO,H) +111 +ÏII -11 0.4

291.90 TaO(P04) +IV +V -16 1.0

292.91 TaO(P04H)

or Ta02(P03H)

+111 or +V

+V +IT1

+8 3

308.90 Ta02(P04H) +v +v +2 3

338.86 Ta(P03)2 +V +111 +27 0.2

354.85 Ta(P04)(PO,) +V +V/+1II +28 0.4

370.86 Ta(P04)2 +v +v +19 1.5

387.86 TaO(P04)(P04H) +VI +v +4 0.2

388.88 TaO(P04H)2 +v +v -48 0.2

504.84 Ta2(PQ2)(PChH) +11 +111 +25 0.2

520.83 Ta2(PO,)(PO,H) +11 +III +13 0.05

538.83 Ta20,(P04H2) +v +v +10 0.1

584.91 Ta202(P04)(PO,H) +1II/+V +v -85 0.2

600.79 Ta2(PO,)(PO,H)2 +III +v +1 0.3

662.75 Ta20(P04> +v +v +1 0.3

681.01 Ta20(OH2)(P04)3 +v +v -360 0.2

892.31 Ta304(P04)2(P04H) +v +v +419 0.2

Class IV: TaaPbOcCdHc

545.16 TaO(0,POC,8H,7)

= TaO-(M-2H)

+III +v +53 0.06

561.18 TaO*(OiPOCi8H^)

= Ta02'(M-2H)

+v +v +2 0.2

Deviation of experimentally observed mass from exact mass of assigned species in ppm

2

Intensity of the secondary ion peaks relative to the most intense peak (P03" = 100%) of the whole

spectrum

106


Table 15: ToF-SIMS: List of the most prominent positive secondary ion masses

(m/e) and assignment to molecular fragment species. M = molecular

mass C8H37OPO3H2.

Secondary Ion

Mass [m/e]

(observed)

Species

Charge: +1

Formal Oxidation

State of

Ta P

Deviation of

observed mass1

[ppm]

Relative

Intensity""

[%]

Class I: CaHbPOc

98.984 H4PO4 +v +6 100

120.97 CtTOPO, +v +6 8

125.00 C,H,OPO,H, +v 0 13

207.08 CsH„OPChH, +v +20 10

214.12 CnH,qOPO +1 +30 10

219.08 CoIinOPChPh +v +15 5

220.11 CmHixOPOoH, +IÏI +53 9

223.12 CgHi7OPO,H, +v +37 5

237.14 Ch>HiqOPO,H, +v -76 5

249.16 C„HÔPO,H, +111 +2 7

263.18 CnH7sOP07H, +TI1 -5 5

275.18 C,4H„OP07IT, +111 -3 3

277.20 CuH77OPO,H, +III -35 4

349.25 CirHÔPO^H^M-H) +v -3 0.3

351.27 CsHÔPO^rM+H) +v -2 0.2

Class II: TaaObHc

180.92 Ta +1 +182 26

181.93 TaH 0 +162 30

196.94 TaO +III 0 30

197.96 TaOH +11 +25 24

212.95 TaO. +v +43 5

230.94 TaO(OH), +v -20 6

231.95 TafOH), +IV -17 0.4

232.95 Ta(OH)?fOKR +111 -44 7

107


Class III: TaaPbOcHd

310.93 Ta(OH)2(P04ll) +V +V -51 0.8

328.94 Ta(OH),(P04H2) +v +v -31 1.3

372.90 Ta(P04H)2 +v +v -10 0.8

390.86 TaOH(P04H)(P04H2) +V IV +59 2

Class IV: TaAOcCdHc

356.96 Ta(OH),(HO,POC2H5) 4-V +V +2 1.0

395.01 Ta(OH)2(03POC6Hn) +V +v 0 1.0

412.98 Ta(OH),(HO,POC6Hn) +v +v +85 0.6

470.98 Ta(P04H)(CHPOC7Hi,) +v +v -1 0.4

1Deviation of experimentally observed mass from exact mass of assigned species in ppm

2Relative intensity of the secondary ion peak relative to the most intense peak (H4P(V - 100%)

The assignment of the masses to molecular ion species is discussed on the

basis of the classes of fragments:

Class 0: CmHn ions: These fragments occur on both the untreated and treated Ta2Os

surface and do not carry specific information about the nature and structure of the

adlayer. They originate both from fragmentation of ODPO4 and from the naturallyadsorbed (contaminant) hydrocarbons on the treated and the bare surface. They are

not included in Tables 14 and 15.

Class I: CaHbPOc ions: Ions of this type are only observed on the Ta2Os/ODP04

surface. Several species are observed in both modes, starting from the pure

phosphate fragments (e.g. P02\ PO3", HP04", H2P04", H4P04+) through various

partly fragmented phosphoric ester species up to the molecular masses ([M-H]~ and

[M±H]+). As expected, a certain amount of reduction of the phosphate (oxidation

state +V) to the phosphonate (oxidation state +III) takes place.

Class II: TaaObHc ions: Fragments of this type occur in both the negative and

positive modes and on both the treated and the bare Ta2Os surfaces. Species

108


corresponding to a formal stoichiometry of Ta(+V) and Ta(+HI) (the most stable

oxidation states in inorganic tantalum compounds) tend to have higher intensities

compared to the others. This is a frequent observation in SIMS spectra of metal

oxides and hydroxides. The general pattern of the TaaObHc ion fragments is similar

on both the bare and the ODP04-modified surface, although the distribution of

intensities is somewhat different.

Class III: TaaPbOcHd ions: The presence of strong peaks characteristic of tantalum

phosphate and phosphonate species in both the positive and negative spectra of the

ODP04-modified surface strongly suggests that the tantalum ions of the Ta20s

layer are actually directly bound to the chelating phosphoric acid group. It is

unlikely that complex tantalum oxide phosphate species would be detected with

high intensity if there were no direct complexation between the phosphate group

and the tantalum(V) cation. In particular, assuming that the phosphoric acid binds

to the oxide surface via hydrogen bonding, species consisting of both tantalum

oxides and the phosphoric acid group are unlikely to survive the emission process

without further fragmentation into pure tantalum oxide and phosphate species. The

fact that complex fragments such as TaQ(P04)(P04H)~, Ta205(P04H2)",

Ta202(P04)(P04H)-, Ta2(P03)(P03H)2", Ta20(OH)(P04)3\ Ta304(P04)2(P04H)",

Ta(P04H)2" or Ta(OH)3(P04H2)~ in the negative spectra and Ta(P04H)2+ or

TaOH(P04H)(P04H2)+ in the positive spectra are observed provides evidence for

the close molecular packing of the self-assembled ODP04 molecules on the Ta2C>5

surface and a binding scheme that involves, at least to some extent, more than one

phosphate head group coordinated to one Ta ion.

Class IV: TaaPbOcCdHe ions: Prominent secondary ion mass peaks correspondingto TaO(03POC18H37y= TaO(M-2H)\ Ta02(03POC18H37y = Ta02«(M-2H)-and

fragmented species, such as Ta(OH)3(H03POC2H5)+, Ta(OH)2(03POC6H13)+ or

Ta(OH)3(H03POC6Hi3)+, again support the presence of a strong bond between the

tantalum (oxide) and the ODP04 molecules. The majorityoftheseclassIVspeciescorrespondtotheformal,moststableoxidationoftantalum,i.e.Ta(+V).TheobservationofTa(P04H)(03POC7Hi5)^positiveions,albeitofweakintensity,suggestsaclosespacingofthephosphoricacidheadgroupsonthesurfaceand,again,thecoordinationoftwoSAMmoleculestothesameTaion.109


a)

900-« .—

O x -

ooCL.

800- oo

700- o

O

-

600-

m

g 500-

°1h

O

400- oa.

-

300-

o

ox ~-o

a ss=..J2J r O

-

200- »

100-

léki.Williy %^mM^,\

,3

230 250

m/z

290

b)

100-

-

o

O

-

H*

-

EN

c?a.

x"

^

XCI

o^

o 9*~ ^x1 "5? oo a.

O jSîcl x:

MW>w \w*^W^

"75Oa.

*W*<

1-,—. -*»

tJ Cla» Ä

—,—,

_—,:

î**

350

m/z

360

Fig. 16a and b: Selected ToF-SIMS spectra (negative secondary ions) of the

ODPCVTaÔs surface in the mass ranges m/z = 200-300 (a) and

300-400 (b). Spectra across the whole mass range investigated

(0-600) are available as supplementary material.

110


NEXAFS:

NEXAFS measurements were performed by D.Brovelli, G.Hähner and

I.Pfund-Klingenfuss and are reported [Brov,1999]. Figure 17 displays NEXAFS

spectra of ODPO4 adsorbed on Ta205 from a 0.5 mM solution in n-heptane + 0.4%

2-propanol for 48 hours at normal and grazing incidence angle of the incoming

X-rays. The sharp transition situated at 287.6 eV close to the absorption step can be

assigned to transitions into C-H* valence orbitals [Outk.,1988]. Its intensity reached

a maximum at normal X-ray incidence (0 = 90°). The two broader resonances

above the step at around 293.1 eV and 301 eV stem from transitions into C-C G

orbitals [Outk,1988]. They are strongest for grazing incidence (0 = 20°). To

emphasize this angular variation the difference spectrum between grazing and

normal incidence has been included in the figure. The transitions into C-C a and

C-H orbitals exhibit an opposite polarization dependence since the corresponding

molecular orbitals are perpendicular for alkyl chains in an all-trans conformation

[Hahn, 1991; Outk,1988]. A comparison with the spectra of a well-ordered SAM of

hexadecane thiol on gold (reference sample) reveals virtually no difference,

indicating a very similar degree of order in the two samples.

A more quantitative analysis [Stöh,1992; Kinz,1994; Fisc,1997] of the order in

the films yields a similar tilt angle, d, to that obtained for hexadecane thiol

monolayers on gold (i.e. u = 33° between the alkyl chain axis and the surface

normal). An error of less than 3° was calculated (for details see ref. [Fisc,1997]).

Ill


-PCO

a

275 285 295 305 315 325

Photon energy [cV]

Fig. 17: NEXAFS spectra of ODPO4 adsorbed on tantalum (V) oxide from a

0.5 mM solution in n-heptane + 0.4% 2-propanol for 48 h at normal and

grazing incidence

AFM:

AFM images were performed by K.Feldman. Figure 18 shows a lateral force

image of an ODPO4 layer on Ta205. Clearly visible is a certain degree of order in

the layer in the form of small patterned areas. However, apart from these ordered

regions, there are regions where no structure is clearly observable and the imagesare 'blurred'. Both structured and non-structured regions are on the order of a few

nanometers in diameter. The small, structured regions display a roughly hexagonal

112


pattern with an average nearest-neighbor distance of 0.49 ± 0.01 nm (see Figure

18). Assuming this arrangement and published [Hand, 1997] bond lengths (0.16 nm

for the O-P, 0.14 nm for O-C and 0.17 nm for O-Ta), densities of the phosphateand hydrocarbon regions of 2 0 g/cm and 1.1 g/cm , respectively, were calculated

(this information has also been used in the calculations of layer thickness and

composition from XPS data). The pattern is similar to that found for alkanethiols

on gold[Nuzz,1983; Sell,1993; Bish,1996; Ulma,1996], and we interpret the

resolved bumps as being the terminal methyl groups of the alkane chains. AFM

images over 1 |im" regions of the untreated Ta20^ surface revealed a very flat,

featureless topography with a roughness (Ra) of 0.2 nm.

Fig. 18: Lateral force image of an ODPO4 layer on a Ta205 surface. The left

image shows a smoothed version of the small inset shown in the image

on the right-hand side.

113


2.1.1.4 Summary, Conclusions and Suggestions

Orientation, stoichiometry, and thickness of the ODP04 SAM:

The results of the angle-dependent XPS investigation of the ODP04/Ta205

surface demonstrate that the terminal phosphate groups are oriented towards the

Ta205 substrate surface. This is strongly suggested by the evolution of the carbon,

oxygen, phosphorus and tantalum XPS intensities as a function of electron-

emission angle (Figure 15). A three-layer model was used to calculate the thickness

and stoichiometry of each layer (hydrocarbon, phosphate head group, and tantalum

substrate) (Table 13). By using the AFM data, densities for each layer could be

estimated and hickness and stoichiometry values for the individual layers could be

subsequantly calculated:

a) The adlayer thickness of 2.2 ± 0.2 nm calculated using the three-layer model is

in excellent agreement with the value of 2.1 ± 0.05 nm calculated from the

total length of the ODPO4 molecule (2.5 nm) and from the (average) tilt angleof 30 - 35° determined experimentally from NEXAFS measurements

[Brov,1999].

b) The O/P atomic ratio of 3.4 ± 0.8, calculated from the deconvolved 0]S and

from the P2p XPS peak areas, is roughly consistent with the expected ratio of 4

: 1 for the phosphate group. The relatively high standard deviation is due to

the low signal-to-noise ratio for the P2p emission.

c) The O/Ta atomic ratio of 2.3 ± 0.3 is close to the expected value for the Ta2Os

stoichiometry.

Substrate Bonding:

In this section we discuss the observations from ToF-SIMS and XPS related to

the question of the bonding mechanism of the phosphate head group of the ODPO4

molecule to the tantalum oxide substrate:

114


The positive and negative ToF-SIMS spectra showing a variety of fragments

corresponding to the classes 1 through IV (see Section ToF-SIMS in previous

chapter) provide conclusive information about the coordination of the phosphatehead group to the Ta cations:

a) The fact that prominent peaks of class III (TaaPbOcHd) and IV (TaaPbOcCdHe)are observed gives strong evidence for a direct coordination of the phosphate

group to Ta cations. If the binding of the phosphate group were weak, such as

hydrogen bonding to tantalum oxides and hydroxides at the surface (e.g.

ROPO(OH)2"-0-Ta), it would be unlikely that complex fragments of type III

and IV would survive the fragmentation and detection process and lead to

mass peaks of relatively high intensity. Moreover, many of the observed

fragments such as m/z = 371 (negative), 391 (positive), 561 (negative) etc.

cannot readily be explained assuming indirect coordination via

P-O-'-H-O-Ta or P-O-H—-O-Ta hydrogen bonding.

b) The fact that a very particular pattern of fragment stoichiometries is observed

supports the findings of a preferred model of coordination, discussed in detail

in chapter 2.1.1.5. The fragmentation types that are experimentally observed

in the positive and/or negative SIMS spectra are summarized in Table 16.

The findings are exactly what one would expect for a model of coordination of

ODP04 on Ta205 involving the presence of both bidentate ODP04 bound to one Ta

ion and of monodentate coordination of two ODP04 molecules to one Ta ion. The

assumption is that the secondary fragments in (static) ToF-SIMS accurately reflect

the original molecular structure of the surface and that the recombination of

fragments that were originally apart from each other is not a likely process during

secondary ion formation.

Based on the structure and preferred coordination of tantalum (V) in oxides,

structures of the molecular ion species corresponding to some of the more

interesting fragments of Table 14 and Table 15 are proposed in Figure 19.

115


Table 16: Observed combinations m/n for secondary ion fragments of the type

Tan(POxCyHz)m in the positive and negative ToF-SIMS spectra of ODPO4

SAM on Ta205.

Ssecondary ion fragments of the type Tan(POxCyHz)min the ToF-SIMS spectra of ODP04 SAM on Ta205

Number n of Ta atoms in

secondaiy ion fragment

Number m of POxCyHz groups

in secondary ion fragment

1 1

1 2

2 1

2 2

2 3

3 3

The predominant structural pattern in the crystalline state of Ta2Os i s

characterized by six-fold coordination of Ta and by edge-sharing octahedra, leadingto the 2:5 stoichiometry [Well, 1991]. Although the sputtered Ta205 substrate used

in this study is amorphous to semi-crystalline, it is still highly likely that the local,

short-range order environment is similar to that of the crystalline state. The species

proposed in Figure 19 can be seen as fragments of the original Ta205 polymericstructure with coordinatively bound phosphate moieties. The fact that the

coordination number of the Ta atoms in these fragments is generally lower than the

preferred value of six, is likely to be a consequence of the preferred oxidation state

of Ta being +V, +IV or +111. In the proposals of Figure 19, the coordination number

of tantalum in the fragments has been maximized subject to physico-chemicalconsiderations and to the hypothesis that in the original SAM layer only mono- and

bidentate phosphate coordination occurs (as opposed to three-fold coordination).

116


Q

O

\r, ...U\\CV//;„,-,

)

CH$ Oil,,.À .^\ùQllli„.rpf .„iirtttCV////,. jfso

J 16OH

m/z = 561 (negative) m/z =539 (negative)

HO3POQ

Xp...u\\0////;„.T .a\\0////,..

o

o

OH

m/z = 893 (negative)

\HO

,..,»l\\0//»J/,.r

HO/Ô'

/OH

Vv..i\\\0////„. r>

dô*^\OH

HO

HO Ta^rÎ

HO

OH

O"*CH,

m/z = 391 (positive) m/z = 413 (positive)

Fig. 19: Proposed structures for some of the observed fragments in the negativeand positive ToF-SIMS spectra of the ODP04 SAM monolayer on

tantalum oxide (Ta205).

117


XPS binding energies and chemical shifts:

The interpretation of the binding energies of the P2p and the Ois electrons (datafrom curve-fitted Ois detail spectra) provides an insight into the binding of the

phosphate head group of the ODP04 molecule at the surface.

Table 17: XPS binding energies of the P2p electrons in reference compounds

depending on the ratio of 'free' O ligands (n) to covalently bound OR

ligands (m, R = H and/or P). Data are from the literature [Gres,1979;

Moul,1992] and are referenced to the Cis of aliphatic hydrocarbons at

285.0 eV.

XPS binding[POn(0

energies [eV] of P2p electrons in structures of the type

R-)m]y" with mean formal charge per O atom (-y/4)

[P04f pP03(OR)f [P02(OR)2]1" [PO(OR)3]°

n = 4, m = 0 n = 3, m = 1 n = 2, m = 2 n= l, m = 3

-y/4 = -0.75 -y/4 = -0.5 -y/4 = -0.25 -y/4 = 0

P043"inNa3P04:132.4 eV

HPO42" in Na2HP04:

133.1 eV

H2PO4" in NaH2P04:134.2 eV

H3PO4:-135 eV

P043"inNa3P04:132.4-132.5 eV

P2074- (P03 52") in

Na4P207:133.2-133.4 eV

H2P03~inNaH2P03:134.2-134.4 eV

P025 inP4O10:

135.3-135.8 eV

The experimental binding energies, ER, of chemical moieties of type

[MOn(OR)rn]y* generally follow a systematic rule: EB increases as the ratio n : m of

'free' O ligands (n) to covalently bound OR ligands (m) is increased stepwise from

4 : 0, to 3 : 1, to 2 : 2, etc. This is a consequence of a systematic dependence of the

partial charge of the M atom on its chemical environment (nearest and next-nearest

neighbor atoms). In our case, M corresponds to P and R is either H or C. This

incremental increase for metal phosphate [Ores, 1979] and POx structures

[Moul,1992] is typically about 1 eV (Table 17).

118


The experimental value of P2P3/2 at 134.7 eV (see Table 1, chapter III 1.2.1)for the free acid (bulk ODP04 powder) agrees with the general trend in Table 17,

being closest to the [PO(OR3)]° ('free acid') case. The corresponding value of

134.2 eV for the ODP04 SAM, however, is definitely lower, suggesting a change of

the chemical structure of the phosphate head group upon coordination to the

surface. The value of 134.2 eV lies close to the reference values for [P02(OR)2]".

However, the P2P electron binding energy is further affected by the observed chargetransfer of approximately 0.5 eV between substrate and adlayer (see Table 11,

chapter IV 2.1.1.3). Assuming no charge transfer (i.e. Ta4f7/2 at 26.4 eV as in the

case of the bare Ta205), the position of the P2p binding energy would be shifted in

the direction of the reference value for [P03(OR)]2\ It cannot be excluded,

therefore, that both [PO(0")(OH)(OCi2H37)]" corresponding to [P02(OR)2]" and

[PO(0")2(OC]2H37)]2" corresponding to [P03(OR)]2~ may coexist at the surface, and

in that respect this is not in contradiction to our preferred model of both

monodentate and bidentate coordination of the alkane phosphoric acid head group

to tantalum cations (Figure 20) as presented and discussed in detail in the following

chapter. However our view is that it is not possible to draw a final conclusion justbased on the XPS P2P signal, and further experimental evidence for one or the other

binding model is needed (e.g. XPS 0]S binding energies and ToF-SIMS data, see

below). What can be definitely excluded based on the observed P2p binding energy

of the ODPO4 adlayer is the presence of a substantial concentration of the free acid.

In such a case the P2p signals would have to be clearly different from the

experimentally observed values.

The P2p signal does not, in fact, show any evidence of asymmetry due to

different chemical states although the P2p binding energies in the type A and type B

environments (Figure 20) would be expected to be different by 1 eV or so. We

believe that the reason is intermolecular hydrogen bonding within the SAM layer,

leading to partial charge transfer between adjacent phosphate groups and a levelingof the differences in partial charge on the P atom in the type A and type B

coordinations.

119


R K

0 ,o-\ //

P

/ \0" -o

00

H \ //

^0—p^0

R

X ,°P CK

o

0_Ta—0_ —0_Ta_o

Fig. 20: Bidentate (type A, left) and monodentate (type B, right) phosphatecoordination to tantalum ions, with the possibility for the formation of

intermolecular hydrogen bonding.

While the Ojs signal of the ODPO4 bulk powder shows two different chemical

states at EB - 532.1 (O type 1) and 533.6 eV (O type 2), respectively (Table 1), the

Ois spectrum of the ODPO4 SAM (Figure 14, Table 11), shows a third component,

due to O from the Ta205 substrate at EB = 530.7 ±0.1 eV (O type 3). Table 18

summarizes the quantitative XPS Ö]S data obtained. Different models of ODPO4

complexation to the surface have been tested with respect to how closely theyreflect the observed experimental data. These include monodentate coordination of

two ODPO4 units (two C18H37OPO2OH") to one Ta cation (type B bond in Figure

20), for which the expected atomic ratio 0(l)/0(2) is 2 : 2; bidentate coordination

of one ODPO4 (C18H37OPO32") to one Ta cation (type A bond in Figure 20), for

which the expected atomic ratio 0(l)/0(2) is 3 : 1; and finally, tridentate

coordination, previously proposed for the coordination of phosphoric acid carriers

to titania by Busca et al. [Buse, 1989] and for which the expected atomic ratio

0(l)/0(2) is again 3:1. None of the simple models assuming a single type of

coordination of the phosphate head group to Ta(V) cations is in close agreement

with the experimental findings (Table 18).

The data listed in Table 18 show the best agreement with a coordination

regime based on a mixed monodentate and bidentate binding of the head group to

the Ta(V) cations. The experimental atomic ratio of the two different O atoms

120


0(l)/0(2) is 1.60 ± 0.20 (average of 11 individual measurements at emission

angles of 15 and 75°) and thus is consistent with the mixed model, for which the

theoretical value is 7:5 or 1.40. Another possibility that would formally be

consistent with the observed 0(l)/0(2) ratio is a mixture of tridentate (type C) and

monodentate (type B) coordination. This, however, is less likely from the point of

view of the preferred coordination number of Ta(V) being 6 or 7, rather than 8.

The difference in EB between 0(1) and 0(2) is somewhat smaller than

expected from the reference data [Gres,1979] (1.7 - 1.9 eV); again we believe this

to be due to intermolecular hydrogen bonding as already discussed above in the

context of the P2p signals.

Table 18: XPS binding energies EB for ODPO4 in bulk form (free acid) and as a

SAM on a Ta205 substrate. Comparison with different theoretical models

for the coordination of the phosphate head groups to tantalum cations at

the oxide surface.

Experimental data Ois EB for O

of type 1

[eV]

OjSEB for O

of type 2

[eV]

Ois EB for O

of type 3

[eV]

Atomic Ratio

ofO(l)/0(2)

ODPO4 bulk powder(0 = 45°)

532.1 533.6 - 0.33 ±0.03

(theor. 0.333)

ODPO4 SAM on Ta205

(0= 15 and 75°)

531.8 533.1 530.7 1.6 ±0.2

Reference data 531.7-532.1 533.1-534.3 530.6-530.8 -

Model calculation for different coordination regimes:

ODPO4 type A coordination (bidentate)" 3 : 1-3.0

ODPO4 type B coordination (monodentate)2 2:2=1.0

ODPO4 type C coordination (tridentate) 3 : 1 = 3.0

ODPO4 type A (1 mol) plus type B (2 mol) coordination 7:5 = 1.40

From literature data [Gres,1979; Moul,]992] and references therein.

"

See Figure 20

121


2.1.1.5 Molecular Model of ODPO4 SAMonTa205

In order to construct a reasonable model of the ODP04/Ta2C>5 system, the

following observations need to be accounted for:

a) The NEXAFS [Brov,1999] evidence for chain order and an average tilt angle,

\), of 30-35°.

b) The AFM evidence of local nearly hexagonal order.

c) The ToF-SIMS evidence for P-O-Ta bonding.

d) The ToF-SIMS evidence for coordination of more than one phosphate to a

single tantalum.

e) The XPS evidence for the tails-up orientation, the possible charge transfer

from substrate to ODPO4, an adsorbed layer thickness of 2.2 nm, the possible

presence of both [P03(OR)]2" and [P02(OR)2]" species, and the inability of a

single type of coordination to account for the observed ratio between different

oxygen environments.

The Ta2Os coating was deposited by physical vapor deposition and is

(according to the manufacturer) semi-crystalline to amorphous. However, even for

an XRD amorphous structure, short-range order is likely to be present, with the Ta

cations in preferred oxide coordination symmetries. The further discussion is based

on the assumption of preferred coordination of Ta cations and a short-range order

deduced from considerations of the structure in crystalline Ta2C>5.

The combination of XPS, ToF-SIMS, AFM and NEXAFS [Brov,1999] results

is believed to constitute conclusive evidence for the presence of ordered

monolayers of ODP04 tilted by an angle, d, of 30 - 35° relative to the surface

normal and for a coordination regime that involves both monodentate and bidentate

direct binding of the phosphate head group to the tantalum cations at the surface of

the Ta205 substrate. Assuming a certain degree of (short-range) order in the oxide

substrate, a model of packing of the phosphate groups on top of the octahedral Ta

ion sites with hexagonal structure is proposed. Ionic radii of 0.14 nm for O"2 and

0.064 nm for Ta+5, an O covalent radius of 0.125 nm and bond lengths of 0.15 -

122


0.16 nm for P=0 and P-O-R(H) have been assumed. There is a particular

arrangement of the phosphate head groups on the square tantalum oxide lattice that

satisfies the assumption of monodentate and bidentate ODPO4 coordination (in the

molecular ratio of 2 : 1) and at the same time leads to an approximately close-

packed phosphate ligand ordering at the surface as shown in Figure 21. The AFM

study provides direct evidence for such a nearly hexagonal, nearly close-packed

adlayer although the order is localized to rather small regions, possibly due to the

generally non-crystalline nature of the surface where only local order can be

expected. Tantalum cation sites not coordinated to phosphate are likely to be linked

to oxygen, hydroxide or water molecules to complete the coordination sphere. For

the sake of clarity this is only partly shown in Figure 21 (as 'free' surface oxide).

O01f m ü mi t /

\ / v» m I V J

O Ta O 0 0 0 Position of alkyl(oxide of ('free'surface (P-Q- coord (P=Qor (P-Q-R) chain attachment

Ta205) oxide) to Ta) P-QH)

Fig. 21: Schematic, idealized view of the arrangement and orientation of

phosphate groups of ODPO4 at a Ta205 surface (with square substrate

lattice). The phosphates are bound to the Ta ions through either

unidentate or bidentate coordination. The P-O-R groups form a nearly

perfect hexagonal lattice with a mean distance between the hydrocarbon

(R) chains of approximately 0.46 nm (corresponding to a specific area of

0.21 nm2 per molecule).

123


A possible arrangement of a row of ÜDPO4 molecules is shown 111 figure 22,

based on a ball-and-stick model fliere is a particular arrangement that allows

hydrogen bonding between two adjacent ODP04 molecules (type A and B,

respectively, see Figure 20) while keeping the hydrocarbon chains at the distance of

approximate!) 0.5 nm, known to be a favorable distance for strong mtermolecular

Van der Waals interactions in long-chain, alkane-based SAMs The tilt angle, v, of

the hydrocarbon chains m the molecular model arrangement of figure 22 is

approximately 30°, as calculated from experimental NEXAPS measurements

[Brov,l999| The parallel ordering of the hydrocarbon chains can be achieved b> a

single gauche 0-CH-.-CH2 conformation of the bidentate ODPO4 molecules

(type A), while the adjacent monodentate ODPOj molecules (type B) have

exclusively trans conformations A consequence of such a conformational

arrangement would be a slight difference m the height level of the terminal methyl

group of ODPO4 molecules A and B respectively This effect is Jikefy to be too

small to be detected by AFM, however

Fig 22 Ball-and-stick model of six adjacent ODPO t molecules w ilh monodentate

and bidentate coordination (111 the ratio of 2 1) to the ITbOs substrate

surface (see Figure 20 for binding of the phosphate group to the

substrate) Only ten carbon atoms of the alkane chain are shown for the

sake of simplicity The tilt angle, u, of the hydrocarbon chain is

approximate!) 30° relative to the surface normal

124


The hydrocarbon chains - attached to the phosphate group at the dark-colored

spots in Figure 21 - form an approximately hexagonal pattern. Assuming an

idealized square substrate (Ta205) lattice of dimensions 0.28 x 0.28 nm2, a [4 x 2]-

overlayer coincidence lattice can be formed. Within this overlayer lattice of

dimension 0.63 nm,3 ODPO4 molecules can be accommodated with each ODPO4

molecule formally occupying an average area of 0.209 nm" at an average

intermolecular distance (parallel to the surface) of 0.49 nm. The intermolecular

distances and area of occupancy per SAM molecule according to AFM and model

calculations are listed in Table 19, together with corresponding literature values for

alkane thiols on gold and octadecyl phosphonic acid on mica.

Table 19: Comparison of structural parameters of ODPO4 on tantalum oxide,

ODPO3 on mica and alkane thiols on gold.

SAM on

substrate

Length1/

[nm]

Thickness2

d

[nm]

Angle3i)

n

Area4

A

[nm2]

Distance5

d'

[nm]

Used technique

ODPO4

on Ta205

2.5 2.2 ±0.3 30-35 0.209 ±

0.04

0.49 ±

0.01

NEXAFS, AFM,

XPS, Molecular

model, SIMS

ODPO3

on mica

2.5 1.8 ±0.2 -40 0.188-

0.220

0.47-

0.506Isotherm, AFM

[Wood, 1996a;

Gao,1996]

Alkane

thiols on

gold

(C6-C22)

1.0 (C6)1.5 (C10)1.7 (C12)2.2 (C16)2.5 (C18)3.0 (C22)

-0.9

-1.3

-1.5

-1.9

-2.2

-2.6

-30

(for all

alkane

thiols)

0.214

(for all

alkane

thiols)

0.506 Contact-angle,FTIRAS, XPS

[Bain, 1989]

1Total length / of extended molecule

2Thickness d of SAM layer (measured perpendicular to surface)

1Average angle v between axis of molecule and surface normal

4Average area A occupied per molecule

5Average intermolecular spacing d' between adjacent molecules (measured parallel to the surface)

6Assuming a hexagonal arrangement of assembled molecules

125


The structural geometric parameters found for the ODP04/Ta20s system are

indeed very close to those reported for alkanethiols on gold [Ulma,1991] as well as

for octadecyl phosphoric acid on mica [Wood, 1996a; Gao,1996] (the adlayer of the

latter is, however, chemically not stable). For the alkanethiol/gold system there is

general agreement that the lateral periodicity of the SAM is directly linked to the

periodicity of the gold surface structure with the terminal sulfur occupying hollow

sites in the gold substrate lattice [Ulma,1991; Feut,1993]. Since the size of the

alkanethiol molecule is too large to occupy every hollow site, an overlayer is

formed that conforms with the steric requirements of the molecule. On Au(lll),the most extensively studied single crystal plane, a V3xV3 R30° overlayer structure

is formed. The observed specific tilt angle of the alkanethiol molecules is believed

to result from a maximization of the Van der Waals attraction between adjacentalkane chains within the SAM, leading to the energetically most favorable

conformation.

It is tempting to use argumentation based on the same or similar principles as

in the gold/thiol system to explain the (local) order observed for the ODP04/Ta205

system. In our proposed model, the periodicity of the Ta cations is again believed

to be the prime factor determining order in the ODPO4 adlayer. In view of the

extremely high strength of the Ta-0 bond (approximately 800 kJ/mol) and the

generally high strength of transition metal-phosphate bonds, one expects deeppotential wells for the phosphate at the Ta cation coordination sites. There are,

however, two main differences between the two SAM systems: (a) there is no

simple coincidence lattice for phosphate groups positioned directly on top of the

cation sites that results in a close-packed adlayer and at the same time maintains a

realistic intermolecular phosphate-phosphate distance; (b) for the phosphate

groups, on the other hand, there are several possible geometries to coordinate to

cations: one-, two- and three-fold which offers more flexibility as regardsgeometric orientation of the phosphate group relative to the cation lattice. Our

preferred model of an adlayer binding based on both monodentate and bidentate

phosphate coordination results in a hexagonal overlayer lattice geometry with an

intermolecular spacing that is almost exactly the same as the one in the thiol/goldsystem. It is then straightforward to assume that the experimentally observed mean

tilt angle, "0, of the ODPO4 molecules of 30 - 35° (relative to the surface normal)

again, results from a maximization of the Van der Waals attraction of the

hydrocarbon chains, as in the gold/thiol case.

126


Order in the ODP04/Ta205 system, as seen in our AFM results, is restricted to

areas of a few nm2 (much smaller than is observed in the case of alkanethiols on

gold). This may be due to the fact that the Ta2Os film is of nanocrystalline to

amorphous nature. If the assumption that the Ta cation lattice geometry is

responsible for adlayer order is correct, then it is straightforward to assume that

order in the ODP04/Ta205 system can only extend over areas comparable to the

size of the nanocrystals. Areas that show no order in the AFM investigation would

then correspond to entirely amorphous Ta205 regions, completely lacking

periodicity in the cation lattice, or to nanocrystalline patches that are too small to

induce measurable order in the adlayer. Nevertheless, NEXAFS data [Brov,1999]

suggest that the overlayer maintains its order over larger distances, essentially

bridging the gap between the ordered tantala patches. A second possibility could be

linked to the different crystallographic planes of the nanocrystals exposed at the

surface, not all of which would be expected to have the appropriate symmetry to

induce order in the ODPO4 SAM. Still another explanation could be the presence of

strongly coordinated ODPO4 molecules with a direct phosphate-Ta(V) bond in

ordered areas surrounded by areas with more weakly bound ODPO4 molecules, e.g.

weakly bound to the oxide surface through hydrogen bonding rather than

chemisorbed by direct coordination. AFM and NEXAFS studies on different single

crystal planes of transition metal oxides would be needed to determine which of

these factors is chiefly responsible for limiting the AFM-observed order to small

areas.

Regarding our proposal for the structure of the ordered ODPO4 areas, other

models, which may be in accord with some of the experimental findings are, of

course, feasible. Our belief in the proposed model is not based on singleobservations, but on the sum of the information from the various techniques appliedto characterize the surface and interface composition and structure. Also, from

purely chemical considerations, the two proposed coordination structures of the

phosphate/ Ta20_s interface may actually make more sense than alternative models,

since:

a) Both types of coordination proposed (A and B, see Figure 20) satisfy formal

charge considerations, in the sense that formally an oxide O2" is replaced byeither two single-coordinated anions of charge -1 or by one bidentate anion of

charge -2. In view of the high O-Ta bond strength (mostly ionic character) of

ca. 800 kJ/mol, it seems very important that a similar gain in enthalpy is

achieved through replacement of the oxide and coordination of the phosphate

127


in order to get a thermodynamically stable adlayer. Therefore coordination

involving anions of the same total negative charge should be preferred over

e.g. single coordination to an anion with charge -1.

b) Taking the simplified model of an octahedral structure of Ta2Os (edge- and

vertex-shared octahedra), the coordination number of the coordinated surface

Ta ion turns out to be 7 for both proposed types A and B of surface bonding.Seven is (in addition to six) a preferred coordination number for Ta(V). The

often proposed tridentate coordination of the phosphate group would, on the

other hand, lead to a coordination number of eight, which we believe is less

likely, in particular for steric reasons.

c) The combination of type A and B coordination bonding at adjacent sites

allows for the formation of hydrogen bonding, which may be important for

further stabilization of the monolayer. Most other models, such as pure

tridentate phosphate coordination or pure type A (bidentate) bonding, do not

provide the possibility for stabilization through hydrogen bonding.

d) Polymerization across the phosphate intermediate layer through condensation

of phosphate groups to form structures such as (R-OP02)oo (similar to NaP03)is unlikely for both thermodynamic and kinetic reasons, since a) there would

be no strong bond of sufficient ionic character (only P=0 Ta type of

coordination) and b) in contrast to silanes, which very easily form polymericsiloxane structures through condensation, phosphates are kinetically much

more inert. Furthermore, no fragments indicative of polyphosphates were

observed in the ToF-SIMS data.

e) Indirect (weak) coordination of the phosphate to the tantalum oxide or

hydroxide surface via hydrogen bonding is believed to be less likely than the

presence of strong phosphate-Ta(V) complexes. Tn the former case the surface

bonds are likely to be too weak to survive as complex secondary ion

fragments. Moreover, the SIMS fragmentation pattern can only be fullyunderstood when assuming direct complex coordination. However it cannot be

excluded that (in localized regions) less strongly bound states are also present,

e.g. hydrogen-bonded phosphoric acid molecules. In fact, one possible

explanation for the local lability of the SAM order as observed by AFM may

be the presence of such weaker surface interactions in localized areas.

128


In terms of reaction mechanism we have no direct evidence for a particularmechanism of surface complex formation. However, in view of the very strong

O-Ta bond, it is not likely that 'free' Ta ions are available at the surface under

ambient conditions. Rather, we assume that an oxide ion has to be replaced by

phosphate(s). Again this is not likely to be directly possible, and we assume that a

low activation energy pathway is only feasible through intermediate structures such

as those proposed in Figure 23 and 24.

H

O O

Il +H+ +1 +ROPOHCV

__0__Ta^0 ^ — 0__Ta__0_ ^

h<w°-R V

H\ /\ A /- \\+/ -HoO '-HP

0___Ta__0__ ^ ___0_^_Ta_0_

Fig. 23: Proposed reaction sequence for the displacement of oxide ligands at the

Ta205 surface by alkylphosphates through intermediate hydroxideformation.

In fact, hydroxylation (or protonation of oxides) has been shown to be an

important initial reaction step prior to the SAM formation of phosphonic acid esters

on alumina. It has been demonstrated that fast adsorption in this system only takes

place if a minimum fraction of hydroxides is present on the alumina surface

[Maeg,1997; Bram,1998; Jung, 1998]. The fact that solvent choice and pH are

important factors for the kinetics of SAM formation may be related to issues of

129


protonation or hydroxylation and to the existence of intermediate surface-

coordinated species as discussed above.

A condensation reaction of ODPO4 (Figure 24) with hydroxy groups at the

Ta205 surface is preferred to a ligand exchange reaction, since in the case of ligand

exchange the substituted negatively charged hydroxy ligand would have to leave

the system and dissolve in the organic non-polar solvent, which is not

advantageous. The new ligand, phosphoric acid, is in its non-charged, acidic form

in solution. This is a reasonable condition for condensation and formation of the

tantalum - phosphate complex.

,R R/ /

? r«crtT50 Hcff~° {

wfX~°

fHO HO

(?H \pH (pil

-3 H 2°

(I (|> \3

Ta2O5 Ta205

Fig. 24: Drawing of the suggested condensation reaction of ODPO4 with the

hydroxy groups at the TaÔ.s surface.

130


2.1.2 Mono- and Bifunctional Phosph(on)ates

Surface modifications with the bifunctional 12-hydroxy dodecyl

phosphate (OH-DDPO4) and 12-(N-ethylamino) dodecyl phosphonate (NEt-

DDPO3) were studied on Ta205, Ti02 and on titanium metal surfaces.

Monofunctional amphiphiles were investigated in parallel for different

reasons: dodecyl phosphonate (DDPO3) as an adsorbent with the same

hydrocarbon length as the bifunctional phosph(on)ates; octadecyl

phosphonate (ODPO3) and octadecyl phosphate (ODPO4) as long-chainreference amphiphiles and as tests for potential differences in behaviour

between phosphates and phosphonates, and finally biphenyl phosphate

(BIPH) as an arylic amphiphile.

2.1.2.1 Applied Methods

Microdroplet density (fidd) and contact angle:

The microdroplet density (jidd) was measured as described in chapterIII 2.2.4. The îdd was measured both before and after rinsing the SAM-

coated surfaces with water. The average of 5 to 10 measurements over

different regions of each surface was determined. The data were used to

estimate the SAM homogeneity on a micrometer scale in the lateral

dimension. Even if the hydrophobicity (from contact-angle measurements) is

not indicative of highly ordered SAMs, as in the case of bifunctional

amphiphiles, the contact-angle results may give some interesting information

in combination with the jidd-data,

Time-of-flight Mass Spectrometry (ToF-SIMS):

Secondary ion mass spectra of the amphipbile-treated metal and metal

oxide surfaces were recorded. Data were correlated to the ToF-SIMS results

for ODPO4 on Ta205 and provided information on the type of chemical bond

between the phosph(on)ate head group and the metal oxide substrate.

131


X-ray photoelectron spectroscopy (XPS) analyses:

Angle-resolved XPS (AR-XPS) measurements were conducted at two

different take-off angles, namely 15° and 75° with respect to the surface

plane, to get depth-dependent information and to determine the thickness of

the monolayer deposited on the metal oxide substrate (using the 2-layermodel, presented in chapter IV 2.1.1.2). The oxygen signals of the different

phosph(on)ate/metal oxide combinations were fit and used to test the

applicability of the phosphate - metal oxide bonding model developed for

the ODP04/Ta205 system.

Near-edge X-ray absorption fine structure spectroscopy (NEXAFS):

NEXAFS measurements were performed by G.Hähner and I.Pfund-

Klingenfuss on different amphiphile/metal oxide combinations to investigatethe order and chain orientation of the phosph(on)ate SAMs.

2.1.2.2 Influence of Different Solvents [Folk,1995; Dann,1998J

The hydrophobicity and the homogeneity of the SAMs were

investigated using contact-angle and jidd data, respectively. A few solvents

were tested to investigate the self-assembly process of the mono- and

bifunctional phosph(on)ates and to find out the best solvent for each of the

amphiphiles. The criteria for the solvents were the solubility of the

amphiphiles and the toxicity of the solvents: The solvents with the best

performance for ODPO4, n-heptane/2-propanol 100:0.4 (No 1) and ethylacetate (No 2) were tested for all amphiphiles. For DDPO3

(n-heptane/2-propanol 100:0.8 (No 1')) as well as for ODPO3 (n-heptane/2-propanol 100:1.5 (No 1")) a slightly higher amount of 2-propanol was

needed to completely dissolve the amphiphiles at 0.5 mM concentration.

Ethanol/water 1:1 (No 4) and methanol/water/ethanol 10:10:1 (No 6) were

chosen as highly polar mixtures. Ethyl acetate mixtures with different

additives such as ethyl acetate/methanol 10:1 (No 5), ethyl acetate/ethanol

10:1 (No 7) and ethyl acetate/methanol/acetic acid 10:l:traces (No 8) were

132


solvents with acceptable solubility. Data from a solution of ODPO4 in ether

(No 3) are included in the list even though the solvent was not used for the

other amphiphiles. To provide a better overview, the solvents are listed in

Table 20.

For the monofunctional amphiphiles (ODPO4, ODPO3 and DDPO3) and

for the aromatic phosphate (BIPH) without a terminal polar functionality,measurement of the contact angle is a good method to determine the

performance of the corresponding solvents by measuring the hydrophobicityof the monolayer surfaces. For a highly oriented and densely packedmonolayer of the bifunctional amphiphiles (OH-DDPO4 and NEt-DDP03) a

higher wettability is expected, due to the polar terminal function. Advancingwater contact-angle data are listed in Table 21.

The jidd data (Table 22) may hint at the homogeneity of the SAM

surfaces independent of their wettability. High (idd data are suggestive of a

high amount of adsorbed clusters providing condensation nuclei for water

vapor at the surface. Washing with water may "clean" such surfaces

resulting in a reduction of the u;dd.

Table 20: List of solvents used for preparing amphiphile solutions

No 1 n-heptane/2-propanol 100:0.4 (v/v)

No 1' n-heptane/2-propanol 100:0.8 (v/v)

No 1" n-heptane/2-propanol 100:1.5 (v/v)

No 2 ethyl acetate

No 3 ether

No 4 ethanol/water 1:1 (v/v)

No 5 ethyl acetate/methanol 10:1 (v/v)

No 6 methanol/water/ethanol 10:10:1 (v/v/v)

No 7 ethyl acetate/ethanol 10:1 (v/v)

No 8 ethyl acetate/methanol/acetic acid 10:1 :traces (v/v)

133


Table 21: Contact-angle data for SAMs on Ta205, Ti02 and Ti (covered by a

native oxide layer) deposited from solutions of different solvents

(see text): The corresponding amphiphile concentrations are

mentioned in Darenthes is .

Adsorbent (concentration) Solvent

No

Advancing water-contact angle of SAMs

Ta205 Ti02 Ti

ODPO4 (0.5 mM) 1 113.7 ±0.2 110.0 ±0.5 115.4 ±0.4

ODPO4 (0.5 mM) 2 113.6±0.4

ODPO4 (0.05 mM) 3 107.9 ±1.2

ODPO4 (0.05 mM) 4 103.5 ±1.5 104.9 ±1.4

ODPO3 (0.5 mM) 1" 111.6± 0.7

DDPO3 (0.5 mM) 5 97.0 ± 1.3 100.7 ±0.8 110.1 ±0.71

DDPO3 (0.5 mM) 2 106.9 ±1.5

DDPO3 (0.5 mM) 1* 111.9 ± 1.5

DDP03(1.8mM) 2 108.2 ±0.6

DDPO3 (0.05 mM) 6 94.2 ±1.4 104.8 ±1.8

DDPO3 (0.005 mM) 1 107.2 ±0.7

OII-DDP04 (0.5 mM) 7 79.1 ±2.8 72.4 ±1.7 77.3 ± 0.5

OH-DDP04(1.3mM) 2 73.1 ±1.1

OH-DDPO4 (0.05 mM) 6 78.3 ±1.6 71.3 ±1.9

OH-DDPO4 (0.001 mM) 1 49.3 ±1.7

NEt-DDP03 (0.5 mM) 4 74.1 ±1.5 65.9 ±0.6 74.0 ±0.8

NEt-DDP03 (0.5 mM) 2 50.7 ±1.1

NEt-DDP03 (0.05 mM) 6 71.5 ± 1.8

NEt-DDP03 (0.001 mM) 1 55.5 ± 1.7

BIPH (0.4 mM) 2 68.9 ±1.8

BIPH (0.25 mM) 8 65.1 ±1.6 64.4 ±1.2 78.9 ±0.5

BIPH (0.05 mM) 6 52.4 ±1.9 70.3 ±1.3

BIPH (0.03 mM) 1 55.0 ±3.7

134


The contact-angle data for SAMs on Ta205 were used to screen the

solvents.

Although the solvent mixture No 5 (ethyl acetate/methanol) did not

result in the most hydrophobic DDPO, SAM, its overall behavior including

transparency of the adlayer and solubility of the phosphonate was better than

for the other solvents.

Solvent No 7 (ethyl acetate/ethanol) was selected for OH-DDPO4 and

solvent No 4 (ethanol/water) forNEt-DDPCV

BIPH was, in general, difficult to dissolve at a 0.5 mM concentration,which was the target concentration for all of the solvents. Using pure ethylacetate/methanol 10:1 or adding of a few micro liters of acetic acid to a

suspension of BIPH in 50 ml acetate/methanol 10:1 did not lead to a

complete dissolution of the amphiphile. In the second case it may be due to

an excess of organic acid. Only if traces of acetic acid were added to the

colloidal solution (e.g. by putting an acid-wetted tissue in the flask neck for

a few minutes) could a 0.5 mM solution be achieved.

Pure ethyl acetate solutions did not give reproducible results and turned

out to be unstable. Sometimes the dissolution of a particular amphiphileworked and sometimes it did not. A reason for this behavior could not be

found.

Table 22: Microdroplet density (|idd) data from phosph(on)ate SAMs on

different substrate surfaces. The self-assembly was performedwith amphiphile solutions in particular solvents (see text). The

samples were measured before and after rinsing with water.

Adsorbent Solvent

N°

(add before/after rinsing with water

Ta20, Ti02 Ti

ODPO4 (0.5 mM) 1 220/130 580/250 800/210

DDPO3 (0.5 mM) 5 350/250 1790/120 850/350

OH-DDPO4 (0.5 mM) 7 1890/90 3000/130 1380/510

NEt-DDPOi (0.5 mM) 4 840/110 1290/790 1820/690

BIPH (0.25 mM) 8 2200 / 820 2800/1250 1330/710

135


The additional rinsing step with pure water resulted in a generalreduction of the |idd, suggesting a higher order and homogeneity of the

adsorbed adlayer after this rinsing step. The removal of adsorbed amphiphileclusters and/or micelles is expected to reduce the average adlayer thickness.

A slight decrease of the phosph(on)ate layer thickness by about 0.1 nm in all

the combinations could be determined by AR-XPS (results not presented in

detail).

2.1.2.3 SAM Orientation and Order

Contact-angle data allow for a first estimation of orientation and order

in a SAM: Hydrophobicity of SAMs formed from monofunctional

phosph(on)ates is due to a tail-up orientation, where the hydrocarbon chains

are oriented away from the substrate and the molecules are attached by the

polar head groups. NEXAFS investigations on ODPO4 SAMs showed order

for monolayers with contact-angle values of 110° ± 2° or higher, indepen¬dent of the metal oxide substrate that was used for the self-assembly process.

For the bifunctional amphiphiles the contact-angle data give less

information. Here it is necessary to get more data from ToF-SIMS, XPS

and/or NEXAFS.

For the studies presented in this section, amphiphile solutions with a

0.5 mM concentration were prepared in the solvents listed in Table 20. The

solutions were used for self-assembly immediately after preparation.

Ta205, TiOa and Ti chips were cleaned in 2-propanol in an ultrasonic

bath followed by 30 minutes of UV cleaning. Chips were transferred into

glass containers, the amphiphile solutions were added, and the samples were

immersed for 48 hours. Samples were removed, rinsed with the pure ethylacetate, blow-dried with nitrogen and stored in a plastic container until

analysis.

ToF-SIMS (negative-ion spectra from all samples and positive-ionspectra from NE1-DDPO3 samples) and AR-XPS (at 15° and 75° with

respect to the surface) measurements were performed.

136


ToF-SIMS:

The most prominent secondary ion masses observed in the static SIMS

spectra are listed in Tables 23 a-f (negative ions). The mass of an entire

amphiphile combined to a metal oxide fragment could not be detected in all

the combinations. Nevertheless, a few fragments of phosph(on)ate/metal ion

and organo-phosph(on)ante/metal ion were present.

From the negative ToF-SIMS spectra showing a variety of fragments

corresponding to the classes II and III (see section IV 2.1.1.3), the bindingmodel can not be proven. Nevertheless, the strong interaction between the

phosph(on)ate and transition metal cations as Ti and Ta is clearly shown bythe presence of such fragments. Furthermore, the detected fragments have

metal-to-phosphoms ratios of 1:1, 1:2, 2:2 and 2:1, which are in agreement

with the structure suggested in Figure 25.

As in the case of ODP04/Ta205 SAMs (see section IV 2.1.1.3), the

prominent peaks of classes II (MnPmOxHy) and III (MnPm0xCyH7) supportdirect coordination ofthe phosph(on)ate group to Ta and Ti cations.

Complete tables of all detected fragments (positive and negative ions)and of all phosph(on)ate/metal oxide combinations are listed in the appendix

(App. 1-13).

The assignment of the masses to molecular ion species is discussed on

the basis of the classes of fragments:

Class I: CnHmPOxions: Species are observed starting from the pure

phosphate fragments (e.g. PCV, HP04", H2P04") through various partlyfragmented phosphoric ester species up to the molecular masses ([M-H]").As expected, reduction of the phosphate (oxidation state +V) to the

phosphonate (oxidation state +III) partly takes place.

Class II: MnPmOxHvions: Species coiTesponding to formal stoichiometrics of

Ta(^V), Ta(+TU), Ti^1 -1 and TiHI) - the most stable oxidation states of inorganictantalum and titanium compounds - tend to have higher intensities comparedto the others, a rale that is often observed in SIMS spectra of metal oxides

and hydroxides.

137


Class III: MnPmOxCyHy ions: The presence of strong emissions of metal

phosphates and phosphonates from the modified surface gives good indica¬

tion that the tantalum and titanium ions of the metal oxide layers are actually

directly bound to the chelating phosphoric or phosphonic acid group. It is

unlikely that complex metal oxide/phosph(on)ate species would form with

significant intensity, if there were no direct complexation between the

phosph(on)ate group and the metal cation. In particular, assuming that the

phosphoric and phosphonic acid head groups bind to the metal oxide surface

via hydrogen bonding, species consisting of both metal oxides and the

phosphoric (or phosphonic) acid group are unlikely to survive the emission

process without further fragmentation into pure metal oxide and phosph-

(on)ate species.

Table 23 a: ToF-SIMS class 1 (POnCmHv) fragments (negative ions) of alkyl

phosph(on)ates adsorbed on Ta205 (x = fragment present;xx = complete alkyl chain present).

ODP04 DDP03 OH-DDP04 NEt-DDP03 BIPH

P02 X X X X X

P03 X X X X X

P04H X X X X X

P04H2 X X X X X

CH3P03 X X

C5H7P03 X

C6H14P03 X

C11H18P03 X X

C12H27P03 XX

C2H4PO4 X X X X

C,8H38P04 XX

Fragments of class I simply prove the presence of adsorbed phosph(on)-ates at the metal oxide surfaces, TfaOs or T1O2 (Tables 23a and b).

138


Table 23 b: ToF-STMS class I (POnCmHx) fragments (negative ions) of alkylphosph(on)ates adsorbed on TiOi (x = fragment present;xx = complete alkyl chain present).

ODP04 DDPOa OH-ÜDP04 NEt-DDP03 BIPH

PO X X X X X

P02 X X X X X

P02H X X X X

P03 X X X X X

P03H X X X X

P04H X X X X

P04H2 X X X X

PO4NII2 X

CH2PO X X X X

CH3NC12H18PO X

C9H14P02 X

C12H26PO2 XX

C2H4PO3 X X X

C2H7P03 X X

C3H5PO3 X X

C5H10PO3 X X

C6H12P03 X X

C12H27PO3 XX X

NC12H28PO3 X

CH3NC2H27PO3 X

C2H5NC12H25P03 XX

CH3PO4 X X

C6H,4P04 X X

C12H11PO4 XX

C8H38PO4 XX

C2H5NC12H25P04 XX

139


Table 23 c: ToF-SlMS class IT (Ta„PmOxHy) fragments (negative ions) of

alkyl phosph(on)ates adsorbed on Ta2Q5 (x = fragment present).

ODP04 DDPOa OH-

DDP04

NEt-

DDPO3

BIPH ratio

(Ta/P)

TaO(P03) X X X X X 1:1

TaO(P03)H X X X X X 1:1

TaO(P04) or Ta02(P03) X X X X X 1:1

TaO(P04)H or

Ta02(P03)H

X X X X * 1:1

Ta02(P04) or

Ta03(P03)

X X X 1:1

Ta02(P04)H or

Ta02(OH)(P03)

X X X 1:1

Ta(P03)2 X X 1:2

Ta(P04)(P03) X X 1:2

Ta(P04)2 X X 1:2

TaO(P04H)2 X 1:2

Ta204(P03) X X X 2:1

Ta205(P04H2) X 2:1

Ta2(P03)2H X 2:2

Ta202(P04)2H X 2:2

Ta2(P03)2(P03H) X 2:3

ïa2(P03)(P03H)2 X 2:3

Ta202(P04)(P04H)2 X 2:3

Ta304(P04)2(P04H) X 3:3

140


Detected fragments with a ratio O/P of 1:1 and 1:2 were present in all

the combinations. The data agree with the model of bidentate and

monodentate attachment of the phosph(on)ate head-group onto the metal ion

(Tables 23c and d).

Table 23 d: ToF-SIMS class II (TinPmOxHy) fragments (negative ions) of

alkyl phosph(on)ates adsorbed on TiQ2 (x = fragment present).

ODP04 DDPOJ OH-

DDP04

NEt-

DDPO3

BIPH ratio

(Ti/P)

TiO(P02)H X X X X 1:1

TiO(P02)NH X 1:1

TiO(P03) X X X X X 1:1

TiO(P03)H X X X X 1:1

TiO(P04) X X X X 1:1

TiO(P04)H X X X X X 1:1

TiO(P04H2) X X X X 1:1

Ti02(P04)H X X X X X 1:1

Ti02(P04H2) X X X X 1:1

Ti(P03)(P02)N X 1:2

Ti(P03)(P03H) X X X 1:2

Ti(P04)(P03)H X X X X X 1:2

Ti(P04)2H X X X X X 1:2

Ti(P04H)2 X X X 1:2

Ti(P04H2)(P04H) X X 1:2

Ti2(P03)(P02) X 2:2

Ti203(P03)2 X 2:2

Ti20(P04)3 X 2:3

Ti305(P03)H X 3:1

Ti306(P03H2) X 3:1

141


Table 23 e: ToF-STMS class III (TaJ^CXQH,) fragments (negative ions) of

alkyl phosph(on)ates adsorbed on Ta205: (x = fragment present;

xx = metal/phosph(on)ate containing complete hydrocarbonmoiety present).

ODPO4 DDPOa OH-

DDPO4

NEt-

DDP03

BIPH ratio

(Ta/P)

Ta(CHP03) X X X 1:1

Ta(CHP04) X X X X X 1:1

TaO(CH3P04) X X X 1:1

Ta02(Ci2H9P04) XX 1:1

Ta02(C18H37P04) XX 1:1

Ta(P03)(CH2P03) X 1:2

Ta(P04)(CH2P03) X X 1:2

TaOH(P03)(CH3P03) X 1:2

Ta2(C5H4P04) X 2:1

Ta203(CH2P04) X 2:1

Ta202(HOC,2H21P04) XX 2:1

Ta2O(PO3)(CH2PO0 X 2:2

Again a series of fragments suggests the validity of the monodentate-

bidentate-model (Tables 23e and f). Coexistence of 1:1 bonding (1 metal

cation : 1 phosph(on)ate anion or 2 metal cations : 2 phosph(on)ate anions,

suggesting bidentate coordination) and 1:2 (1 metal cation : 2 phosph(on)ateanions, suggesting monodentate coordination) agrees with the model (formore details see also the previous chapter and Figure 25).

142


Table 23 f: ToF-SIMS class III (TinPmOxCyHz) fragments (negative ions) of

alkyl phosph(on)ates adsorbed on TiC>2 (x = fragment present;xx = metal/phosph(on)ate containing complete hydrocarbonmoiety present).

ODPOj DDP05 OH-DDP04 NEt-DDP03 BTPH ratio (TUP)

TiO(C2H3PO) X X X

TiO(C2H4PO) X X X

TiO(NCH3PO) X

TiO(CH2P02) X X

TiO(C2H2P02) X X

TiO(CH3P03) X X X X

TiO(C2H3P03) X X

TiO(C3H3P03) X X X

TiO(CH3P04) X X X

TiO(C2H4P04) X X X

TiO(C3H4P04) X X X

TiO(C,2H9P04) XX

Ti02(C2H3P04) X X

Ti02(C2H4P04) X

Ti(P02)(CHP03) X X X 1:2

Ti(P02)(CH3P03) X X 1:2

Ti(P02)(C2H,P03) X X X 1:2

Ti(P02)(NCH3P03) X 1:2

Ti(PO0(CH3PO3) X X X 1:2

Ti(PO,)(CH2P04) X X 1:2

Ti(P03)(C4H,,P04) X X 1:2

Ti(CH2P03)(C12H„P04) XX 1:2

Ti(P04H2)(CH3P04) X X 1:2

Ti(P04)(C2H3P04) X X 1:2

Ti(P04)(C4H13P04) X X 1:2

Ti204(HOC,2H29P04) XX 2:1

Ti20,(HOCl2HOTP04) XX 2:1

Ti206(CH2P04) X 2:1

Ti20(PO4H2)(CH3PO4) X 2:2

Ti20(P04)(C2H6P04) X 2:2

Ti305(CH2P03) X 3:1

143


XPS:

Angle-dependent X-ray photoelectron spectroscopy was performed on

the different self-assembled phosphates and phosphonates. In all cases, the

amphiphiles were shown to be oriented such that the polar head group was

adsorbed on the metal oxide surface.

The thickness of the organic part of the adlayer was calculated from the

attenuation of the photoelectron intensity of the Ta4r and the Ti2p- signals,(Table 24), using a two-layer model (for more information see chapter IV

2.1.1.2).

The measured SAM thicknesses of the monofunctional phosph(on)ates(DDPO4, ODPO3, ODPO4 and BIPH) on metal oxide substrates such as

Ta205 and Ti02 were found to be close to the expected value for a perfectmonolayer with an average tilt angle, t), of 33° between the molecules and

the surface normal (Figure 22). SAM thicknesses being significantly lower

than the calculated theoretical value for perfect monolayers were determined

for the bifunctional phosph(on)ates (OH-DDPO4 and NE1-DDPO3) self-

assembled on Ta205 or Ti02. The implied lower degree of coverage may be

due to the solvents used or due to the polar, end-standing functionality(hydroxy or N-ethylamino). These effects may prevent the formation of an

ordered, close-packed layer due either to steric repulsion which could

prevent molecules from reaching energetically preferred separation of ca.

0.5 nm, or to disadvantageous dipolar interactions.

In general, lower adlayer thicknesses (or lower degree of coverage)were determined for all amphiphiles adsorbed on metallic titanium (i.e.natural oxide film).

144


Table 24: Calculated SAM thicknesses in nm using the Ta4f or Ti2p data from

AR-XPS. The lengths of the extended single molecules were

determined using the CS Chem3D Pro software program. The

length was measured from the outermost oxygen atom of the

phosph(on)ate group to the terminal carbon atom (or oxygen atom

of the hydroxy group in case of OH-DDP04) after energyminimization to a minimum rms gradient of 0.100, considering the

bond order and the steric constraints. The theoretical thickness (d)of a complete monolayer was calculated with an assumed tilt angleof 33° with respect to the surface normal (see NEXAFS, chapterTV 2.1.1.3).

Amphiphile Measured adlayer thickness1 d [nm]on corresponding substrate

Molecule

length"/

[nm]

Calculated

thickness3

<f (33°)

[nm]Ta205 Ti02

r »

DDPO3 1.14 ±0.1 1.31±0.1 0.70 ±0.1 1.70 1.43

ODPO4 2.10 ±0.2 2.04 ±0.2 1.40 ±0.1 2.54 2.13

ODPO3 1.63 ±0.2 2.41 2.02

OH-DDPO4 0.86 ±0.1 0.95 ±0.1 0.79 ±0.7 1.90 1.59

NEt-DDP03 0.85 ±0.1 1.01 ±0.1 0.77 ±0.1 2.07 1.74

BIPH 0.90 ±0.1 0.92 ±0.1 0.62 ±0.1 1.10 0.92

Calculated according to the 2-layer model using XPS measurements at two different

take-off angles (15° and 75°)

2Total length, /, of extended single molecules

Theoretical monolayer thickness. d\ of a perfect SAM with a tilt angle, x>, of 33°

between the molecules and the surface normal

145


2.1.2.4 Chemical Bond

The results of the AR-XPS investigations of the SAM-covered surfaces

clearly prove that the terminal phosph(on)ate groups are oriented towards

the metal oxide substrate surface and that the hydrocarbon chains (or arylmoiety in the case of biphenylphosphate) are oriented towards the vacuum.

At grazing angle detection (9 = 15°, corresponding to an information depth3Àsinl5° of approximately 2.5 nm) the calculated atomic ratio C/P is (as

expected) far higher than the stoichiometric ratio (except in the case of the

OH-DDPO4 SAMs). An increase of the take-off angle (9 = 75°,

corresponding to an information depth of approximately 9 nm) results in a

value much closer to the theoretical value (Table 25).

Table 25: Comparison of C/P ratios calculated from XPS studies of self-

assembled mono- and bifunctional phosph(on)ates on Ta20s, Ti

(covered with a natural oxide layer) or TiQ2.

Araphiphile Metal C/P

15° 75° theor.

ODPO4 Ta2Os 66.0 28.4 18

ODPO4 Ti 51.9 28.8 18

ODPO4 Ti02 63.2 24.1 18

DDPO3 Ta2Os 21.0 15.8 12

DDPOi n 25.1 17.2 12

DDPO, Ti02 25.3 14.3 12

OH-DDPO4 Ta20, 12.8 9.3 12

OH-DDP04 Ti 12.1 10.1 12

OII-DDPO4 Ti02 12.4 8.6 12

NEt-DDPO^ Ta2Os 39.6 21.0 14

NEt-DDP03 n 32.8 22.0 14

NEt-DDPOi Ti02 26.2 17.4 14

BIPH Ta2Os 23.1 12.4 12

BIPH Ti 24.2 12.7 12

BIPH Ti02 24.2 13.2 12

146


The oxygen signal Ois in the XPS spectra shows three different bindingenergies (four in case of OH-DDPO4 and BIPH) due to differences in the

binding of the corresponding oxygen. On the basis of literature reference

data two binding energies can be assigned to P=0 and P-O-R (R=C or H)and the third to the metal oxide. The fourth 0Js binding energy peak for OH-

DDPO4 and BIPH originates from the alcohol group (-CH2-OH) and the

phosphoric acid ester functionality (biphenyl-O-P), respectively. Curve

deconvolution supports the expected atomic ratios of 0(l)/0(2) - 1.4 for

phosphates and 3.5 for phosphonates (Table 26). The theoretical value for

the ratio is calculated according to the binding model of bidentate (A) and

monodentate (B) coordination (see also chapter IV 2.1.1.5) of the

phosph(on)ate bond to the metal cations (Figure 25).

Table 26: Ratio of the XPS oxygen peak intensities of 0(1) and 0(2)

originating from the phosph(on)ate head groups. The data are

taken from the deconvolved signal of the spectra measured at the

take-of angle of 15°.

Amphiphile/metal oxide

system

Ratio 0(l)/0(2)

experimental theoretical

ODP04/Ta205 1.3 ± 0.1 1.4

ODP04/Ti02 1.3 ±0-1 1.4

ODPO4/IÏ 1.7 ±0.2 1.4

DDP03/Ta205 3.5 ±0.3 3.5

DDP03/Ti02 2.4 ±0.2 3.5

DDPO3/TI 3.7 ±0.4 3.5

OH-DDP04/Ta205 1.5 ±0.2 1.4

OH-DDP04/Ti02 1.1 ±0.1 1.4

OH-DDPO/Ti 1.4±0.1 1.4

NEt-DDP03/Ta20, 3.2 ±0.3 3.5

NEt-DDP03/Ti02 2.6 ± 0.3 3.5

NEt-DDP03/Ti 2.6 ± 0.3 3.5

BIPH/Ta205 1.7 ±0.2 1.4

BIPH/Ti02 1.2 ±0.1 1.4

BIPH/Ti 1.5 ±0.2 1.4

147


In the spectra measured with a take-of angle of 75° the phosph(on)atcoxygen signal is very small compared to the large metal oxide oxygen

signal, resulting in a higher error for the phosph(on)ate Ois signal (data not

presented). Therefore these spectra were not included in the evaluation.

A B

O

Ta-Q O(0

©

0(11)

O O

Ta or Ti

530.0 - 530.6 531.3 -531.8 532.5 - 533.2

eV eV eV

C P-O-H

(hydrocarbonchain or arryl)

Fig. 25: Model for the mixed bidendate (A) and monodentate (B)coordination to tantalum ions of the phosph(on)ate head group of

the amphiphiles. The ratio 0(l)/0(2) is 7/5 or 1.4 for phosphatesand 7/2 or 3.5 for phosphonates. The binding energies are slightlyshifted depending on the metal ion and the organic moiety of the

corresponding amphiphile. The oxygen atoms in brackets are pre¬sent in the case of phosphates but not in the case of phosphonates.

148


The experimental ratio of the two different oxygen atoms 0(l)/0(2)shows a large difference between the phosphates and the phosphonates. The

values are close to the model of both mono- and bidentate phosph(on)atecoordination with a theoretical value of 1.4 and 3.5 for phosphates and

phosphonates, respectively.

Hydroxide groups or water bound to the metal oxide surface may be

present even after the self-assembly process. They may affect the ratio of

intensities at the binding energy positions of 532 and 533 eV. However, we

do not believe that this would completely alter the conclusions presentedabove.

2.1.2.5 Summary

The interpretation of the binding energies of the deconvolved 0ls

spectra allows further insight into binding aspects of the phosphate and

phosphonate head groups at the surface.

NEXAFS studies [Hofe,2000b] suggest a homogeneous monolayer in

terms of order only for the long-chain alkyl phosphate ODPO4 self-

assembled on Ta205 or TiO? (HDPO4 and TDPO4 not measured). The

hydrocarbon chain of DDPO3 seems to be at or below the critical chain-

length limit to form ordered SAMs from solutions of the free acid in organicsolvents. Terminal functionalities such as ethylamino or hydroxy groups as

well as biphenyl seem to prevent the formation of ordered SAM.

Nevertheless, comparison of the amphiphiles concerning differences in chain

lengths, terminal functionalities or head groups (phosphate or phosphonate)is difficult since small variations (e.g. purity of the substances, pH of the

solutions or properties of the used solvents) may influence the SAM quality.

The calculation of the adlayer thicknesses for samples that have been

immersed for 48 hours agree well with the findings of the previouslypublished NEXAFS study [Brov,1999] of ODPO4. This long-chain alkylphosphate has been shown to form ordered structures, similar to the

149


thiol/gold system, with a tilted orientation of the unfolded hydrocarbon chain

and a mean tilt angle, 1), of 30 - 35° relative to the surface normal. This tilt

angle agrees with the theoretical value for maximal monolayer adsorptionwith dense arrangement of phosph(on)ate groups at the substrate surface and

an optimal Van der Waals distance between the unfolded hydrocarbonchains. The experimentally determined (perpendicular) adlayer thickness of

2.1 ± 0.1 ran agrees well with the theoretical value of /•cost» (/ = 2.5 nm)and is also strong evidence for the formation of a densely packed monolayeras opposed to submono- and/or multilayer formation. For o-functionalized

amphiphiles OH-DDPO4, NEt-DDP03 and for the aryl phosphate BIPH such

an order could not be reached on tantalum oxide or titanium oxide by the

self-assembly process, as mentioned above. Although a sufficiently highadlayer thickness could be calculated for BIPH/Ta205, BIPH/Ti02 and

DDPCVTiOi, no large areas of uniform order (or tilt angle orientation) could

be measured with NEXAFS for these combinations [Hofe,2000b]. The Van

der Waals interaction between the hydrocarbon chains is believed to be

needed for the stability of the SAM order. The relatively short hydrocarbonchain length and the electrostatic repulsion between the terminal ethylaminogroups (R-NH2+-CH2CH3) are believed to be the reasons for the missingorder in the NEt-DDP03 SAMs. In the case of OH-DDPO4, hydrogenbonding between the hydroxy groups could result in an additional

stabilization of SAMs, but in the case of this amphiphile as well as in the

case of NEt-DDP03 steric hindrance among the solvated terminal groups

may occur. An overall impression is that all tested amphiphiles adsorb in an

oriented way (phosph(on)ate group oriented towards the metal oxide

surface), with the SAMs formed from alkyl phosphates being more highlyordered than those formed from alkyl phosphonates. In all cases that did not

show any order of their SAMs it is possible that the optimum solvent has not

yet been found. SAMs formed from aqueous solution of the ammonium salts

of DDPO4 and OH-DDPO4 do have the potential to form ordered

monolayers. NEXAFS and XPS studies of such SAMs are in progress and

will be published [Tosa,2000].

All of the experimental results obtained so far are compatible with the

model for the complex binding of the phosph(on)ate head group to the metal

ion based on both mono- and bidentate binding.

150


2.1.2.6 Toxicology

Since surface coating with phosph(on)ate amphiphiles may be

interesting for corrosion protection and/or adhesion promotion for lacquersin the metal industry as well as for coating of implants to improvebiocompatibility, it is essential that there is no cell toxicity. Even if the

surface coating would be used for high tech materials without any contact to

living cells, toxic substances are always more difficult to handle.

Toxicity of these amphiphiles is inherently probable because

phosphoric acid esters are a prominent group of nerve poisons, blocking the

Cholinesterase reaction in nerve synapses.

0.05 mM solutions of five different phosph(on)ates were prepared(DDPO3, OH-DDPO4, NEt-DDP03 and BIPH in methanol/water/ethanol

10:10:1 and ODPO4 in ethanol/water 1:1), and SAMs were formed on TiCV

coated glass chips according to the following protocol:

Cutting of chips : Four samples of Ti02 coated glass were cut into a

13 x 51 mm format for each amphiphile

Cleaning protocol : Sonication for 15 minutes in ultra-pure water

(removal of inorganic impurities and glass powder).

Sonication for 15 minutes in 2-propanol (removal of

polar organic impurities)

Sonication for 15 minutes in n-heptane (removal of

non-polar organic impurities)

Sonication for 15 minutes in ether (removal of

possibly adsorbed n-heptane)

Blow-drying in nitrogen gas stream (evaporation of

possibly adsorbed ether)

151


Self-assembly

Rinsing

Storage in ultra-pure water for 3 days (desorption of

any alkali-metal ions from the substrate)

Sonication for 15 minutes in methanol/water (1:1)(preconditioning of the immersion container)

UV cleaning for 30 minutes (activation of the

substrate)

Subsequent transfer of the chips into the immerison

container, addition of one of the 0.05 mM amphiphilesolutions, and immersion of chips for 48 hours

Removal of the chips and rinsing with ca. 5 ml of

methanol/water/ethanol (10:10:1) each

The samples were sent to a toxicology lab (Dr. Hendrich,Universitätsklinik Würzburg). The chips were eluted for 72 hours accordingto the DIN norm procedure. The extracts of three to four chips of each

phosph(on)ate SAM were investigated by the DIN/ISO 10993 technique:cytometer determination of the number of cells after incubation in the

extracts and performance of a proliferation assay (WST-1 corresponding to a

XTT-test). m the WST-1-test the extinction was recorded as a function of

substrate turnover. Both tests were examined for the BALB/3T3 fibroblast

clone A31 and for hFOB osteoblast cells.

None of the extracts (performed on each phosph(on)ate/metal oxide

combination) that were tested with the two toxicology test systems showed

any cell-toxic reaction.

152


2.1.3 Ammonium Salts of Alkyl Phosphates

Metal oxide surfaces have been coated by self-assembled monolayers(SAMs) of dodecyl phosphate (DDPO4) and 12-hydroxy dodecyl phosphate(OH-DDPO4), by means of adsorption of the alkyl phosphate ammonium

salts from aqueous solution. The SAM formation of DDPO4 has been

successfully demonstrated on anodized aluminum (A1203) as well as on

smooth thin films of tantalum oxide (Ta205), niobium oxide (Nb205),zirconium oxide (ZrC>2), iron oxide (Fe203), titanium oxide (Ti02) and

titanium silicon oxide (Tio.4Sio.0O2) deposited on glass substrates, resulting in

highly hydrophobic surfaces with water-contact angles > 110°. SAM

formation does not occur on silica surfaces under the same conditions. The

formation of SAMs based on the adsorption of hydroxy-terminated alkylphosphates (OH-DDPO4) onto Ta205 and Nb2Os has been studied. DDPO4

and OH-DDPO4 were co-deposited onto Ta205 from aqueous solutions of

their ammonium salts at different concentration ratios. It is shown that the

water-contact-angle can be precisely controlled within the range of 110° to

55° by adjusting the molar ratio of the two different molecules in the SAM.

The SAMs were characterized by water-contact angle (wettability),microdroplet density (condensation to judge homogeneity and order) and X-

ray photoelectron spectroscopy (orientation and adlayer thickness).

The use of organic solvents in the deposition process of self-assembled

monolayers has three main disadvantages: 1) in biomaterial applications,organic solvent molecules may be trapped within the adlayer, reducing the

biocompatibility of the surface; 2) for the formation of mixed adlayersystems, it can be difficult to find an organic solvent that is suitable for both

adlayer components; and 3) on an industrial scale, organic solvents are

increasingly falling into disfavor due to both air-pollution and water-

pollution issues. We demonstrate in this chapter that both pure and mixed

alkyl phosphates with different terminal functionalities can be depositedfrom aqueous solution by converting the free (water-insoluble) alkylphosphoric acids into the corresponding water-soluble salts. Mixed SAMs

on metal oxide surfaces are of particular interest to the biosensor and

biomaterial communities, since they allow surface properties such as

153


wettability, polarity, surface charge, etc. to be tailored in a precise manner.

Such surface tailoring may prove to be highly relevant for controlling the

interaction between the SAM-modified surface and biological systems such

as proteins, antibodies and cells.

2.1.3.1 Applied Methods

Microdroplet density (jidd) and contact-angle:

The microdroplet density (p.dd) was measured as described in chapterIII 2.2.4. The (add of the SAM-coated surfaces was measured before and

after a rinsing step with water. The average over 5 to 10 different regions of

the surface was determined. The data were used to determine the SAM

homogeneity in a lateral micrometer dimension. Even if the hydrophobicityis not indicative of a high order of the SAMs formed from bifunctional

amphiphiles, the contact-angle results may give some interesting information

in combination with the fidd data.

X-ray photoelectron spectroscopy (XPS) analysis:

Angle-resolved XPS (AR-XPS) measurements were conducted at

different take-off angles (11.5° to 75° with respect to the surface plane) to

get depth-dependent information.

154


2.1.3,2 Sample Preparation

DDP04(NH4)2 and OH-DDP04(NH4)2 were precipitated as described in

chapter III 1.2.1.

Preparation of 0.5 mM amphiphile solutions:

a) 15.0 mg of DDP04(NH4)2 were dissolved in 5 ml of high-purity water byheating up to ca. 50°C and adjusted to a volume of 100 ml with water.

b) 15.8 mg of OH-DDP04(NH4)2 were dissolved in 5 ml of high-puritywater by heating up to ca. 80°C. The solution was cooled to room

temperature, filtered through a 0.22 jim filter (MILLEX-GV,MILLIPORE, Bedford, MA) and adjusted to a volume of 100 ml.

The alkyl phosphate and hydroxy-alkyl phosphate solutions, a) and b),were mixed in different ratios from 0 to 100 vol.-% with respect to the

amount of OH-DDP04(NH4)2, in steps of 10% leading to 11 different

solutions at a constant total phosphate concentration of 0.5 mM.

The substrates were sonicated in high-purity water for 15 minutes

followed by a second sonication in 2-propanol for another 15 minutes. The

chips were removed from the cleaning solvent, blow-dried with nitrogen and

transferred to an oxygen plasma. After 3 minutes of oxygen-plasmacleaning, the chips were transferred to 5-ml glass vials, and amphiphilesolution was subsequently added. The chips were immersed for 48 hours in

the solution for self-assembly, removed, rinsed with 10 ml of high-puritywater each and blow-dried with nitrogen.

155


2.1.3.3 Investigation of Surface Properties

DDPO4 on different metal oxides:

Self-assembled monolayers of dodecyl phosphate were produced on

AI2O3, Ta205, Nb205, Zr02, Fe203, Ti02, Tio.4Sio.0O2 and Si02 by immersion

in 0.5 mM DDP04(NH4)2 as described in the previous section.

Contact angle:

Advancing water-contact angles (CA) were measured immediatelyfollowing sample preparation (Table 27). The results show that highly

hydrophobic SAMs form on all the metal oxide surfaces used in this work,with the exception of silica and Tio.4Sio.6O2. The contact-angle values of

> 110° are typical for well-defined alkyl phosphate SAMs [Brov,1999]. The

isoelectric points (IEP) of the different bare metal oxides, which vary from

2.7 - 3.0 for Ta205 to 7.0 - 8.6 for Fe203, do not seem to be an importantparameter for SAM formation with this system (Table 27). On Si02, the

contact angle remains in the same range as it was prior to immersion in the

amphiphile solution. This suggests that DDPO4 does not adsorb on Si02with this approach.

The Ti0.4Si0.6O2-coated glass chips consist of a heterogeneous (two-phase) structure of Ti02 and Si02 nanoparticles at the substrate surface.

Tio.4Sioè02 samples immersed in the amphiphile solution show reduced

wettability. The contact-angle value of 64° suggests the formation of an

incomplete alkyl phosphate monolayer. This is consistent with the

observation that DDPO4 self-assembles onto Ti02 but does not adsorb on

Si02 from an aqueous solution of the amphiphile ammonium salt.

156


Table 27: Literature values of isoelectric points (IEP) of the investigatedbare oxides and experimental advancing water-contact-angles

(CA) (± standard deviation) and microdroplet density values ((add)

(± standard deviation) following self-assembly of DDPO4 on the

oxide surfaces.

Substrate IEP Ref(IEP) CA (± stdev) (idd (± stdev)

Ta205 2.7-3.0 [Bous,19911 114.6 ±0.48 129 ±19

A120, 7.5-8.0 [Matt, 19301 111.4± 1.07 152 ±22

Nb20, 3.4-3.6 [Park, 1965] 109.7 ±0.63 120 ± 18

Zr02 4.0 [Verw,19411 110.1 ±0.61 258 ±39

Fe2C>3 7.0-8.6 [Haze, 1931;Matt, 19341

110.8 ±0.64 197 ±30

Ti0 4Sio e02 3.6 [Kena,20001 63.8 ±1.93 334 ±50

Ti02 4.7-6.2 [Joha, 19571 111.4=4= 1.18 221 ± 33

Si02 1.8-2.2 [Park, 19651 10.0 ±3.42 >3000

Microdroplet density:

In order to remove any particles that had adsorbed on the surfaces, the

samples were rinsed with high-purity water just before measuring the

microdroplet density (jidd). DDPO4 SAMs on pure metal oxide surfaces

show a low |idd value of 150 - 250 droplets/mm2 (Table 27) and a

homogeneous distribution of the growing droplets of condensed water

(Figure 26). Si02 remains hydrophilic (as mentioned above) and micro-

droplets coalesce to a complete layer of condensed water before they can be

detected as single droplets (The maximum value that can be detected with

the used setup corresponds to about 3000 droplets/mm"). The jadd value of

> 3000 droplets/mm" is consistent with the suggestion based on the contact-

angle data, that DDPO4 does not adsorb onto Si02 from aqueous amphiphilesolution. On the DDP04-covered Ti04Sio602 substrate, a Ltdd of 350

droplets/mm"" was measured. This value is significantly higher than the

157


values (typically less than 300 droplets/mm ) for alkyl phosphates self-

assembled onto the bare metal oxide surfaces. It reflects lower order in the

adlayer structure as a consequence of the incomplete monolayer formation.

Fig. 26: Microdroplet-density measurements of DDP04 SAMs on different

smooth metal oxide films (image size: 1 mm2): a) AI2O3, b) Ta205,

c) Nb2Os, d) Zr02, e) Fe20,, f) Ti02, g) Ti0 4Si0 602 and h) Si02.The homogeneous distribution of growing water droplets reflects

the homogeneity of the hydrophobic SAMs on a Jim scale. Si02remains hydrophilic and is completely covered by a layer of

condensed water. The Ti0 4Si0 602 surface shows a slightly higher\xàà, suggesting homogeneous, but only partly covered by the

alkyl phosphate.

158


Fe203 was coated onto a Tio.4Sio.6O2 glass slide in the form of a stripewith a width of 2 mm. During the rinsing step after the immersion in the

amphiphile solution it became clear that the Fe203 area had become

hydrophobic, while the surrounding area (Tio.4Sio.6O2) remained more

hydrophilic. The resulting water-repellent stripe, 2 mm in width,

corresponded to the Fe203-coated area (Figure 27).

Fig. 27: DDPO4 SAM on a Tio.4Sio.6O2 surface with a 2 mm large stripe of

Fe203. The Fe203 surface is covered by DDPO4 and hydrophobic(low (J.dd) while the substrate (Tio.4Sio.0O2) is more hydrophilic due

to a lower degree ofDDPO4 coverage.

XPS:

After immersion of the metal oxide and silica samples in a 0.5 mM

DDP04(NH4)2 solution, the surfaces were investigated by X-rayphotoelectron spectroscopy (XPS) at two different take-off angles, 0. At a

take-off angle of 15° (with respect to the surface plane) the analysis is highlysurface sensitive, while a take-off angle of 75° yields more information on

the substrate.

The atomic concentrations of C, O, P and of the corresponding substrate

metal cations were calculated following standard procedure (Tables 28a and

b). With the exception of iron oxide, the different metal oxide coatings are

thicker than the information depth of the XPS technique and therefore no

signals of the underlying waveguide layer were detected. With the exceptionof iron oxide, the different metal oxide coatings are thicker than the

information depth of the XPS technique and therefore no signal of the under-

159


laying waveguide layer were detected. On the iron oxide samples, traces of

the Tio.4Sio.6O2 substrate beneath the sputter-coated Fe203 could be detected

at the 15° take-off angle (information depth ~ 2.5 nm). The titanium was

quantified and presented in Table 28a and b. At the 75° take-off angle(information depth ~ 9 nm) the atomic concentration of titanium was much

higher, showing that the Tio.4Sio.6O2 is effectively located beneath the Fe203.XPS data with the determined binding energies are summarized in the

appendix (App. 14).

Tables 28a and b): XPS-determined atomic concentrations of DDP04, self-

assembled onto different metal oxide substrates. The

measurements are performed at take-off angles of 15° and

75° with respect to the surface plane. The values for the

atomic concentration of titanium originate from the

substrate underneath the sputter-coated metal oxide in the

case of Fe203, and for the atomic concentration of silicon

in the case of the mixed TiSi02 substrate are also

presented.

Substrate

M0X

Atomic Concentration (15° take-off angle)

%C % 0 (tot) %M %P

Ta205 67.8 23.2 6.71 2.36

AI2O3 67.6 21.1 8.85 2.53

Nb205 59.7 29.6 8.5 2.26

Zr02 72.2 22.0 3.06 2.77

Fe203 66.1 27.5 Fe: 3.09/Ti: 0.41 2.06

Ti02 68.6 23.0 6.13 2.23

Si02 6.2 70.8 22.9 0.0

Tio.4Sio.6O2 26.2 54.5 Ti: 3.04/Si: 15.1 1.13

160


Substrate

MOx

Atomic Concentration (75° take-off angle)

%C % 0 (tot) %M %P

Ta205 27.3 53.7 17.4 1.61

A1203 25.3 47.1 25.5 2.10

Nb205 31.9 49.4 17.3 1.30

Zr02 35.9 49.0 12.9 2.28

Fe203 28.9 56.5 Fe: 10.2 /Ti: 2.46 1.93

Ti02 31.2 49.4 17.6 1.83

Si02 n.d.a) n.d.a) n.d.a) n.d.a)

Tio 4Sio 6Û2 14.9 60.2 Ti: 9.97/Si: 14.1 0.85

a) not detected

The amount of adsorbed amphiphile on the surfaces is in the same

range for all investigated metal oxide substrates and points to a coverageclose to one monolayer [Text, 1999]. Consistent with the CA and fidd results,no adsorbed alkyl phosphate could be detected on the Si02 substrate surface

(no phosphorus and very low carbon). On the mixed Ti/Si samples(Tio.4Sio.0O2 substrate) the titanium concentration was about half of that of

the bare Ti02 substrate. Correspondingly, the detected surface concen¬

trations of P and C are also scaled down by approx. 50% compared to the

results obtained on the pure Ti02 surface. This supports the hypothesis that,from aqueous solutions, DDP04(NH4)2 is selectively adsorbed on the Ti02rather than on the Si02 regions. As a consequence of the lower degree of

coverage, the photoelectrons from the 0)s in the substrate are less affected

by inelastic scattering by the adlayer, and the atomic concentration of

oxygen is therefore significantly higher for DDP04/Tio4Sio602 in

comparison to DDP04/Ti02.

161


OH-DDPO4 on Ta205 and Nb205:

12-hydroxy dodecyl phosphate (OH-DDPO4) was self-assembled on

Ta2Os and Nb205 chips by immersion in a 0.5 mM OH-DDP04(NH4)2aqueous solution, as described in a previous section.

Contact angle and XPS:

The advancing water-contact angle was measured immediately after the

self-assembly process. The contact angle was approximately 50°, i.e. the

surface is much more hydrophilic than in case of the methyl-terminatedDDPO4 SAMs (the water-contact angles of the clean substrates were close to

0°). This is good evidence that the terminal hydroxy1 groups are indeed

exposed at the SAM surface. To prove this hypothesis, XPS spectra were

recorded at different take-off angles (The information depth scales linearlywith the sine of the take-off angle). The variation of the signal intensity as a

function of the take-off angle may give information about the position of the

corresponding element. The Ois signal of the spectra at the two grazingangles 11.5° and 20.5° are shown in Figure 28. Compared to the non-

functionalized dodecyl phosphate, the oxygen signal of the hydroxy groupshows an additional shoulder at 533.4 eV (Figure 28) with a relative

intensity that is significantly higher at 11.5° compared to 20.5° (Peakdeconvolution presented in App. 15). This is due to the fact that the Oissignal from the external hydroxy1 group is not affected by inelastic scatteringwithin the alkyl monolayer, whereas the Ois signals from the interfacial

phosphate and the covered metal oxide are strongly reduced at grazing take¬

off angles.

No nitrogen could be detected by XPS, demonstrating that the

ammonium cations are not incorporated in the self-assembled adlayer.

162


7000

>!"! 512 511 510

Hiiuliii!' Fnenn»

529 528 527 M4 53"! 512 531 510 52')

Binding Energv [eV|

Fig. 28: Ois signals of XPS measurements of self assembled OH-DDPO4

onto Ta20s (left) and NboOs (right). The data from two different

take-off angles (20.5° top and 11.5° middle) are compared to the

oxygen signal of a DDPO4 SAM on the corresponding substrate,measured at a take-off angle of 15° (bottom).

It has been shown that water-contact angle data reflect the degree of

coverage of long chain alkyl phosphate SAMs [Brov,1999]. However,

contact-angles values of a complete monolayer of hydroxy-terminatedamphiphiles have been reported to be 50 - 80° [Folk,1995]. To test whether

OH-DDPO4 forms densely packed SAMs, we compared the carbon signalsfrom DDP04-SAMs with those of OH-DDPO4 measured by XPS at different

take-off angles (Figure 29). The results show similar carbon signalintensities for both amphiphiles on Ta205 as well as on Nb205, suggestingthat hydroxy-terminated dodecyl phosphate also forms densely packedmonolayers.

A collection of all the different XPS Ois detail spectra (from different

take-off angles) is presented in the appendix (App. 16).

163


0 50

sin 0

Fig. 29: Comparison of carbon concentrations determined by XPS of OH-

DDP04 and DDP04-SAMs on Ta205 (left) and Nb205 (right),measured at different take-off angles (0). A linear trend line is

added for the OH-DDPCVcarbon values (squares). DDPCVcarbonvalues are represented by triangles.

The atomic concentrations of C, O, P and the corresponding substrate

metal cations Ta or Nb were calculated after quantification of the elements

(Tables 29a and b). The data are consistent with our model of a surface

architecture where the alkyl phosph(on)ates adsorb onto metal oxide

substrate surfaces with a "tails-up" orientation.

The Ois signal was fitted and separated into three different types of

oxygen: a) the metal oxide (530.2 eV), b) the phosphate oxygen (531.4 eV

for P-O-Metal and P=0; 532.6 eV for R-O-P and P-OH) and c) the terminal

hydroxyl group (533.4 eV). The assignment of different binding energies to

the different species in the substrate and coordinated phosphate groups has

been described in chapters IV 2.1.1.3/4/5 and IV 2.1.2.4 [Text, 1999]. The

data are again consistent with an oriented adsorption (phosphate head-groupstoward the metal oxide surface) of the OH-DDPO4 molecules (Figure 30).

164


01 01 05 07 û 1 01 03 05 07 09

sin 0 sin 0

Fig. 30: Peak area of Nb3d, Cts and P2p (left) as well as peak area of

different oxygen types (right), determined by XPS of an OH-

DDP04 SAM on Nb205. The 0,q signal was deconvolved into 3

different types of oxygen [0(Nb2Os): squares; 0(PC>4): triangles;

O(ROH): dots], The atomic peak areas are represented as a

function of the sine of the take-off angle, 0, Similar results were

obtained from OH-DDPO4 self-assembled on Ta2Os.

The rise of the substrate oxygen signal as the take-off angle approachesthe surface normal is due to the increase of the sampling depth. The

thickness of the metal oxide layer can be considered as semi-infinite for XPS

measurements. The phosphate oxygen 0,s signal decreases slightly as the

take-off angle increases. This can be explained by our model, where the

phosphate molecules build an interfacial monolayer between the substrate

and the alkyl chain layer. The Ojs signal from the hydroxy group decreases

slightly with increasing take-off angle, 0.

The rise of the metal (Ta, Nb) signal originating from the semi-infinite

oxide substrate with increasing take-off angle, 0, is again, as in the case of

the oxygen substrate signal, clearly due to the corresponding increase of the

sampling depth, which varies with the sine of the take-off angle, 0. This is

paralleled by a corresponding decrease of the carbon signal of the organicadlayer. All signals, on the other hand, that originate from a purely two-

dimensional, roughly mono-atomic surface or interface layer (PO4 interface

and terminal OH groups) drop to an intensity that shows much less

165


dependence on the take-off angle. The same observation has been reportedfor the similar system of octadecyl phosphate self-assembled on tantalum

oxide [Text, 1999]. The reason behind this is the fact that the XPS intensities

of elements contained in two-dimensional layers are not proportional to the

mean free path, X, of the electrons, in contrast to intensities of elements

within three-dimensional structures (thin films or substrates). X in the case of

such two-dimensional atomic arrangements only enters as far as the

reduction of intensities by overlying films is concerned (here the reduction

of the PO4 intensities due to the alkane chains), and this effect is rather small

in our case as a consequence of the rather small alkane chain length (about1.4 nm corresponding to 1.2 nm thickness at a tilt angle, t), of 30 - 35°).

Table 29a and b: Atomic concentrations determined by XPS of OH-DDPO4self-assembled onto TaÔs and ND2O5, respectively. The

measurements were performed at different take-off angles.The sine of the take-off angle, 0, was increased linearly,corresponding to a linear increase of effective adlayerthickness, which has to be penetrated by the photo-electrons.

Take-off

Angle

0

sin 0 Concentraition [at-%]

%Ta %0 %P %C

11.5 0.20 3.05 26.6 3.69 66.72

20.5 0.35 6.27 33.0 3.91 56.79

30.0 0.50 9.21 39.8 3.12 47.78

40.5 0.65 11.9 45.8 2.83 39.46

53.1 0.80 13.42 50.3 2.00 34.29

71.8 0.95 14.88 53.3 2.12 29.65

166


Take-off

Angle

0

sin© Concentration [at-%]

%Nb %0 %P %C

11.5 0.20 4.64 29.3 3.86 62.21

20.5 0.35 6.74 33.9 3.32 55.97

30.0 0.50 9.53 38.5 3.07 48.62

40.5 0.65 11.97 43.6 2.78 41.97

53.1 0.80 14.05 47.7 2.01 36.24

71.8 0.95 16.38 51.6 2.07 29.98

Mixed SAMs of OH-DDP04/DDP04 on Ta205:

Mixed solutions of 0.5 mM OH-DDP04(NH4)2 and 0.5 mM

DDP04(NH4)2 were prepared and Ta205 samples were cleaned and

immersed in the mixed amphiphile solutions as described in a previous

chapter (IV 2.1.3.2).

Contact angle and microdroplet density:

The advancing water-contact angle was measured immediately

following sample preparation. The results show that the wettability of

the resulting SAM correlates with the volume (or molar) ratio

OH-DDP04(NH4)2/DDP04(NH4)2 of the correspondingly used amphiphilesolution (Figure 31).

167


15

105

a»

1^ 95

i85

o03+-*

75oo

r-< 65a>4-»CO

£ 55

45

Fig. 31:

3000

2500 |

2000

40 60 80 100

amount of OH-DDP04(NH4)2 in solution [%]

&

a»

1500 "o

©.—i

1000 p4

O

500 g

Contact angle (squares) and u,dd (dots) data from mixed

OH-DDPO4/DDPO4 SAMs on Ta205. The results are presented as

a function of the amount of OH-DDPO4 in the amphiphilesolutions in vol-%. The total alkyl phosphate concentration is keptconstant at 0.5 rnM.

The microdroplet density images appear homogeneous on the

micrometer scale, suggesting that hydroxy-functionalized and non-

functionalized alkyl phosphates in the monolayer are homogeneouslydistributed at this scale. The difference in wettability does not significantlyaffect the u.dd. An increase of the droplet density would therefore be related

to an increase in roughness and/or charge at the surface or a decrease in

orientation or building of molecule clusters at the surface [Hofe,2000a]. The

low |add value of 100 - 200 droplets per mm" suggests the presence of well-

defined smooth adlayers [Hofe,2000a] (Table 30). The (idd does not show a

dependence on the amount of OH-DDPO4 at the surface. This suggests that

the incorporation of OH-DDPOt into the DDPO4 layer does not affect the

organization of the SAMs.

168


Table 30: Contact-angle and microdroplet density (jidd) (± stdev) data of

mixed SAMs on Ta20v

% OH-DDP04

in the solution

Contact angle (± stdev)

[°]

jidd (± stdev)

[mm" ]

0 110.1 ±0.8 144 ± 29

10 105.6±1.7 126 ±26

20 103.6 ±1.0 104 ±21

30 86.6 ±1.4 167 ±33

40 81.6 ± 1.4 125 ±25

50 73.0 ±1.8 102 ±21

60 70.9 ±1.9 72 ±30

70 64.1 ±2.3 161 ±32

80 58.2 ±1.5 123 ±25

90 57.3 ± 1.2 229 ± 46

100 54.3 ±3.5 159 ±32

2.1.3.4 Exposure to Inorganic Phosphate Solution

The influence of inorganic phosphate solution on alkyl phosph(on)ateSAMs has to be discussed from two perspectives:

a) Is it possible to exchange self-assembled alkylphosph(on)ate molecules

by inorganic phosphate anions (substitution)?

169


If there is a substitution reaction of inorganic and organic phosphates, a

significant increase of the surface wettability of hydrophobic alkyl

phosph(on)ate SAMs would be expected as a consequence of loweringthe monolayer order. Simultaneously, the microdroplet density would

increase due to lowering the surface homogeneity.

b) Do inorganic and organic phosphates competitively adsorb onto clean

metal oxide surfaces (competitive adsorption)?

If the driving force for self-assembly of alkyl phosph(on)ates is the

strong phosph(on)ate - metal cation interaction, one would expect a

lower degree of coverage by organic phosph(on)ates if inorganicphosphates are present in the amphiphile solution during the SAM-

formation process. If the driving force is the Van der Waals interaction

between the hydrocarbon chains, the SAM-formation process is expectedto be less influenced.

To answer these questions, hydrophobic DDPO^NTL^ SAMs were

stored in PBS buffer solution, and water-contact angle, |idd and XPS were

measured (see also chapter IV 2.1.1.2). XPS data were performed at a

detection angle of 15° with respect to the surface plane for high surface

sensitivity. In a second experiment, the SAM formation of DDP04(NH4)2 on

clean Ta20s from a solution in PBS was monitored by measuring the water-

contact angle since the hydrophobicity has been shown to be a good marker

for the determination of coverage for methyl-terminated alkyl phosphateSAMs [Text, 1999].

For the phosphate - alkyl phosphate exchange reaction studies, Ta205

chips were cleaned according to the standard cleaning procedure (sonicationin 2-propanol followed by UV-cleaning) and coated by DDP04 from a

0.5 mM aqueous solution of DDPO^NFLi)? as described in previouschapters. The SAM-coated chips were subsequently stored in PBS for

1 hour, 3 hours, 20 hours, 50 hours, 14 days and 16 days and continuouslystirred during this period. The samples were removed after the

corresponding storage time, rinsed with water and analyzed. Contact angleand [idd data are shown in Figure 32.

170


1200

1000

800

600 "3

400

200

0

Fig. 32: Water-contact angle (triangles) and microdroplet density (dots)data of DDPO4 self-assembled on Ta205 as a function of storagetime (logarithmic scale) in PBS buffer. A logarithmic trend line is

applied on both data series (note that the x-axis is presented in

logarithmic scale).

The results from contact-angle and (idd analysis suggest a desorption of

DDPO4. XPS results show a decrease of the carbon Cis signal. This agrees

with the assumption of alkyl phosphate desorption. The phosphorus P2psignal intensity tends to increase, suggesting adsorption of inorganicphosphate (results presented in App. 17).

Regarding these data, we assume that there is a substitution reaction of

inorganic and organic phosphate.

For the study of competitive inorganic and organic phosphateadsorption, a 0.5 mM DDPC^NHOi solution in PBS buffer was prepared.Clean Ta205 chips were incubated for self-assembly. Samples were removed

after 15 seconds, 50 seconds, 3 minutes, 10 minutes, 80 minutes, 16 hours

and 4 days, immediately rinsed with water and blow-dried with nitrogen.Water-contact angle was measured, and the results are shown in Figure 33.

0 1 10 100

incubation time [hj

171


/J>

21

/""N 19O

(!) 17

"Ehr~r 15CO

13

CO-ti 11t-t

ÜÜ 9

7

5

0.1 100 1000 10000

time (min)

Fig. 33: Contact-angle data of DDP04-SAM formation on Ta20^ surface

from 0.5 mM DDPC^NH^ solution in PBS. The results are

plotted as a function of immersion time. The x-axis is presented in

logarithmic scale.

The results show that almost no DDPO4 adsorbs at the metal oxide

surface from the 0.5 mM DDP04(NH4)2 solution in PBS, suggesting that

inorganic phosphate adsorbs at the Ta205 surface and reduces the

accessibility for organic phosphate. The amphiphile solution in PBS (total

phosphate concentration in PBS is 11 mM) has an organic-to-inorganic ratio

of 0.5/11. If the sticking coefficient is the same for both the organic and the

inorganic phosphate, the results can be explained on the basis of competitiveadsorption.

Contact-angle and microdroplet density data from SAM stability analysis of

DDPO3, ODPCh and ODPO4 (0.5 mM in n-heptane/2-propanol) as well as of

DDPO4 (0.5 mM DDPO^NR^ in water) self-assembled on Ta20^ are

summarized in App. 18 and 19.

170


2.1.4 Poly-(L)-Lysine Polyethylene Glycol) (PLL-PEG)

Having a pKa of about 9.5, the PLL-backbone of the co-polymer is positivelycharged in pH-neutral solutions. However, metal oxides such as Ta2Os and TiC>2 are

negatively charged at this pH. The adsorption mechanism and the driving force for

self-assembly is therefore completely different from SAM formation from

phosph(on)ates on such substrates. One can consider the bonding between the

polymer molecules and the substrate to be a multiple Coulomb interaction.

Optical waveguide lightmode spectroscopy (OWLS) was used for the

monitoring of PLL-PEG adsorption, protein resistance and specific streptavidinbinding onto PLL-PEG/PEGbiotin coated surfaces. The OWLS measurements were

performed by J. Vörös.

PLL-g-PEG self-assembly, protein resistance [Prim, 1993]and strepatvidin binding:

Solutions of PLL(20)-g-PEG(2) derivatives having grafting ratios of g[2],g[3.5] and g[5], respectively, were prepared in 10 mM HEPES buffer (pH 7.4), at a

concentration of 1 mg/ml and filter-sterilized (0.22 (im Durapore Millex, Sigma-CH). Self-assembly of the different non-functionalized copolymers on T1O2(3 samples per polymer type) was allowed for proceeding for 90 minutes and the

adsorbed mass was determined. Samples were then rinsed extensively with filter-

sterilised HEPES buffer (pH 7.4), and protein solution (1 mg/ml in HEPES buffer)was applied for non-specific protein-adsorption studies (Figure 34).

173


«r-

Of.=5

-/<

«

U ZS

DPU-PKilB Set ura

0 2 — — — —

40 15 -

IIBR!^

0 1

IBS0 05

A mifaa,—,

3 5

graftme ratio

Fig 34 Mass of PLL-PFG copolymer adsorbed onto aTiÖ2 surface (tig/cin2) and

protein (serum) resistance of three different PLL-g-Pl G polymers as

determined by the OWL S technique

When the PFG mter-cham spacing is smaller or equivalent to the PI G radius

of gyration (for details see publication (Kena,20()()]) non-specific protein adsorp¬tion is efficiently prevented This is reached tor the two PLI -PKi copolymcishaving giaftmg ratios of 2 or 3 5 On the othet hand, an mter-cham spacing largerthan the PPG radius of gyration leads to protein adsorption levels comparable to

those of the bare TiOz surface In addition, one can also note the difference m the

polymer mass adsorption, which is highei for the PI 1 -g[3 5|-PFG compared to the

PLL-g[2]-PFG This is probabfy related to the smaller number of anchoring NIL,1groups available m the PLL-g[2]-PI G polymer since a larger number oi these side

chains have been denvati/ed by PF G Although the mass of copolymer adsorbed at

the surface of the waveguide is lower, the density ol PLG-chams is probably highenough to lead to PPG chain o\erlappmg and hence sufficient to yield a protein-

repellent PL1 -g-PPG mteiface Nevertheless, one can argue that the lower polymermass adsoibed as well as the few et anchoring points m this polymer could be

detrimental to the stability of the interface Therefore, the PI 1 -g|3 5J-PFGderivative seems to be optimum in oui PI L -g-P! G polymer series

174


Based on results of protein-resistance studies of the PLt -PPG modified

surfaces, biotin-functionahzed copoljmers with a grafting ratio of g[3 5| were

synthesized with different biotin concentrations (see chapter III 1 2 2 i)

Solutions of PLL(20)-g[3 5|-PI G(2) (non-fimctionah/ed) as well as of

PLL(20)-gf3 5]~PFG(2)/PFGbiotin(3 4)10% and 50%, respectively, were preparedat a concentration of 1 mg/ml m HEPFS buffet and filter-sterilized tor each

copolymer derivative, a clean Ti07 waveguide chip was mounted in the OWLS

apparatus, and self-assembly of the corresponding P1L-PLG derivative was

measured for 90 minutes The chips were rinsed with HEPFS buffer followed by

application of streptavidm solution (1 mg/ml m I If PL S buffer) The adsorbed

masses of the PLL-PI G derivatives and of the specifically bound streptavidmthereon were calculated (Figure 35)

0 4

0 35

o i

>

I 0 25

tot3

0 2

w~ . -

o i

0 05

0

no biotin 10%biotin 50% biotini

_

Fig. 35 OWLS results of PL1 -PFG copolymer mass adsorption and subsequentstreptavidm binding thereon (measured m jig/cm") Comparison of

biotin-functionahzed PI I-PIG derivatives (10% and 50% biotin,

respectively) with non-functionalized PLI ~PLG Grafting tatio of g[3 5\was kept constant for ail three polvmer derivatives

PLG-PLL'

Streptav idin

175


The amounts of adsorbed PLL-PEG mass per area reflects the differences in

molecular mass of the derivatives. The density of adsorbed streptavidin layer on the

50% biotin-functionaiized derivative (PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4)50%)of about 0.3 mg/cm corresponds to about one monolayer.

An example of on-line monitoring of PLL-PEG copolymer adsorption, proteinresistance measuring and streptavidin binding using the OWLS technique is shown

in Figure 36.

s

"woJEL

a

ou

o

ViVi

400 H

300

200

100

0

BufferSerum

Buffer

PLL-PEG/

PEGbiotin

50%

Streptavidin

20 40 60 80 100 120 140 160 180

T* I • i

îme |min|

Fig. 36: On-line monitoring of mass adsorption with the OWLS technique:Adsorption of PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4)50% followed by

rinsing with a buffer for a stability measurement of the copolymer

adlayer. Application of serum to study non-specific protein adsorptionfollowed by rinsing with buffer to investigate the stability of non-

specifically bound protein. Finally, application of streptavidm to study

specific binding onto interfacial biotin.

176


The curve illustrates the very fast self-assembly kinetics of PLL-PEG onto

TiOi surfaces. Rinsing with buffer only results in a small decrease of the adsorbed

mass, showing the stability of the copolymer adlayer. A slight increase of the mass

adsorption curve suggests a small amount of non-specific binding from the

subsequently applied serum. The adsorbed serum seems to be very loosely bound

since an additional rinsing step with buffer results in a fast decrease of the curve

back to the mass-level of the clean copolymer adlayer. However, application of

streptavidin yields a big "jump" in the mass adsorption curve, illustrating the highly

specific binding of streptavidin to biotin conjugated PEG. The near flatness of the

curve during the third rinsing step with buffer suggests the stable binding of

streptavidin.

177


2.2 Adsorption of Thiols onto Gold Surfaces

The motivation behind the use of gold chloride layers on metal oxide

surfaces was the combination of the evanescent field measuring techniquesand the large spectrum of commercially available alkyl thiols: Opticalevanescent field techniques need highly transparent waveguide layermaterial such as Ta2Û5. Thiols, on the other hand, preferably adsorb onto

gold (or silver) surfaces. Since an adlayer of gold (Au0) on a waveguide

layer would completely absorb incoupled light, the highly transparent goldchloride (Aum) was used for preliminary coating experiments.

2.2.1 Coating of Ta2Os with Gold Chloride

The pH-dependence of the gold chloride solution during the coating

process of the Ta2Os chips and the amount of self-assembled alkane thiol on

the resulting layer was investigated as follows:

A 0.4 % gold chloride solution was neutralized with a 0.1 M NaOH

solution (Figure 37). The stepwise neutralization of the acidic gold chloride

with NaOH and its proposed major chemical structure are shown in Figure38.

178


pH titration-curve of 0.25 mM HAuC14 with NaOH

5 10 15

mlO.lMMaOH

20

Fig. 37: Experimental pH-curve of neutralization of a gold chloride

solution with NaOH.

[AuClj]" + NaOH ~> [AuCl3(OH) ] + NaCl (deep yellow)

[AuCl,(OH) ]" + NaOH -> [AuCl2(OH)2]- + NaCl (yellow)

[AuCl2(OH)2]" + NaOH -» [AuCl (OH),]" + NaCl (light yellow)

[AuCl (OH),]" + NaOH -> [Au(OH)4]" + NaCl (colorless)

Fig. 38: Stepwise neutralization of gold chloride with NaOH: The color

changes after every partial ligand exchange reaction.

179


For the determination of the optimum pH for the gold chloride coatingprocess, the TajOj chips were cleaned with 2-propanol in an ultrasonic bath

for 15 min followed by UV cleaning for 30 min. In each [AuCl4.n(OH)n]"solution of different pH, two chips were immersed for 48 hours followed byrinsing with water, drying in a N2 stream and storing at 80°C for 1 hour.

One of the chips from each [AuCl4.n(OH)n]" solution was immersed in

an octadecyl phosphate (ODPO4) solution (0.05 mM in ethyl acetate), which

is known to strongly adsorb onto Ta20s. In parallel, another chip from each

gold chloride solution was immersed in a solution of octadecyl thiol (ODT,0.05 mM in ethyl acetate) which selectively adsorbs onto gold surfaces. The

protocol is schematically drawn in Scheme 16. The chips were removed

after 24 hours, rinsed with ethyl acetate and blow-dried with nitrogen.

If the tantalum oxide surface is completely covered with gold (or goldions), alkyl thiols (ODT) should adsorb thereon. If there is no adsorbed gold(or gold ions) on the tantalum oxide surface, then no alkyl thiol adsorption is

expected, but ODP04 should build a SAM. A partial coating of the tantalum

oxide surface would result in a partial SAM formation in both series, the

alkyl phosphate and the alkyl thiol. The wettability is therefore a useful

parameter to test qualitatively the coverage of the tantalum oxide with gold.

In addition to measuring the advancing water-contact angle of the

amphiphile-treated chips, microdroplet density, XPS-spectra and AFM

images of the samples were optained.

180


Determination of pH-optimum

vuntreated chips

V

i incubation in Au-solution having different

pH-\ alues IpH 7

pi I 6

pi I 5

pH 4

incubation in

ODT

JZZZZZZ

incubation m

ODPO4 1

J-ê^-^^t adsorption depending on C^Z^Z=Au-co\ erase p

p^-^H^ ^ .v.i*V»(i1.4k,1

c

COo

Hxpected wettability

pH

Scheme 16: Determination of pH-optimum. The adsorption of ODP04 (red)or ODT (blue) is controlled by the advancing contact angledata.

181


Contact-angle and iidd;

Contact-angle data from the samples of the series described in the

schematic drawing (Scheme 16) show the expected wettability curve

(Figure 39).

120

110

100CD

3 00

«

-t-1

U 80CS

-+*J

1-1 70oo

60

50

40

/"^0— —-•»

j/"S

.—

\

-ODPÖ4IODT J

\

—

2 5 4 5 6 5 8 5 10 5

pH of gold chloride solutior

Fig. 39; Contact-angle data for octadecyl phosphate (ODPÜ4) or octadecylthiol (ODT) self-assembled onto tantalum oxide surfaces, which

were immersed in 0.2 % (w/v) HAuCLpsolutions of different pll-values. The contact-angle data semi-qualitatively reflect the

amount of adsorbed amphiphile.

The contact-angle results show that over the whole range of goldchloride solution pH values there is higher hydrophobicity after immersion

of the gold chloride-treated chips in the amphiphile solutions compared to

that of bare tantalum oxide, which is typically 5 - 10°. This is what we

expected for ODPO4 since alkyl phosph(on)ates are known to adsorb on

different metal cations such as Cu(ll), Au(Il), Ti(IV), Ta(V) etc. [Folk,1995;Brov,1999]. Adsorption of ODT may occur on Ta20>, forming incompletemonolayers (only small increase of the hydrophobicity) outside the ideal pFl

182


range for gold chloride adsorption, The region between pH 6 and pH 8

shows the highest contact angle for ODT-treated chips and the lowest

hydrophobicity for ODPCVtreated chips. This is a strong indication that the

pH-neutral region constitutes the best condition for gold adsorption usingthis system. The relatively high contact-angle value of 104° is close to a

complete monolayer of alkyl thiols on ultra-flat gold but may also rise from

incomplete SAMs on nano-structured surfaces showing the so-called lotus

effect [Bart,1997].

The stability of the modified surfaces was controlled by additional

rinsing of the chips with water for 5 min. The data are shown in Table 31.

Table 31 : Comparison of advancing contact angle between chips before and

after additional water rinsing.

pH ODP04 before

tÔ-rinsing

ODPO4 after

H20-rinsing

ODT before

H20-rinsing

ODT after

H20-rinsing

3.20 97 89 56 50

4.20 107 105 64 64

5.10 107 99 66 51

5.90 104 98 67 53

6.20 97 92 79 67

6.90 90 91 104 100

7.80 95 94 99 90

The small decrease of the contact angle after washing with water shows

that the hydrophobic ODT-surface is stable if it is in contact with water.

Long term stability studies were not performed.

183


Drying the chips after the immersion in the gold chloride solution (80°Cfor 1 hour) seems to be an important step, since gold chloride is very

hygroscopic. Some of the chips were investigated without this drying stepand a significant decrease of the contact angle was observed after rinsing the

ODT-treated chips with water (results not reported).

If the gold chloride solution is neutralized to pH 7.0 with NaOH, the

gold is in a metastable state, and the solution becomes orange within a few

days. Gold adsorbs on Ta2Os under these conditions.

The contact-angle of 104° was close to the value reported for perfectODT-SAMs on gold (111) [Hard, 1998]. However, the jidd techniqueshowed that the ODT-adlayers formed at the gold/gold chloride interface

were heterogeneous with a high jJLdd value (Figure 40). The (idd pictureshows local heterogeneities and suggests an overall low ordered ODT-

adlayer formation [Hofe,2000]

Fig. 40: Microdroplet-density picture of a gold chloride-modified Ta20^

chip coated with an octadecyl thiol SAM. The image size is 1.2 x

1.3 mm.

184


XPS:

Au4f signals are present of gold in two different oxidation states: non-

charged Au (Au4f7/2 at 84.9 eV; Au4t5/2 at 88.4 eV) and gold cations An1

(Au4f7/2 at 86.8 eV; Au4f5/2 at 90.4 eV), which probably anse from

adsorbed gold chloride (Figure 42). An overlap of two XPS-survey spectrameasured at take-off angles of 15° and 75° is shown in Figure 41. The

quantitative data are listed in Table 32. The attenuation of the tantalum

signal at the grazing take-off angle (15°) suggests gold adsorption on top of

the tantalum oxide layer.

Fig. 41: XPS survey spectra of a gold chloride coated Ta20 chip. The

spectra are measured at a 15° take-off angle (blue) and at a 75°

take-off angle (red).

185


Au4t signal of gold chloride coated Ta205chip

Au,t7/2

from Au4

94 92 90 88 86 84 82

binding energy (eV)

Fig. 42: XPS detail spectrum of the Au4f signal. The binding energies

correspond to two different oxidation states of gold which are

noted above the peaks.

The Auf peak (Au4r7/2) exhibited a chemical shift of+1.95 eV relative

to that of Au° (see Figure 42). This value corresponds exactly to that which

was found in the literature for Au1* [Sala, 1995].

Ta2Os chips that were immersed for 48 hours in a 0.4 % (w/v) goldchloride solution at a pH of 2.5 were investigated by XPS in parallel. No

gold could be detected.

186


Table 32: XPS data from TaÔs chips immersed in a aqueous 0.2 % (w/v)solution of gold chloride measured at two different take-off

angles.i

element binding energy

[eV]

atomic concentration [%]

75° take-off angle 15° take-off angle

Cls 285.00 26.3 ±1.6 51.4 ±0.2

Ou 530.50 47.2 ± 0.9 23.6 ±8.6

Ta4f 27.10 13.4 ±2.2 2.2 ±1.4

Au4f (total) 12.5 ±0.6 21.8 ±7.5

Au4f7/2 (Au°) 84.85

Au4f5/2 (Au°) 88.45

Au4f7/2 (Au,+) 86.85

Au4f5/2 (Au1+) 90.45

Cl2p 198.40 0.38 ±0.09 0.68 ±0.40

Relatively high Cis signal (contamination) was measured on the goldchloride coated Ta202 chips. The tantalum signal decreased drastically with

the grazing take-off angle, suggesting a gold chloride adlayer at the top of

the Ta2Os substrate surface.

ATM:

AFM examination indicated that gold adsorbs in the form of 10 - 20 nm

diameter islands (Figure 43) with a homogeneous distribution of gold or

gold ion islands and without building crystallites.

187


Fig. 43: AFM image of a gold/gold chloride-coated Ta2C>5 chip: Gold

chloride adsorbs onto Ta2Os in the form of 10 to 20-nm diameter

islands.

In summary, one can conclude that gold ions adsorb out of an aqueous

gold chloride solution at a pH of 6-8 onto Ta2C^ and grow in the form of

islands. The Au-islands are able to adsorb octadecyl thiol molecules in a

self-assembly process from a solution of amphiphile dissolved in ethylacetate. It is likely that Ta20, (which adsorbs ODP04 but not ODT) is

generaly present between the islands. Contact-angle data give some

information about the coverage with gold ions.

188


The waveguide properties of the Ta20s were completely changed after

the modification with gold chloride. The adsorbed gold/gold chloride

resulted in a complete absorption of the incoupled laser light. This makes the

system non-usable for the optical bio-affinity sensor method FOBIA.

Nevertheless, a more profound study of the applications of a gold ion

(or silver ion) interface in between the metal oxide waveguide layer and the

alkyl thiol SAM would be interesting. If the adsorbed noble metal (Au or

Ag) could be stabilized in the oxidized form (e.g. as a complex), thiols could

be self-assembled on evanescent field optical sensors. The advantages of

thiol chemistry (large number of commercially available functionalized alkylthiols) and the good waveguide properties of metal oxides such as TaiOs

could be combined in such a system.

Since the phosph(on)ate and the PLL-PEG chemistry constitute two systemsthat turned out to be very successful for the surface modification of metal

oxide-based chips, no further experiments were done with the gold/oxide

system.

189


2.3Stability of Modified Surfaces

Amphiphiles that were considered to be the most promising candidates

for sensor applications were investigated as regards the stability of their self-

assembled adlayers. The systems for the stability tests in water are octadecyl

phosphate (ODPO4), octadecyl phosphonate (ODPO3), dodecyl phosphate

(DDPO4) and dodecyl phosphonate (DDPO3), and the adlayer amphiphilesfor the stability studies in different buffers are ODPO4, ODPO3 as well as

the PLL(20)-g[5]-PEG(2) copolymer.

2.3.1 Stability of Hydrophobic Phosph(on)ate SAMs in Water

Ta205-chips were sonicated in 2-propanol for 15 minutes followed by

UV-cleaning for 30 minutes (standard cleaning protocol). Chips were then

immediately transferred into one of the 0.5 mM alkyl phosph(on)atesolutions for self-assembly. (Solutions: ODPO4 in n-heptane + 0.4%

2-propanol, ODPO3 in n-heptane + 1.5% 2-propanol, DDPO3 in n-heptane +

0.8% 2-propanol and DDP04(NH4)2 in water). The samples were removed

after 48 hours of immersion time, rinsed with 2-propanol and blow-dried

with nitrogen.

The stability of the hydrophobic alkyl phosph(on)ate SAMs on TaiOs

surfaces in contact with water were studied by rinsing the amphiphile-coatedchips with pure water and after an immersion time of 3 days.

Results of the SAM stability in water are listed in the appendix (App.18 and 19). Contact angle and (idd data are presented, showing similar

results as for ODPO4 (for details of ODPO4 SAM-stability in water or in

PBS, see chapter IV 2.1.1.2): A short rinsing results in a small decrease of

the contact-angle and a strong reduction of the u,dd. This is believed to be

due to removal of non-specifically adsorbed molecules. Longer storage in

water (e.g. 3 days) at room temperature reduces the average monolayerthickness by a few % and affects the homogeneity and the order within the

amphiphile adlayer, resulting in a significant increase of the (idd.

190


If the hydrophobic SAM is coated by macromolecules such as proteins(see chapter TV 3) a larger amount of alkyl phosphate molecules is

connected by hydrophobic interaction with the adsorbed macromolecules,resulting in a higher stability of the SAM.

2.3.2 Stability of Phosph(on)ate SAMs and PLL-PEG Adlayers in

Different Buffer Systems

Since the modified metal oxide surfaces may be used for DNA/RNA

sensor application studies, one candidate out of each adsorbent group (alkylphosphate, alkyl phosphonate and PLL-PEG) was investigated concerningits adlayer stability on Ta205. Amphiphile-coated Ta205 chips were, there¬

fore, immersed in different relevant buffer systems (Table 33) that are

widely used for surface chemistiy such as DNA hybridization, washing stepsand regeneration if working with immobilized recognition molecules.

Adsorbent solutions were prepared as follows: 0.5 mM ODPO4 in

n-heptane + 0.4%) 2-propanol, 0.5 mM ODPO3 in n-heptane + 1.5%

2-propanol and 1 mg/ml PLL(20)-g[5]-PEG(2) in HEPES-buffer pH 7.4.

Ta2Û5 chips were cleaned according to the standard cleaning protocolmentioned above and subsequently transferred into one of the preparedadsorbent solutions. Samples from PLL-PEG coating were removed after 2

hours from the solution, rinsed with water and blow-dried with nitrogen.Samples from phosph(on)ate self-assembly were removed from the solutions

after 48 hours, rinsed with 2-propanol and also blow-dried with nitrogen.

Two samples of each coating variant were now stored in one of the

buffer systems (see list below) for different times and different temperatures,according to the conditions that are needed for the correspondingapplications, such as oligonucleotide hybridization and regeneration.Samples were removed from the buffer solutions and extensively rinsed with

pure water (ca. 200 ml each).

191


List of applied buffer systems for adlayer stability experiments with their

used abbreviation in Tables 33 - 35 and an example of applicationconditions. Since the composition of some buffer systems is confidential, the

supplier of ExpressHyb1M is mentioned only:

Buffer Abbre¬

viation

Composition Typical conditions

MB 1,

pH7.5

MB 1 600 mM NaCl

60 mM sodium citrate

1 % Tween 20

first 15 min at 75°C

second l-2dat45°C

ExpressHyb'M

Ex.Hyb Product of Clontech

(Cat. #8015-1)

first 15 minat 75°C

second 1-2 d at 45°C

Regenerationbuffer 2,

pH7.5

Reg.2 15 mM NaCl

1.5 mM sodium citrate

1% Tween 20

50% formamid

30 min at 75 °C

WashingBuffer 3,

pH7.5

Wash 3 15 mM NaCl

1.5 mM sodium citrate

1 h at 45°C

WashingBuffer 2,

pH7.5

Wash 2 30 mM NaCl

3 mM sodium citrate

1 h at 45°C

WashingBuffer 1,

pH7.5

Wash 1 300 mM NaCl

30 mM sodium citrate

1 h at 45°C

192


Advancing water-contact angle and XPS analyses were performed. XPS

spectra were measured at a take-off angle of 15° with respect to the surface

plane to get surface-sensitive information. The results are presented in

Tables 33-35.

Table 33: Contact-angle and XPS data from ODPO4 self-assembled from a

0.5 mM solution in n-heptane/2-propanol on Ta205. Samples were

measured before and after immersion in different buffer systems

(see list of buffers in text) and under different conditions.

Buffer conditions contact angle

[°l

XPS: atomic ratio [%] Add.

Elem.time temp. [°C] C Ta O P C/Ta

before 112.2 ±0.8 85 2.9 11 1.4 29.3 no

MB1 15 min 75° C 85.3 ± 1.5 54 9.4 36 1.4 5.7 no

2d 45° C 47.5 ±2.3 42 10.0 48 0.9 4.2 Na

Ex.Hyb 15 min 75° C 56.9 ±1.7 43 10.1 44 2.8 4.3 no

2d 45° C 47.0 ± 1.0 43 9.3 45 3.1 4.7 Na

Reg.2 30 min 75° C 100.4 ±1.4 70 6.4 23 1.3 10.9 no

Wash 3 60 min 45° C 108.4 ±0.9 80 4.4 14 1.4 18.2 no

Wash 2 60 min 45° C 98.5 ±3.5 66 7.5 27 1.5 8.8 no

Washl 60 min 45° C 82.7 ±1.9 55 9.2 35 1.5 6.0 Na

193


Table 34: Contact-angle and XPS data from ODPO3 self-assembled from a

0.5 mM solution in n-heptane/2-propanol on Ta20s. Samples were

measured before and after immersion in different buffer systems

(see list of buffers in text) and under different conditions.


L°l

XPS: atomic ratio [%] Add.

Elem.time temp. [°C] C Ta O P C/Ta

before 111.7 ±1.1 83 3.5 12 1.6 23.7 no

MBl 15 min 75° C 78.7±1.8 51 9.8 38 1.3 5.2 no

2d 45° C 61.5 ±2.1 46 9.5 43 0.7 4.8 Na

Ex.Hyb 15 min 75° C 62.3 ±1.5 49 8.9 40 2.7 5.5 Na

2d 45° C 43.6 ±1.2 49 8.0 41 2.6 6.1 Na

Reg.2 30 min 75° C 102.1 ±1.4 74 5.8 19 1.6 12.8 no

Wash 3 60 min 45° C 104.1 ±1.3 72 5.9 21 1.5 12.2 no

Wash 2 60 min 45° C 105.3 ±1.3 71 6.4 22 1.5 11.1 no

Washl 60 min 45° C 84.0±1.1 54 9.5 36 1.4 5.7 Na

(traces)

The hydrophobicity (contact angele) of the phosph(on)ate SAMs is

closely related to the detected atomic concentration of carbon, tantalum and

oxygen (XPS): The stronger the decrease of the contact angle, the lower the

carbon signal and the higher the tantalum and oxygen concentration,

compared to the SAMs before immersion in the corresponding buffer. This

agrees with the suggestion of partial desorption of self-assembled alkyl

phosph(on)ate. Very high amphiphile desorption is shown when the MBl or

the ExpressHyb buffer is added. In addition strong temperature dependentadlayer corrosion is detected when samples are stored in MBl buffer. Low

desorption effects were measured for Regeneration buffer 2, Washingbuffers 2 and 3. Higher salt concentration is believed to be the reason for

higher phosph(on)ate SAM corrosion effect in Washing buffer 1.

194


The observation of a significant increase of the phosphorus concen¬

tration after immersion in ExpressHyb agrees with the assumption that the

alkyl phosph(on)ates are partially substituted by inorganic phosphate, which

is a component of the buffer. An additional peak from sodium, arising from

the buffer, could be detected with XPS on sample surfaces with stronglyreduced alkyl phosph(on)ate adlayer density.

Table 35: Contact-angle and XPS data from PLL(20)-g[5]-PEG(2) self-

assembled on Ta2Os from a 1 mg/ml-solution in HEPES buffer

before and after immersion in different buffer systems (see list of

buffers in text) and under different conditions.


I°]

XPS: atomic ratio [%] Additiomal

elementstime temp. [°C] C Ta O C/Ta

before 36.7 ±1.3 59 4.1 37 14.6 no

MB1 15 min 75° C 38.6 ±1.2 51 7.3 42 7.0 Na (traces)

2d 45° C 39.6 ± 1.2 48 7.9 45 6.1 Na (traces)

Ex.Hyb 15 min 75° C 35.2 ±2.0 36 11.6 53 3.1 P/Na

2d 45° C 29.5 ±1.8 39 10.7 51 3.6 P/Na

Reg.2 30 min 75° C 36.2 ±1.4 57 5.4 38 11.2 no

Wash 3 60 min 45° C 38.4 ± 0.9 52 7.1 41 7.3 no

Wash 2 60 min 45° C 39.3 ± 0.9 52 8.3 41 6.3 Na (traces)

Wash I 60 min 45° C .1o,/ i l.O 46 9.5 45 4.8 Na

Contact-angle data from PLL-PEG-coated chips are not a useful method

to estimate the adlayer quality. XPS on the other hand, is a surface-sensitive

method and gives adequate results for this adlayer system. XPS data

(measured at a take-off angle of 15° with respect to the surface plane) show

195


similar results as for the hydrophobic alky I phosp(on)ate SAMs. Increased

tantalum and oxygen concentrations as well as significantly lower carbon

concentrations were detected after immersion in MB1 and ExpressIIybbuffer, respectively. Smaller, but still significant desorption effects could

also be detected for the other buffer s\ stems. Appearance of phosphorus and

sodium after immersion in ExpressHyb again points to phosphate-inducedmodification of the TaiCVPLL-PEG surface,

We conclude that the adlayers (especially hydrophobic SAMs) are

reasonably stable in water but less stable in buffer systems containingdetergents.

Since the surface modifications are predominantly performed for

further immobilization of proteins (enzymes or antibodies) without

additional chemical processes, the results of the ad layer stability in different

buffer systems will not be further discussed.

Nevertheless, the data pro\ ide results that have to be kept in mind for

the biosensor surface performance of the different oxide surface modifi¬

cations (see chapter IV 3.2.4). SAMs from ODPO3 and ODPO4, are formed

from a similar solvent and under the same conditions. Quantitative C/Ta

ratio XPS results from the coated chips before their incubation in the

different buffer solutions show a significant different between the two

octadecyl phosph(on)ate systems. The contact angle shows similar

wettability for both the ODPOi and ODPO, SAMs, but the C/Ta ratio for

alkyl phosphate SAMs is significant!} higher than that for alkyl phosphonateSAMs. This cannot be explained b\ the difference in length of the unfolded

molecules, but is believed to be due to differences in the densit> of

monolayer packing.

196


3 Immobilization of Biomolecules and Quantification of

Target Molecules

Recognition molecules were immobilized via the specific binding of

streptavidin conjugated enzymes onto biotinylated surfaces or by usingstreptavidin as an interface between biotin activated surfaces and biotin

conjugated antibodies.

The densities of immobilized recognition molecules were not measured

(e.g. by the OWLS technique) since we were essentially interested in their

activity, rather in the amount of adsorbed capture molecules. The activity of

immobilized biomolecules was determined by the quantification of

converted substrate in the case of enzymes and by measuring the targetmolecule concentration that binds to the antibody sensor surface.

As described in the introduction there is a large number of different

methods for the quantification of target molecules. We decided to use two

generally different methods: an enzyme/substrate and an antibody/antigen(immunoassay) system. For the enzymatic test setup we measured the

horseradish peroxidase activity of adsorbed biomolecules on the different

self-assembled monolayers. The results were used to screen the large list of

phosph(on)ates and PLL-PEG derivatives and to select the best candidates in

terms of dose-response for future implementation.

The adsorbents with the highest sensitivity were then used for the

optical biosensor application tests. The immunoassay that was chosen is a

simple rabbit IgG - goat anti rabbit IgG system.

197


3.1Enzymatic Activity of Horseradish Peroxidase

All the enzymatic activity tests were executed according to the protocol

explained in Scheme 17:

General test protocol: Metal oxide samples were cleaned by one of the

standard cleaning methods followed by self-assembly of the adsorbents. The

samples that were coated with PLL-PEG/PEGbiotin derivatives were readyfor the immobilization of streptavidin conjugated enzyme while the

hydrophobic phosph(on)ate SAMs had to be coated with an additional layerof biotinylated bovine serum albumin (BSA-biotin). This additional layer is

needed to activate the surfaces for the immobilization of the enzyme via the

conjugated streptavidin and to create a passivation layer for non-specific

protein adsorption. The coated metal oxide samples were then transferred in

a common glass container filled with a solution of streptavidin-conjugatedhorseradish peroxidase in HEPES buffer pH 7.4. After the enzyme-

immobilization step, the samples were rinsed with HEPES buffer followed

by a second rinsing with citrate buffer pH 6.0. All samples were then

separately stored in 0.8 ml of citrate buffer until the reaction was started. 0.8

ml of citrate buffer containing EbOi (substrate) and the indicator 3,3',5,5'-

tetramethylbenzidine (TMB) were used (reaction start). The solutions,

containing the enzyme-coated samples, were gently shaken during the

reaction (solution turned slowly to blue). After 10 minutes the metal oxide

slides were removed and the reaction was immediately stopped by adding100 (il of H2SO4 (solution turned yellow). The color intensity, which is

directly proportional to the enzymatic activity, was determined by measuringthe light absorption at X = 450 nm.

198


^^^^^^P'^^^P'

samples in streptavidin conjugatedhorseradish peroxidase solution

in HEPES buffer pH 7.4

X rinse with HEPES buffer

L rinse with citrate buffer pH 6

I

each samples stored separately in

0.8 ml of citrate buffer pH 6

until reaction start

addition of 0.8 ml citrate buffer

containing EbO? and TMB

incubation by weak shaking

removal of chips and

addition of 100 ul of H2SO4

measurement of absorption at X = 450 nm

Scheme 17: Protocol of the horseradish peroxidase activity test, (TMB

tetramethyl benzidine, indicator)

199


3.1.1 Peroxidase Activity on PLL-PEG Coated Chip

Buffers and solutions were prepared as follows:

HEPES buffer pH 7.4: 10 mM HEPES, adjusted to pH 7.4 with 1 M NaOH.

Enzyme solution: Streptavidin-conjugated horseradish peroxidase

(0.3 mg peroxidase/ml and 0.1 rag streptavidin/ml)was purchased as an aqueous solution.

Enzyme dilutions: The enzyme solution described above was diluted

with citrate buffer to 20 ppm, 10 ppm, 2 ppm,

1 ppm, 0.2 ppm, 0.1 ppm and 0.02 ppm.

Citrate buffer pH 6.0: 8.2 g of sodium acetate were dissolved in 1 liter of

water (0.1 mo El) and the pH was adjusted to 6.0

with citric acid.

EECE solution: 26 pi of hydrogen peroxide (30 %) were added to

100 nil of citrate buffer (0.008 %).

TMB solution: 21.3 mg of tetramethyl benzidin were dissolved in

2.13 mfof DMSO (42 mrnol/1).

Start-solution: 20 ml of EECE solution were mixed with 400 pi of

TMB solution.

Stop-solution: 1.1 ml EESCE (98%) were diluted with 8.9 ml of

water (2 mol/1).

The activity of the immobilized enzymes (specific binding) on self-

assembled functionalized copolymer derivatives PLL(20)-g[3.5]-PEG(2)

/PEGbiotin(3.4)10% and PLE(20)-g[3.51-PEG(2)/PEGbiotin(3.4)50% was

investigated in comparison with the non-specific binding on self-assembled

non-functionalized polymer PEL(20)-g[3.5 ]-PEG(2):

200


TiOi chips (10 x 10 mm) were sonicated in 2-propanol for 15 minutes

followed by an oxygen-plasma cleaning for 2 minutes. One glass container

was prepared for each of the PLL-PEG/PEGbiotin derivatives. The metal

oxide samples were transferred into the glass containers and covered with a

lmg/ml-solution (in HEPES buffer pH 7.4) of the corresponding adsorbent.

Samples were incubated in the solutions for self-assembly over 12 hours,rinsed with HEPES buffer and transferred in a common container for

enzyme immobilization.

A solution of streptavidin-conjugated horseradish peroxidase (dilution =

1/1000 with HEPES buffer pH 7.4) was added and the samples were

immersed for 40 minutes for enzyme immobilization. The samples were

removed and rinsed with 2 ml of HEPES buffer, followed by a rinsing with

2 ml of citrate buffer pH 6.0. Each sample was now stored separately in

0.8 ml of citrate buffer until reaction start.

Each sample was covered with 0.8 ml of freshly mixed "start-solution".

The containers were weakly shaken during the incubation step. The metal

oxide chips were removed after an incubation time of 10 minutes and the

enzymatic reaction was immediately stopped, by adding 0.1 ml of "stop-solution".

A standard series of diluted enzyme solution were prepared and 0.8 ml

of each dilution were filled in an empty container instead of the 0.8 ml

portions of citrate buffer at the first step of the procedure. The procedurewas followed according to the protocol exactly the same way as for the

samples containing the metal oxide chips.

The light absorption at the wavelength of 450 nm was measured for

each sample as well as the standard series. The results of the coated metal

oxide chips were divided by the corresponding metal oxide surface areas.

The absorption, corresponding to enzyme activities of the different surface

coatings, are plotted as a function of the enzyme concentration of the

standard series (Figure 44).

201


0 00 2 00 4 00 (> 00 8 00 10 00

enzyme concentration (ppm)

Fig, 44: Peroxidase activity measurements: The activity of the immobilized

enzymes on PLL-PEG-coated TiO: chips is determined by measu¬

ring the absorption (k ~ 450 nm) of the sample solutions in com¬

parison with a standard series of different dilutions of the en/yme.

The results reflect the strong protein repelling effect of the non-functi-

onalt/ed PLL(20)-g|3.5]-PHG(2) adsorbed on TiCb while surfaces of both

PI L(20)-gl3.5j-PHG(2VPEGbiotm(3.4)10% - derivatives (synthesis I and II)show the highest activity of immobili/ed enzymes within this test series. The

PLL(20)-g[3.5]-PHG(2)/PEGbiotin-(3.4)50% derivative coated samples still

shows a higher activity of the specifically adsorbed peroxidase enzymes than

the non-specifically adsorbed enzymes on the bare metal oxide sample,

These results demonstrate that it is possible to specifically adsorb

biomolecules via streptavidin - biotin binding and to strongly reduce non¬

specific adsorption using the PI 1 --PEG copolymer as an interface between

the metal oxide waveguide layer and the capture molecules,

from streptavidin adsorption studies measured by the OWES technique,we knew that functionahzed PEL(20)-g|3.5|-PHG(2)/PEGbiotin(3.4)derivatives with a PEGbiotin concentration of l°o or less would not adsorb

enough streptavidin conjugated peroxidase to be interesting adsorbents for

202


biosensor applications. The optimum amount of PEGbiotin was therefore

believed to be higher than 1% and lower than 50%.

The conjugation ratio of streptavidin/enzyme is 1 mg streptavidin(60'000 g/mol)/3mg peroxidase (44'000 g/mol), corresponding to a

molecular streptavidin/enzyme ratio of about 1/4. The area that is covered byone streptavidin-enzyme conjugate is about 50 nm~. However, the metal

oxide substrate area that is covered by one PLL-PEG/PEGbiotin copolymermolecule is about 20 nm2. Assuming a 100% accessibility of the biotin

groups on self-assembled PLL-PEG/PEGbiotin adlayers, and no overlap of

the copolymer molecules, a concentration of 0.4 biotin groups per

copolymer would be needed to bind one monolayer of streptavidin-enzymeconjugate (occupation of 1 of the 4 streptavidin binding sites by biotin; or

0.8 biotin groups per copolymer, if 2 of the 4 streptavidin binding sites are

used for the binding). The calculation predicts an excessively high biotin

concentration of the 50-%-PEGbiotin derivative (14 biotin molecules per

copolymer molecule). An excess of biotin functionalities allows for highlydense packing of immobilized recognition elements (surface overload). This

is in good agreement with the enzyme-activity tests, where an enzyme'sstress (or steric hindrance) is believed to reduce their activity. On the other

hand, a chip that is covered by the 1-%-PEGbiotin derivative (0.3 biotin

molecules per copolymer molecule) is not able to bind enough enzyme-

conjugate to form a complete enzyme monolayer. The average biotin

concentration of 2.7 biotin groups per PLL(20)-g[3.5]-PEG(2)/PEG-biotin(3.4)10% copolymer molecule is still significantly higher than the

calculated optimum of 0.4 to 0.8 biotin functionalities per copolymer.Considering that only a fraction of the biotin molecules is accessible, which

is reasonable for relatively long side-chain PEG polymers, the calculated

value for the 10-%-PEGbiotin derivative approaches to the optimum. This is

also reflected in the experimental results.

New derivatives such as PLL(20)-g[3.5J-PEG(2)/PEGbiotin(3.4)5%,15% and 20% were synthesized according to the protocol PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4)10% II synthesis. These were directly tested with the

FOBIA system.

203


3.1.2 Peroxidase Activity on Phosph(on)ate Coated Chip

Buffers and solutions were prepared as follows:

HEPES buffer pH 7.4:

BSA-biotin:

Enzyme sol. (strept-hrp):

Enzyme dilutions T:

Enzyme sol. (hrp):

Enzyme dilutions II:

Citrate buffer pH 6.0:

H202 solution:

TMB solution:

Start-solution:

Stop-solution:

10 mM HEPES, adjusted to pH 7.4 with 1 M

NaOH.

1 mg/ml in ultra pure water

Streptavidin-conjugated horseradish peroxidase(0.3 mg peroxidase/ml and 0.1 mg streptavi-din/ml) was purchased as an aqueous solution.

The enzyme solution (strept-hrp) described

above was diluted with citrate buffer to 20 ppm,

10 ppm, 2 ppm, 1 ppm, 0.2 ppm, 0.1 ppm and

0.02 ppm for the standard series.

Non-conjugated horseradish peroxidase solution

( I kU/ml) was diluted to 1 U/ml.

The enzyme solution (hrp) described above was

diluted with citrate buffer to 20 ppm, 10 ppm,2 ppm, 1 ppm, 0.2 ppm, 0.1 ppm and 0.02 ppmfor the standard series.

8.2 g of sodium acetate were dissolved in 1 liter

of water (0.1 mol/1) and the pH was adjusted to

6.0 with citric acid.

26 (il of hydrogen peroxide (30 %) were added

to 100 ml of citrate buffer (0.008 %).

21.3 mg of TMB were dissolved in 2.13 ml of

DMSO (42 mmol/1).

20 ml of H2O2 solution were mixed with 400 julIof TMB solution.

1.1 ml H2S04 (98%) were diluted with 8.9 ml of

water (2 mol/1).

204


The activity of the immobilized enz\mes on hydrophobicphosph(on)ate SAMs was investigated: The specific binding of enzymes was

performed via activation of the SAMs with biotinylated BSA (BSA-biotin)and binding of streptavidin-conjugated horseradish peroxidase thereon. The

non-specific binding of the enzyme was studied by immersing

phosph(on)ate-coated and BSA-biotin-activated samples in a non-

streptavidin-conjugated horseradish peroxidase solution.

Four different phoph(on)ate solutions were prepared: a) 0.5 mM

ODPO4 in n-heptane + 0.4% 2-propanol. b) 0.5 mM ODP03 in n-heptane +

1.5% 2-propanol, c) 0.5 mM DDPCh in n-heptane + 0.8% 2-propanol and

d) 0.5 mM DDPOi(NTf)2 in water. T'a20^ chips (8x8 mm) were sonicated

in 2-propanol for 15 minutes followed by a oxygen plasma cleaning for

2 minutes. Five chips (three for SB- and two for NSB-studics) were

transferred in one of each phosph(on)ate solutions described above and

incubated for 48 hours for self-assembly. The chips were removed and

rinsed with 2-propanol (ODPOj, ODPO, and DDPO3) or water

(DDPO^MIfb) followed b> blow-dn ing with nitrogen.

The coated, hydrophobic chips were transferred into a common

container and covered with a BSA-biotin solution (lmg/ml in water) for

activating with biotin and passivating to improve protein resistance. The

samples were removed from the solution after 30 minutes and rinsed with

HEPES buffer.

Three samples of each phosph(on)ate SAM type were immersed in a

streptavidin-conjugated horseradish peroxidase solution (dilution of enzyme

sol. (strept-hrp) = 1/1000 in HE PES buffer pi I 7.4) for specific enzyme

immobilization (SB-stud\). Two chips of each SAM variant were immersed

in a horseradish peroxidase solution (dilution of enzyme sol. (hrp) = 1/1000

in FIEPES buffer) for non specific enz\me immobilization (NSB). Sampleswere removed after 1 hour and rinsed with 2 ml of HEPES buffer each,

followed by rinsing with 2 ml of citrate buffer pH 6. Sample were now

stored separately in 0.8 ml of citrate buffer until reaction start.

Each sample was covered with 0.8 ml of freshly mixed "start-solution".

The containers were weakly shaken during the incubation step. The metal

oxide chipsw7ereremovedafteranincubationtimeof30minutesandthe205


enzymatic reaction was immediate!} stopped by adding 0.1 ml of "sfop-solutiorf.

Two bare l^bOs samples were treated exactly the same way as the

SAM-coated chips to study the NSB of streptavidm-conjugated enzyme onto

non-coated, metal oxide surfaces,

Two standard series of diluted en/yme solutions (enzyme dilutions I

and 11) were prepared and 0.8 ml of each dilution were filled in an emptyreaction container instead of the 0.8 ml portions of citrate buffer at the first

step of the test series. The procedure was followed according to the protocolin exactly the same way as for the samples containing the metal oxide chips.

The light absorption at the wavelength of 450 nm was measured for

each sample as well as the standard series. Both standard series {strept-hrpand hrp) were correlated to one standard curve and the corresponding values

of the metal oxide samples were compared (Figure 45).

Fig. 45: Peroxidase activity measurements: The activity of the immobilized

enzymes on phosph(on)ate SAMs on TaÔ. chips is determined bymeasuring the absorption (k = 450 nm) of the sample solutions in

comparison with a standard series of different dilutions of the

en/yme. (HRP = Horseradish peroxidase)

206


Very little enzyme activity could be measured for the BSA-biotin-

activated samples that were immersed in the non-streptavidin-conjugatedhorse radish peroxidase solution {Enzyme sol, (hrp)). The average values are

indicated by orange to brown triangles. The results reflect the strong proteinresistance of hydrophobic phosph(on)ate-SAMs after passivation with BSA

(low NSB) (Scheme 18).

Mo significant difference could be detected between the enzyme activityof specific immobilized horseradish peroxidase on the different SAM-coated

and BSA-biotin activated samples. The average values are indicated with

yellow and green dots. The results demonstrate the constant degree of

immobilization on all different types of biotin activated phosph(on)ateSAMs and the high enzyme activities of the surfaces with specific bound

recognition elements (high SB) (Scheme 18).

îmr^

SB NSB

Scheme 18: Scheme demonstrating specific binding (SB) of streptavidinconjugated enzyme (strept-hrp) and non-specific binding (NSB)of bare enzyme (hrp), respectively, onto a BSA-biotin coated

phosph(on)ate SAM surface.

207


3.2Optical Biosensor Application

This chapter covers investigations of the applicability of modified metal

oxide surfaces as biosensors - immunoassay platform with the aim of

showing the high performance of such interfaces at least for one system.

Quantitative assays using rabbit IgG (RIgG) were performed with the

FOBIA system and Ta205-planar waveguide chips (for details see chapter III

1.1.4 and III 2.4.2). The samples were cleaned by sonication in 2-propanolfor 15 minutes followed by an UV-cleaning for 30 minutes. Samples were

subsequently transferred in the adsorbent solutions for self-assembly. Chipswere removed from the phosph(on)ate solution after an immersion time of

24 hours and from the PLL-PEG solution after an immersion time of 2

hours, respectively. The coated chips were then rinsed with 2-propanol

(phosph(on)ates) or water (PLL-PEG) and blow-dried with a nitrogenstream.

The concentrations (0.5 mM for the phosph(on)ates and lmg /ml for the

PLL-PEG-derivatives) as well as the solvents for the adsorbent solutions

were kept constant for all the sensor application studies: ODPO4 was

dissolved in n-heptane + 0.4% 2-propanol, ODPO3 solution was prepared in

n-heptane + 1.5 % 2-propanol, DDP04(NH4)2 was dissolved in water and

PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4) in PBS buffer.

3.2.1 Labeling of Biomolecules

The quantification of biomolecules with the optical analysis apparatusFOBIA needs a fluorescent marker. Cy5 is such a molecule that is widelyused for the fluorescence labeling of cells [Ball, 1995], proteins and other

biomolecules [Yu,1994]. The advantage of Cy5-dye is its high quantum

yield and the simple labeling protocol.

Cy5 is commercially available and the labeling has been done according to

the protocol of the supplier [Amer,2000]:

208


For the labeling of the target molecules, 1.0 mg RIgG was dissolved in

1.0 ml of carbonate buffer pH 9.2 (Na2CO, 0.1 M and MaHCO, 0.1 M mixed

in a volume ratio of 1 : 9). The RIgG solution was transferred into a Cy5-vial (already prepared for use from the supplier), well shaken and stored for

30 minutes in the dark at room temperature (vial shortly shaken once every

10 minutes). The Cy5-conjugated RIgG was then cleaned by size exclusion

chromatography using a pre-conditioned PD-10 Sephadex G-25 M column

(pre-conditioning with 50 ml PBS). The pure product was then eluted by

adding portions of I ml PBS. The 1-mi-fractions were collected separately in

plastic tubes.

The determination of Cy5-RlgG concentration (mol/1) as well as the

tracer-to-biomolecule ratio were performed by measuring the absorption at

two different wavelengths (280 nm and 650 nm) and applying the simpleequations 6 to 8 with molar extinction coefficients of 250'000 and 170'000

for Cy5 and RIgG, respectively.

[Cy5-dye] = (A650)/250'000 equation 6

IRlgG] = [A28o-(0.05*A6<„)l / 170'000 equation 7

The determination of Cy5/RIgG ratio is a simply division of the

calculated concentrations of the two components (equation 8).

Cy5/RIgG = [Cy5 d\ ej / [RIgG] equation 8

The fractions No 3, 4 and 5, containing the product, were mixed and

aliquots of 100 pi were frozen and stored at -5° until use. The Cy5/RIgGratio has been calculated to 4.0.

209


3.2.2 Specific and Non-Specific Binding of Target Molecule

The performance of the different self-assembled interfaces (phosph(on)-ates and PLL-PEG/PFXTbiotm dematives) was tested by measuring the

fluorescence of Cy5-labeled target molecule. Cy5-RlgG, The signal intensityis proportional to the amount of bound Cy5-RlgG, where the binding onto

its immobilized antibody is called specific binding (SB) and all the bindingonto other surfaces are called non-specific binding (NSB). One of the

important performance criterion is the ratio SB/NSB,

The test series is executed according to the following protocol:

Four types of interfaces were prepared by self-assembly of 01>J\

)WOx and MDIVMNlhV as well as

onto freshly cleaned chip surfaces.

V *u

„*,i\]U ..., v* .»V

"fhe pruNph(onMU'-coatcd chips were activated by immersing in a BSA-

solution (1 mg/ml m PBS) for 2 hours and rinsing with water. Possible

remaining non-coated areas of the chip were back-filled with BSA for

passivation by immersing m a BSA solution of 10 mg/ml in PBS.

The chips were then mounted m the FOB1A apparatus, where a flow

cell is pressed against the sample surfaces.

210


Subsequently, 0.9 ml of a C\5-RlgG solution

(0.25 nM in PBS) was pumped through the flow

cell. The recorded signal reflects NSB of targetmolecule to BSA-biotin and PLL-PEG/PEGbiotin,

respectiveh (Since the assay is the same for both

the phosph-(on)ates and the PLL-PEG/PEGbiotin

coated chips, the adsorbents are indicated by the

only, for simplicity.)

0.9 ml of a solution of • \ (100 ug/ml in

PBS) was pumped through the How cell. The chipsurfaces were coated with i

^ >

,which builds

an interface between the activated chipsurface and the biotmylated antibody.

Again, 0.9 ml of the C\5-RIgG solution (0.25 nM in

PBS) was pumped through the flow cell. This time

the recorded signal corresponds to NSB of targetmolecule to o ^ \ v .

To prepare a real, active biosensor chip, 0 9 ml of a

biotin-conjugated antibody (biotin-aRlgG, Y)solution (70 (.ig/ml in PBS) was pumped through the

How cell. The antibody is believed to be immobil¬

ized on the surface in an oriented way via the

specific and stable biotin- binding.

The SB of target molecule onto the antibodyactivated sensor surface was tested by pumping0.9 mi of the Cy5-RlgG solution (0.25 nM in PBS)

through the flow cell.

211


Figure 46 shows the results of test series for the ODPCVcoated chip.The concentration of 0.25 nM (= 2.5 x 10"1() M) Cy5~RIgG is close to the

upper detection limit. Application of 2.5 nM solution of labeled targetmolecule gave SB-signals that were outside the linear response regime of the

optical device of the FOBIA system.

Data points 1-19 are registered during the preconditioning phase, which

is a simple rinsing step with assay buffer solution (PBS). It is essential to use

exactly the same buffer mixture for the assay run cycles as well as for

preparing the solutions (Cy5-RIgG, streptavidin and aRIgG) to get a

quantitative signal not disturbed by change of the refractive indices of the

applied solutions. Any change of refractive index of the solution would

affect the result.

Data points 20-37 show the total fluorescence intensity from both the

bulk within the evanescent field volume and the molecules attached at the

sensor surface. We call this cycle time interval adsorption.

Data points 38-50 are registered during a second rinsing step with assaybuffer. A decrease of the signal intensity within this time interval may occur

due to desorption of fluorescent molecules (e.g. labeled tracer molecules)and bleaching of their fluorescence. We did not distinguish between these

two effects and call this time interval desorption.

The most interesting time window of the cycles is therefore that around

the data points 39-41, where bulk solution has already been replaced byassay buffer (no bulk signal) and a minimum bleaching effect has to be

considered.

212


0)o

Su"A

Ü

pû-r

Fig, 46:

base hue

..—2 5e-IO\lCy5~RlgG

streptaudm

—-2^6-10 \HV5-RlgG

——biotm-aRlgG

~ 2 5e-lOMCV5-Rl«G

data points (time)

TOB1A data from an ODPOreoated TaÔ^ planar waveguide chip:The different steps of the sensor surface preparation and the

application of labeled target molecule (Cy5~RIgG) solution after

every modification is described in the text. Non-specific binding(NSB) of Cy5-RlgG is determined on the BSA-biotin-activated

surface (light blue) and after functionalization with streptavidin(deep blue). The specific binding (SB) after immobilization of

capture molecule. aRfgG, is indicated by the brown line. Baseline

stability of the assay buffer (green) as well as self-fluorescence of

streptavidin (pink) or aRfgG (black) are also shown.

The base line (PBS rinse) reflects the stability of the optical system.The line (yellow) is hardly visible. Data points of the first ran with Cy5-RlgG (light blue line) are almost congruent with the base line, demonstratingthe effective protein resistance of albumin. Conjugation with biotin (ca, 11

biolin molecules per albumin) does not affect the protein-repelling behavior.


Streptavidin (pink line) shows a small signal increase from self-

fluorescence. This effect could be detected in all experiments. The signalonly increases at the beginning of the streptavidin adsorption and the

fluorescence intensity returns to the base line level before the end of the

streptavidin solution flow. The conclusion is a fast bleaching of the self-

fluorescence within the evanescent field. Once the surface is completelycovered with a streptavidin layer no more streptavidin adsorbs and no

additional fluorescence occurs.

The small signal increase at the beginning of the subsequent cycle with

Cy5-RIgG (deep blue line) suggests a small target - streptavidin NSB.

However, it must be a weak binding, for the signal returns to the base line

level before the desorption step.

The immobilization of capture molecule, aRIgG, (black line) does not

show any significant change in the signal intensity (no self-fluorescence).

The strong fluorescence signal in the adsorption interval of the last test

cycle with Cy5-RIgG (brown line) reflects the high specificity of the sensor

in combination with the FOBIA system.

Similar results were obtained from sensors which were prepared with

self-assembled ODPO,, DDP04(NH4)2 and PLL(20)-g[3.5]-PEG(2)/PEG-biotin(3.4)l 0%. The results are presented in the appendix (App. 20 - 22).

All investigated surfaces show relatively low NSB- and high SB-

signals. Very low fluorescence intensity at the first cycle of Cy5-RIgGapplication suggests got protein resistance for BSA passivated phosph-(on)ate SAMs as well as for the PTL-PEG-derivative.

214


3.2J Specific and Non-Specific Binding of Streptavidin

As mentioned in the previous chapter, streptavidin is used as an

interface between the BSA-biotm coated phosph(on)ate SAMs (or PLL-

PEG/PEGbiotin modified surfaces) and the biotin-conjugated antibody,biotin-aRIgG (Scheme 19).

cizzzzZ22zzrö22ZZzzz/

MJMMUUUJMIL,

biolin-antibody

streptavidin

BSA-biotin

hydrophobic SAM

Scheme 19: Schematic drawing of streptavidin interface between biotin-

conjugated BSA coated hydrophobic SAM and biotin-

conjugated antibody

The streptavidin accessibility of biotin-activated (or biotinylated)surfaces and the protem resistanee of the same non-biotinylated surface

variant was investigated with one example of hydrophobic SAM (01)P04)and with aPLL-PhG/PEG-biotin modified surface-

ODP04: Ta2Os sensor chips were cleaned according to the standard

procedures (someation in 2-propanol followed by UV cleaning),The chips were subsequently immersed m a 0.5 mM solution of

ODP04 in n-heptane + 0.4% 2-propanol for 24 hours, rinsed

with 2-propanol and dried with nitrogen.

215


For biotin activation and specific streptavidin binding (SB) one

chip was incubated in a BSA-biotin solution (1 mg/ml in PBS)for 1 hour, rinsed with PBS and immersed in a BSA solution

(10 mg/ml in assay buffer) for back-filling of any free spaces at

the hydrophobic sensor surface.

For the non-specific binding (NSB) test, another ODPO4-SAM-

chip (preparation as above) was directly incubated in BSA

(without biotin) solution (10 mg/ml in assay buffer) for 1 hour.

PLL-PEG: Ta20s sensor chips were cleaned according to the standard

procedures.

For the investigation of SB, one chip was subsequentlyimmersed in a PLL(20)-g[3.51-PEG(2)/PEGbiotin(3.4)10%solution (1 mg/ml in PBS) for 1 hour and rinsed with assay

buffer.

For the NSB test, one TÔ, chip was incubated in a PLL(20)-g[3.5]-PEG(2) (without biotin) solution for 1 hour and

subsequently rinsed with assay buffer.

The chips were then mounted in the FOBIA apparatus and three cyclesof Cy5-streptavidin (3.3 pM in assay buffer), 0.9 ml each, were executed

(Figure 47). A rinsing step with pure assay buffer was introduced before

(pretreatment) and after (desorption) each Cy5-streptavidin application.

216


a Ovcle 1

H Cycle 2

SB fo! ODP04 NSB lot ODP04 SB lot PT L- NSB loi PI T -PL G

PBObiotmi

Fig 47 TOBTA data of specific binding (SB) and non-specific (NSB)studies of fluorescence-labeled streptavidm on a BSA-biotm

activated ÜÜPO4 SAM (SB for ODP04), a BSA passivatedODPO4 SAM (NSB for ODP04). a PLI (20)-g[3 5]-PT G(2)/PPG~biotm(3 4)I0% la\er (SB for PL I-PLGbiotin) and a PU (20)-

gL3 5]-PFG(2) layer (NSB for PI L-PFG)

I he NSB signal for streptavidm is very low (0 12 % and 0 08 % of the

SB signal intensity in the case of ODP04 and PU -PFG, respectively), while

the SB signal is high for both modified surfaces Fhe total signal intensitydoes not reilect a complete streptavidm surface coverage but can be referred

to the relative amount of adsorbed protein A complete Cy5-streptavidmcoverage of the sensor surface would exceed the upper detection limit of the

system Since the streptaMdm adsorption studies has been perlormed on

individual waveguide chips, having different optical properties (such as

difference m light diffusion or m grating quality), it is not possible to

quantitatively compare the two different sut face modification models

©

a

0>

217


3.2.4 Detection Limit of Target Molecule

The performance of the different self-assembled interfaces

(phosph(on)ates and PLL-PEG/PEGbiotin) in terms of the lower detection

limit was tested by measuring the fluorescence of Cy5-labeled target

molecule, Cy5-RIgG. The sensor surfaces are activated by a layer of

immobilized, biotinylated antibodies, biotin-aRIgG.

The antibody immobilization was executed according to two different

methods, corresponding to the two different types of surfaces: The

hydrophobic phosph(on)ate SAMs were coated by BSA-biotin followed byformation of a streptavidin interfacial adlayer. Self-assembled layers of

PLL-PEG/PEG-biotin derivatives can be directly coated with the

streptavidin adlayer. Both types of the phosph(on)ate SAMs as well as the

PLL-PEG/PEGbiotin samples are now ready for the antibody immobili¬

zation via the biotin - streptavidin biding. The process is described more in

detail in the previous chapter.

The test setup does not correspond to a real assay, because in a real

application it is not possible to directly label the target molecule.

The chemicals and dilutions that were used for the investigation of the

detection limits with their abbreviations are listed in Table 36.

Ta205 planar waveguide chips were sonicated for 15 minutes in

2-propanol followed by an UV-cleaning for 30 minutes. Subsequently, the

chips were separately incubated in one of each adsorbent solution.

The PLL-PEG/PEG-biotin coated chip was removed from the adsorbent

solution after 1 hour, rinsed with water and was ready for use.

The three phosph(on)ate chips were removed after 24 hours from the

amphiphile solution and rinsed with 2-propanol (ODP04, ODP03) or water

(DDP04(NH4)2). The hydrophobic SAM-chips had to be activated by

coating with a BSA-biotin adlayer (immersion of the chips in BSA-biotin

solution for 1 hour) followed by rinsing with 5 ml of PBS. Additional

immersion in a BSA solution was executed for back-filling of any free

218


adsorption areas at the hydrophobic SAM surfaces (passivation for non¬

specific protein binding).

Table 36: List of chemicals and dilutions with their abbreviation.

abbreviation Reagents / Delution

ODPO4 solution 0.5 mM in n-heptane + 0.4% 2-propanol

ODPO3 solution 0.5 mM in n-heptane + 1.5% 2-propanol

DDP04(NH4)2 solution 0.5 mM in water

PLL-PEG/PEGbiotin10%

solution

PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4)10%1 mg/ml in PBS

BSA-biotin 1 mg/ml PBS

BSA 10 mg/ml assay buffer

Assay buffer PBS + tween 20 (0.5 ml/L)

Streptavidin 0.1 mg/ml in assay buffer

a-RIgG-biotin 0.12 mg/ml in assay buffer

Sample 1 Cy5-RIgG, 50 fM in assay buffer



Sample 4 Cy5-RIgG, 2.5 pM in assay buffer

Sample 5 Cy5-RIgG, 5 pM in assay buffer




Chips were mounted in the FOBIA apparatus and the test program

could be started.

219


The solutions and Cy5-R]gG samples were pumped through the flow

cell one after the other. An additional rinsing with assay buffer was executed

before (precondition) and after (des orpt ion) each solution or sampleapplication (adsorption), fhe test series was executed according to the

following protocol:

cycle N° Sample description Surface reaction

1 Assay buffer preconditioning

2 Streptavidin actuation for antibody immobilization

3 oc-RlgG-biotin aeti\ ation w ith antibody

4 Assay buffer blank (0.0 IM standard)

5 Sample 1 50 fM standard



8 Sample 4 2.5 pM standard

9 Sample 5 5 pVJ standard

10 Sample 6 25 pM standard



The data point No 18 (end ofprecondition) was used to define the zero

point of the scale and standard cycles are plotted in parallel. Figures 48a

and b show an example of such an immunoassay on an ODPOi-coated chip.The other adsorbent-covered sensors gave similar results (curves presentedin the appendix, App. 23 - 25). The average value of measuring pointsNo 40 - 45 within the desorption-time window were used to determine the

dose - response values. Figure 49a and b show an overlay of the dose-

response curves from two ODPO»-coated chips and two PLL-

PEG/PEGbiotin-coated chips.

220


a

vi

.a

o

SO

a>

O

tllittilttWllilMM

>>

CD

O

o

-blank

-M)(M

-2^0 (TVI

-500 I'M

-2 5 pM

-5 0 pM

-2^p\[

o0p\l

-2^0 pM

measuring points

measming points

Figure 48a and b: FOBIA results from bioün-aRIgG/Cy5-RIgG immuno-

assa} on streptavidin-activated ODPOj-sensor surface

(application of Cy5-RTgG different concentrations:

0 - 250 pM). Data point No 1 8 (end of precondition) is

used to define the zero point of the y-scale. Figure 48b

shows the zoomed loAver standard concentration region(zoom factor = 40).

221


a

Ol

Soo

Ö<a

o'A

<0n

O

M

-fr

0>

o

r-t

©o!»

<D•H

O

-PLt-PFGbiotina

-PLL-PtGbiotmb

-0DP04a

-ODP04b

50 100 ISO 200

C>5-RIgG concent!ation (pM)

250

-PIL-PIGbiotina

-o-PLL-PEGbiotmb

-»-ODPCMa

-0-ODPO4b

Cy5-RIgG concentiation (pM)

Figure 49a and b: Dose -

response curve from biotin-aRIgG/Cy5-RIgGimmunoassay on streptavidm-activated sensor surfaces.

The net signals of two chips of each coating types

(PLL(20)-g[3.5]-PEG(2)PEGbiotin(3.4)10% and ODP04,respectively) with the concentration of tracer labeled

target molecule (Cy5-RJgG). Figure 49b shows the lower

pM to fM region.

222


The fluorescence signals are significantly higher than the background(intensity measured at data points 1 to 18), even for the lowest standard

concentration. However, the dose - response curve is not linear in this low

concentration region (Figure 49b).

Within the series of phosph(on)ate SAMs, the ODPOrcoated sensor

showed much lower sensitivity compared to the other hydrophobic-coatedmetal oxide surfaces (Figure 50). The stability of ODP03 SAMs in water

(App. 18 and 19) and their wettability (Table 21) have been shown to be

similar to the corresponding test data from self-assembled monolayers from

ODPCH and DDPOt(NH4)2. Ho\\e\er, the C/Ta ratio (XPS data, Tables 33

and 34) showed a significant difference between ODPO3 and ODPO4 SAMs.

Lower packing density of the OÜPO, amphiphile monolayers is concluded.

The existence of a correlation between the difference in the head group

(phosphate or phosphonate) and the SAM quality is not yet proven and will

be investigated in the future. However, since we already had two successful

hydrophobic systems (ODPO4 from organic sohcnt amphiphile solution and

DDPO4 from aqueous solutions of the ammonium salt) we decided to con¬

tinue the investigations without further stud> of the ODPO3 coating option.

0 50 100 |so 200 250

C\ ^-RlgG coneentiation (pM)

Fig. 50: Dose - response cune from biotin-aRlgü/Cy5-RlgG immuno¬

assay on strepta\idin-acti\ated phosph(on)ate-scnsor surfaces. The

net signals are plotted as a function of the concentration of tracer-

labeled target molecule (Cy5-RJgG).

223


The detection limit for ODP04 or DDP04 coated sensors is in the low

femtomolar region if directly labeled target molecules are applied. The PLL-

PEG/PEGbiotin coated sensor is a little bit less sensitive, compared to the

two hydrophobic SAM sensors, but the detection limit is still in the

femtomolar region.

As mentioned above, the assay described in the preceding paragraphdoes not provide the effective detection limits of a real immunoassay. But

the results give important inputs for the decision which interfaces are most

promising as a platform technology to build real sensor systems.

3.2.5 Optimization of Biotin Concentration in PLL-PEG Derivatives

The optimum of grafted PEGbiotin in the PLL-PEG/PEGbiotin

copolymer has been shown to be between 1% and 50% for the

immobilization of horseradish peroxidase, where the PLL(20)~g[3.5]-PEG(2)/PEGbiotin(3.4)10% derivative showed the highest enzyme activity(see chapter IV 3.1.1). Similar to the procedure of the enzymatic activitymeasurements (chapter IV 3.1) the performance of the different PEGbiotin

concentrations in the copolymer derivatives were studied using the biotin-

ocRIgG/RIgG-CyS immunoassay on streptavidin activated PLL-PEG/PEG¬

biotin surfaces.

Solutions of PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4)l%, 10% and 50%

were prepared (1 mg/ml in PBS) and clean Ta205 waveguide chips (standardcleaning procedure as mentioned above) were immersed for 2 hours. The

chips were washed with pure water, blow-dried with nitrogen and introduced

into the FOBIA apparatus.

Chemicals and dilutions as well as the assay protocols were executed

according to the information given in Table 36 and in the text of the

previous protocol.

224


Similar to the enzymatic activity measurements, the sensor chip coated

with the 10% PEGbiotin copolymer (with respect to the total number of PEG

chains) showed the best results within this series of experiments (Figure 51).

0 25 50 75 100 125 150 175 200 225 250

Cy5-RIgG concentration (pM)

Fig. 51 : Dose - response curve from biotin-aRIgG/RIgG-Cy5 immuno¬

assay on streptavidin activated PLL(20)-g[3.5]-PEG(2)/PEG-biotin(3.4)l%, 10% and 50% surfaces. The net signals are plottedas a function of tracer-labeled target molecule concentration (Cy5-RlgG).

Significantly higher sensitivity for the 10% derivative compared to the

1% and the 50% and a linear dose - response curve down to the femtomolar

to low picomolar concentration was obtained.

To increase the sensitivity and/or to find the optimum PEGbiotin

concentration, three new copolymer derivatives were synthesized: PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4)5%, 15% and 20% (synthesis protocol see

chapter III 1.2.2).

225


Ta205-chips were coated with the new copolymers and the biotin-

ocRlgG/RIgG-Cy5 immunoassay was performed on the sensors after

activation with streptavidin as described above (application of a standard

series consisted of a blank followed by a 50 fM, 500 fM, 5 pM and 50 pMRIgG-Cy5 solution). The results were compared to these of the PLL-

PEG/PEGbiotinl%, 10% and 50%) derivative coated sensors. The relative

fluorescence intensities of the 50 pM RIgG-Cy5 standard solution are

presented in Figure 52.

1 5 10 15 20 50

PEGbiotin concentration (%)

Fig. 52: Results of biotin~aRIgG/RIgG-Cy5 immunoassay on streptavidinactivated PLL(20)-g[3.5]-PEG(2)/PEGbiotin(3.4) derivative

surfaces with different PEGbiotin concentrations. The relative

fluorescence signal of the 50 pM Cy5-RIgG solution is plottedwith the PEGbiotin concentration.

The results demonstrate the highest sensitivity for the 15-%~PEGbiotin

derivative-coated sensor.

226


3.2.6 Immunoassay Design

As mentioned above, the recognition and quantification of labeled

target molecules do not correspond to a real immunoassay, since it is not

realistic to perform a direct fluorescence molecule conjugation of targetmolecule traces in a matrix such as blood or urine samples.

Two protocols that form the basis for potential bioassays are the

following (Scheme 20):

a) The sample, including target molecules (e.g. unlabeled RTgG), is appliedat the recognition molecule acthated sensor surface. Target molecules

bind to the immobilized recognition molecules (e.g. aRIgG antibody)and the surface is subsequent!}' rinsed with buffer to remove non-

specifically adsorbed target molecules. Fluorescently labeled tracer (e.g.

Cy5-cxRIgG) is now applied, which specifically binds to the free

accessible moieties of the previoush bound target molecules at the sensor

surface. The surface is rinsed again to reduce bulk signal.

b) The sample is first incubated with an excess of tracer (e.g. Cy5- oRIgG

antibody) where RIgG reacts with the antibody and forms a

RIgG/aRIgG-Cy5 - complex. This reaction solution is subsequently

applied to the sensor surface (consisting of immobilized recognitionmolecules, aRTgG). The sensor surface is rinsed to reduce bulk signal.

Both designs correspond to a sandwich immunoassay system.

">">7


#*#*#

0

+ fmm

0 ¥

YYYYYYY YYYYYYY

a) b)

RIgG/aRlgG-Cy5 complex (e.g. polyclonal Ab)

• cxRlgG-Cy5 tracer (e.g. polyclonal Ab)

o Target molecule (e.g. RlgG)

p^^n Immobilized aRIgG (e.g. monoclonal Ab) at sensor surface

Scheme 20: Schematic drawing of two immunoassay designs: a) Targetmolecules are first bound to the antibody (e.g. monoclonal Ab)functionalized sensor surface. Subsequently an excess of tracer

molecules (e.g. Cy5-labeled polyclonal Ab) is added for bindingto the surface adsorbed target molecules, b) Tracer (e.g. Cy5-polyclonal Ab) and target molecules are first incubated and the

target/tracer complex is added to the sensor surface for bindingto the (monofunctional) antibody-functionalized sensor surface.

22o


Ta2Os planar wave guide chips were cleaned according to the standard

cleaning procedure and coated with a DDP04 SAM by immersion of the

chips in a 0.5 mM DDPO^NH^ solution for 24 hours. The hydrophobic

chips were then immersed in a BSA-biotin solution (0.1 mg/ml in PBS) for

1 hour, rinsed with water and passivated by an additional immersion in a

BSA solution (10 mg/ml in assay buffer) for 10 minutes. The sensor chipswere subsequently introduced in the FOBIA apparatus.

The series of applied solutions within the assay design a) were

performed as follows: (Table 37)

Table 37: Cycle description for testing the immunoassay design a)

Cycle N° Sample description

1 Assay buffer

2 Streptavidin (0.1 mg/ml in assay buffer)

3 biotin-ocRIgG (12 |JLg/ml in assay buffer)

4 RIgG standard 1 (blank)

5 Cy5-ocRIgG tracer ( 667 pM in assay buffer)

6 RlgG standard 2 (67 fM in assay buffer)

7 Cy5-otRIgG tracer ( 667 pM in assay buffer)

8 RIgG standard 3 (667 fM in assay buffer)

9 Cy5-ooRlgG tracer ( 667 pM in assay buffer)

10 RIgG standard 4 (6.7 pM in assay buffer)

11 Cy5-ocRIgG tracer ( 667 pM in assay buffer)

12 RIgG standard 5 (67 pM in assay buffer)

13 Cy5-aRIgG tracer ( 667 pM in assay buffer)

229


The series of applied solutions within the assay design b) is much

shorter since the tracer has not to be introduced after every cycle, but is

mixed with the samples. Standard solutions with the same RIgGconcentrations as in test series a) were mixed with a 667 pM solution of

Cy5-aRlgG tracer in a ratio of 1:1 (v:v). The tracer/sample mixture was

stored in the dark for 1 hour for pre-incubation, and the same volume

(corresponding to half of the concentration) as in the test series a) was

pumped through the flow cell.

The test series of assay design b) was executed according to Table 38.

Table 38: Cycle description for testing the immunoassay design b).

Cycle N° Sample description

1 Assay buffer

2 Streptavidin (0.1 mg/ml in assay buffer)

3 biotin-aRlgG (12 p,g/ml in assay buffer)

4 Standard 1: RlgG/Cy5-aRIgG (blank/667 pM) 1/1

5 Standard 2: RIgG/Cy5-aRIgG (67 fM/667 pM) 1/1

6 Standard 3: RIgG/Cy5-aRlgG (667 fM/667 pM) 1/1

7 Standard 4: RIgG/CyS-aRIgG (6.7 pM/667 pM) 1/1

8 Standard 5: RIgG/Cy5-aRlgG (67 pM/667 pM) 1/1

In both series of different assay design, a rinsing step with assay buffer

was applied before and after every cycle.

The fluorescence intensity was measured after each cycle and the

results of both test series were compared (Figure 53). Due to the dilution

effect by the mixture with tracer solution at the pre-incubation step in test

series b), just half of the RIgG concentration was present at the correspon¬

ding standard solutions compared to these of the test series a). Nevertheless,the signal was dramatically higher if the assay design b) was applied.

230


0 HL 20^

30 JO 50 60,

7J0

RIgG concentration (pM)

RIgG concentration (pM)

Fig. 53: Fluorescence intensity of tracer (Cy5-labeled aRIgG antibody)bond to target molecules (RIgG) that have been adsorbed at the

sensor surface in a previous step (assay design a), triangles)compared to the fluorescence intensity of target/tracer - complexafter pre-incubation of RIgG and aRlgG-Cy5 (assay design b),squares). The x-axis represents the initial target molecule

concentration of the sample solutions.

231


In comparison it is obvious that the pre-incubation of target molecules

with fluorescently labeled antibody (tracer) molecules followed by the

reaction of the target/tracer - complex with the sensor surface is much more

efficient than the stepwise application of target molecules at the antibody-functionalized sensor surface followed by the reaction of added target and

finally binding the tracer at the sensor. Concerning the RIgG concentration,one has to note that the pre-incubation in design b) leeds automatically to a

dilution of 1:1. The effective applied target molecule concentration is

therefore twice as much at the assay design a) compared to the design b).Nevertheless, the sensitivity for design b) is much higher and the detection

limit (ca. 20 fM) is significantly lower than that for design a). The curve

shows a relatively linear dose - response dependence.

Since the pre-incubation method (design b) is more sensitive and less

time consuming (almost half of the cycles within a test series) we decided to

continue with this assay design for the final step: the investigation, if the

sensor set-up also works for total serum applications.

3.2.7 Application of the Sensor for Serum Samples

The last step, namely the analysis of real biological samples such as

serum or urine is often a real challenge for the applicability of a sensor. Non¬

specific adsorption of proteins, the difference in ion composition of the

measured solutions or the difference in viscosity and diffusion behavior maybe factors that result in lower sensitivity, specificity and/or reproducibility.

Since such a sensor should be usable for different individuals (differentanimals or different patients) an excess of BSA was added to the previouslyused PBS assay buffer. An addition of 1 gram BSA per liter assay buffer

guarantees low relative deviation of the total protein concentration from

sample to sample (minimization of matrix effect). Tween 20 is another

additive, which is introduced to prevent protein adsorption as well as bubble

formation on the walls of the tubes and the flow cell within the apparatus.The new assay buffer was therefore: PBS containing 0.5 ml Tween 20 and 1

g BSA per liter.

232


The three adsorbents for the metal oxide modification showing the

highest sensitivity in the screening FOB!A tests (ODPO4, DDFO^NH^ and

PLL(20)-g[3.5]-PEG(2)/PEGbioti"n(3.4)15%) were chosen for this Unat

evaluation and real assay proof-of-eoncept.

Amphiphile solutions of ODPO4, 0.5 mM in n-heptane 4 0.4%

2-propanol; DDPO^NHfli, 0.5 mM in water and a tmg/ml - solution of

PEE-PEG/PEGbiotin in PBS were freshly prepared. Planar waveguide chips

(TaaO^) were cleaned according to the standard cleaning procedure

(sonication in 2-propanol followed by UV-cleaning) and subsequentlytransferred into one of the corresponding adsorbent solutions for self-

assembly.

The PLE-PEG/PFGbiotin coated chips were removed after I hour and

rinsed with water.

The chips incubated in the ODP04 or DDPO^NH^ solution were

removed after 24 hours and rinsed with 2-propanol and water, respectively.The hydrophobic surfaces were actis ated with biotin by immersion in a

BSA-biotin solution (0.1 mg/ml in PBS) for 1 hour, washed with water and

immersed in assay buffer (containing BSA) for passivation to reduce non¬

specific binding.

The chips were then mounted in the FÖB1A apparatus and the test

series was performed according to the assay design b) described in the

previous chapter, except that the RlgG solutions are now prepared in 10%,

25%, 50% or 100% newborn calf serum. The samples (RIgG dilutions) were

mixed with fluorescent labeled aRlgG solution (670 pM in assay buffer

without BSA) in a 1:1 ratio. The sample - tracer mixtures were stored in the

dark for RIgG/aRIgG-Cy5 complex formation for 30 minutes. Test series

were performed as described in Table 38 of the previous chapter.

The results of the assay series from 10%, 50% and 100% sample serum

concentration, respectively, were compared to the results of the same RIgGdilutions prepared in pure assay buffer containing no calf serum. The dose-

response curves are plotted for all newborn bo\ine serum concentrations in

parallel for the two types of interface architecture (figures 54 and 55).

1j j


3.2.7.1 PLL-PEG Sensor Interface

The results of PLL(20V[3 5]-PEG(2)PHGbiotm(3.4)15% modified

sensors are discussed and presented in Figure 54-

0% Sei um

s 10% Seium a)

i?t* 10% Serum b)

% 10% Setum c)0}

-A- 50% Sei um.3

100% Sei umo

CD

5^

O

ê \ï^ „^, i-.ww.wW««wm**

0 10 20 30 40 50 60

RlgG concentration (pM)

Fig. 54: Dose - response curve for a RlgG immunoassay on streptavidinactivated PLL(20)-L3.5|-PHG(2)/Pi:übiotin(3.4)15% surface, fhe

different serum concentration reflect the amount of newborn calf

serum within the RlgG dilutions, followed by 1:1 mixture with

Cy5~aRIgG~traeer solution of each sample.

The results presented in Figure 54 indicate non-specific adsorption of

components from the newborn calf serum and/or slower adsorption kinetics

due to increased viscosity. The reduction of sensitivity is not linear. Addition

of 10 % serum to the assay buffer already results in a decrease of the 67 pMstandard signal of about 50 %. The dose-response curve from the 50 %

serum series is almost congruent with that from the 100 % serum series, fhe

fluorescence intensity data of all series showed an almost linear dependenceon the RlgG concentration and the detection limit is still in the low pMregion. Even if the detected signal intensities of the 100 % serum series is

only about half of the corresponding signal intensities of the 10 % scries,

234


one has to note thai the net sensitivity referred to the target/serumconcentration (the present assav data show the target/sample concentration)

is about five times higher for non-diluted samples (100% serum series), For

a real patient sample application, the target/serum concentration is the

determining factor.

3 2 7.2 H} drophobic Alkyl Phosphate Sensor Interfaces

The immunoassay results using DDPCH SAMs (front aqueous

DDP04(NPIf)> solution) or ODPOj SAMs as platforms are presented in

Figure 55 and App 26, respective!}

0 °o Set urn

\ 10° o Set urn — _ ^

25% Set urn

-A-SO0 o Set urn

0)

q

100° o Set urn L0)

5

O

ë

ST-"L _

<) 10 20 10 40 SO

R gG concentration (pîvf)

--Î-

60

Fig 55 Dose - response curve for a R[g(i immunoassay on BSA-biotm -

streptavidin activated I)DP04(NHfb surface Ihe different serum

concentrations reflect the amount of newborn calf serum within

the RlgG dilutions, followed b\ 1 1 mixture with Oy5-ccRIgG~tracer solution of each sample The 10 % serum curve correspondsto the average of 4 different series on separate sensors with the

error bars

235


The dose-response curve for a RIgG immunoassay on BSA-biotin -

streptavidin activated ODPO4 surface (see appendix, App. 26) gave similar

results as the test data from the DDPO4 based sensor surface, which are

presented in Figure 55. The dose-response deviation between the 0%, 10%,

25% and 50% serum concentrations is within the error range of the system

(differences from chip to chip, predominantly due to limits in the

reproducibility of the waveguide layers). Both sensors prepared via

hydrophobic phosphate SAMs of DDPO4 and ODPO( showed a decrease of

the fluorescence signals for the 100 % serum series compared to the signalintensities of lower serum concentration series. Nevertheless, the net

sensitivity referred to the target/serum concentration is the highest for non-

diluted serum samples. That means that 100 % of patient serum can be used

for the determination of the concentrations of similar target molecules.

3.2.7.3 Comparison of PLL-PEG and Alk\ I Phosphate Sensor Interfaces

The limiting factors for the sensitivity of a sensor are the signal-to-noiseas well as the SB/NSB ratio. These factors are indirectly represented in the

dose-response curves (Figures 54 and 55, and App. 26).

Both systems (PLL-PEG and alkyl phosphate SAMs) showed excellent

tolerance to biological material such as serum. The dose-response curves are

linear down to the low picomolar to femtomolar region, independent of the

serum concentration. The inherent protein resistance and specific activityagainst streptavidin make the PLL-PEG'PEGbiotin derivatives interestingfor biosensor applications. The good solubility in water is a further behavior

that simplifies its handling. Nevertheless, the alkyl phosphate SAM-based

sensors showed significantly higher sensitivity for the 100 % serum series

(about three times higher signal in the case of DDPO4 compared to the PLL-

PEG sensor surface). Since the increased viscosih with increasing serum

concentration and therefore the kinetics for specific adsorption is the same

for both systems we believe that the BSA-biotin - streptavidin activated (and

236


BSA passivated) surfaces have a better resistance against non-specific

protein adsorption, in comparison to PLL-PEG/PFG-biotinl5% surfaces.

The detection limit for RIgG (concentration referred to newborn bovine

serum), measured with the phosphate sensor chips, was found to be in the

femtomolar region and measured \\ ith the PT 1 -PFG based sensor chip in the

region of 1 to 10 pM.

3.2.8 Mixed SAMs for Biosensor Applications

First tests were executed to inxestigate the influence of the SAM

hydrophobicity on the activity of immobilized antibodies.

Planar waveguide chips (TaiOs) were cleaned and coated by immersion

in different DDPOt(NH4yOH-DDP04(NFI4)2 mixtures as described in

chapter IV 2.1.3.2, where the water-contact-angle data are presented in

Figure 3 1 and Table 30: 0.5 mM aqueous solutions from the two different

alky I phosphate ammonium salts were prepared. The solutions were mixed

in a ratio of 8:2, 6:4, 4:6 and 2:8 (corresponding to OH-DDPO^NHf)?concentrations of 80%, 60%, 40% and 20%, referred to the total amount of

alkyl phosphate, respectively).

Clean Ta2Os chips were separately incubated in the mixed amphiphilesolutions as well as in pure 0.5 mM DDP04(NITt)2 solution (0% hydroxy-terminated phosphate) and in pure 0.5 mM OH-DDPOt(NTT4)2 solution

(100% hydroxy-terminated phosphate). The resulting SAMs presentingdifferent hydrophobicities were then used for first biosensor tests.

For the determination of the activity of immobilized antibodies, we also

used the Cy5-RlgG/(xRIgG model system. Similar to the determination of

specific and non-specific binding of target molecules (chapter TV 3.2.2), the

SAM-coated sensor chips were incubated in a BSA-biotin solution

(0.1 mg/ml in PBS) for 1 hour, rinsed with water and immersed in a BSA

solution (I mg/ml in PBS) for 10 minutes, fhe chips were again rinsed with

water, blow-dried with nitrogen and mounted in the FOB1A apparatus.

237


The test series is again presented in the form of a table (Table 39).Similar to the first measurements of PLL-PEG/PEGbiotin or hydrophobic

alkyl phosph(on)ate SAM coated chips, we used PBS buffer as a solvent for

the assay as well as for the dilutions.

Table 39: Test series to determine the performance of bio sensor chips based

on SAM platforms with different hydrophobicities.

Cycle N° Sample description Surface reaction

1 Assay buffer (PBS) preconditioning

2 Streptavidin (50 fig/ml in PBS) Activation for immobilization

of antibodies

3 Cy5-RlgG (250 pM) Non-specific binding

4 biotin-oeRIgG (12 p:g/ml in PBS) Immobilization of antibodies

5 Cy5-RIgG (250 pM) specific binding

The results of the test series to determine the activity of immobilized aRIgGantibodies an biosensors based on SAM platforms with different

hydrophobicities are presented in Figure 56.

238


100% 80% 60% 40% 20% 0%

OH-DDP04(N 114)2 concentration

Fig. 56 Cy5-RlgG/aRIgü immunoassay data from chips based on mixed

DDPCVOH-DDPOj SAM platforms with different hydropho-bicities The total amphiphile concentration of the mixed alkylphosphate solutions which were used for the self-assemblyprocesses was kept constant at 0 5 mM The ratio of OH~

DDP04(NH4)2/DDP04~(NH4)2 is indicated in % OII-DDP04 I he

highest fluorescence signal is set to 100% and the relative signalintensities are compared to each other

SAMs from the solution mixture of 60 % OH-DDPO4(NFl0> and 40%

DDP04(N1I4)2 (both from 0 5 mM stock solutions) gave the best assay

performance within this tests series fhe water-contact-angle of this mixed

SAM was 70 9 ± 19 corresponding to a surface that is neithet stronglyhydrophilic nor hydrophobic

The non-specific binding signal was very low on all the different

surfaces, suggesting stable adsorption of BSA on highly hydrophobic as well

as on less hydrophobic alk\ I phosphate SAMs (results not presented)

239


More detailed analyses of the mixed SAMs based on techniques such as

XPS, ToF-SIMS and AFM are in preparation and will be published[Tosa,20001.

3.2.9 Summary of the Immunoassay Performance

A quantitative and generali} valid comparison between the three

different surface architectures in terms of their performance as immunoassayplatforms is difficult because onl} one test, the rabbit IgG - goat anti rabbit

IgG model system, has been investigated in detail. Another problem is the

currently low reproducibility of the planar waveguide chips which show

differences in their quality, especiall} from batch to batch. Nevertheless,there are several qualitative to semi-quantitative compared criteria that can

be used to judge the sensor performance of the different approaches.

The total concentration range for a RlgG immunoassay using our

system reaches from 1 to 250 pM (or higher1) and from 10 to 250 pM (orhigher1) for hydrophobic alky I phosphate SAMs and PLL-PEG based

sensors, respectively. The upper physical detection limit of the opticaldevice (including amplifier, filters, etc.) within a dose-response test series is

about 1 nM.

Concerning immunoassay applicability, the mixed SAMs show the best

performance in terms of a system without further functionalization, i.e. usingphysical adsorption of the recognition units. Nevertheless, PLL-PEG may

represent a platform which provide a surface architecture that is chemicallymore versatile (e.g. functionalization with succinimidyl group for

subsequent covalent binding of recognition elements) or for oligonucleotidesensor applications.

PLL-PEG derivatives and the ammonium salts of DDPO4 and 011-

DDPO4 are highl} suitable for industrial sensor production, for three

reasons: the substances are not toxic; aqueous solutions (no need for organicsolvents); cost-effective spontaneous adsorption by simple dipping processnot requiring multistage, expensive and difficult surface functionalization.

1) The upper detection limit or the linearitj. limit of the dose-response curve in the higherconcentration range were not tested.

240

V Conclusions and Outlook

Conclusions and Outlook

Two different principles of adsorbents (alkyl phosph(on)ates and poly-(L)-lysine - poly(ethylene glycol) derivatives) for the formation of self-

assembled adlayers were used to prepare biospecific surfaces and to studytheir performance for biosensor applications.

Metal oxide surfaces such as Ta2t>5, T1O2 and Nb205 have been coated

with different alkyl phosphates and phosphonates. The monofunctional

(methyl-terminated) phosph(on)ates adsorb as two-dimensionally ordered

self-assembled monolayers (SAMs) and form highly hydrophobic surfaces.

SAM formation kinetics were studied with a grating coupler method

and water-contact-angle. It has been shown that self-assembly of alkyl

phosphates on TaiOs is a fast process, where 90% of the monolayer has

formed after a few minutes. Alkyl phosph(on)ates have been shown to be

stable in air for long periods (months). Slow alkyl phosph(on)ate desorptioncould be measured when SAM coated metal oxide samples were stored in

water or phosphate buffer saline solution (PBS).

Based on water-contact angle measurements, X-ray photoelectron

spectroscopy (XPS), time-of-flight secondary-ion mass spectrometry (ToF-

STMS), near-edge X-ray absorption fine-structure spectroscopy (NEXAFS)and atomic force microscopy (AFM) we have developed a molecular model

describing orientation and order of the SAMs. The driving force for the

process is believed to be the strong complex binding of the polarphosph(on)ate head group to the metal cations. Strong evidence for this

conclusion is the fact that SAMs do not form if an excess of inorganic

phosphate is present in the amphiphile solution.

Octadecyl phosphate (ODP04), self-assembled on TajOs has been

studied most extensively. A binding model has been proposed, where the

phosphate (ligand) forms monodentate and bidentate complexes with the

tantalum cation. The model has been applied on the different

amphiphile/metal oxide combinations of monofunctional amphiphiles

241


(dodecyl phosphonate and biphenyl phosphate) or bifunctional amphiphiles(hydroxy dodecyl phosphate and N-ethylamino dodecyl phosphonate) self-

assembled on Ta20s or TiCb. The ratio of two different oxygen bindingtypes (0(1) : P-0 M and P=0; 0(2) : PO-R and PO-H) that were detected

by XPS analysis support the validity of the monodentate - bidentate model.

Ammonium salts of a monofunctional (DDPO4) and a bifunctional

(OII-DDPO4) phosphate were prepared by simple precipitation of the

products. Formation of the ammonium salts improves water solubility of the

alkyl phosphates.

The observation that DDPO4 selectively self-assembles (from aqueoussolution of its ammonium salt) onto metal oxide surfaces (such as TJO2,Ta20s, Fe203, ND2O5, ..) but not onto silicon oxide it will be possible to

produce hydrophobic/hydrophilic patterns on correspondingly preparedmetal oxide/silicon oxide surfaces (prepared by e.g. lithographic methods).Within the framework of the thesis it has been shown that it is also possibleto produce more hydrophilic SAMs using aqueous solutions of the

ammonium salt of hydroxy-terminated dodecyl phosphates (OH-DDP04(NH4)2). Furthermore, the hydrophobicity imposed onto a metal

oxide surface can be varied in a controlled, predictable way by the formation

of mixed methyl-/hydroxy-terminated dodecyl phosphate SAMs. Such

mixed SAMs have been shown to serve as sensor platforms for

immunoassay applications with a maximum recognition element activity for

SAM-coated samples that were produced from a solution with a

DDP04(NH4)2/OH-DDP04(NH4)2 ratio of 0.66.

Since hydrophobic SAMs from alkyl phosphates provide reproducible,homogeneous surfaces that can be easily functionalized by, for example,biotin-conjugated albumin, the system becomes an ideal substrate for

biosensor miniaturization.

For the investigation of these hydrophobic surfaces, a novel method

named microdroplet density (|add) has been introduced: Samples are cooled

down in air of high humidity and water vapor condenses at the surface in the

form of droplets. The distribution of growing droplets has been used to

estimate the homogeneity of the SAMs (qualitatively) on the jim scale.

Quantification (droplets per unit area) leads to a value (microdroplet density)that allows for the quantitative estimation of the surface quality. The method


is believed to be useful in a large field of application in materials science

and industry.

Polyethylene glycol)-grafted poly-(L)-lysine (PLL-g-PEG), with a PLL

backbone that is inherently positively charged at neutral pH, spontaneouslyadsorbs onto negatively charged surfaces, such as metal oxides, havingisoelectric points (IP) typically lower than 5, by electrostatic interaction,

PEG-grafted copolymers are not limited to molecules based on poly-lysinebackbones. Negatively charged backbone polymers such as poly-carboxylates would provide copolymers that could be selectively adsorbed

onto positively charged surfaces (e.g. AI2O3 or Fe203) at neutral pH.Patterned mixed adlayers of positively and negatively charged PEG-

copolymers could be realized on correspondingly prepared mixed metal

oxide surfaces.

Biotin-functionalized PLL-g-PEG copolymers have been successfullyused for the immobilization of recognition elements either by specificbinding of streptavidin-conjugated capture molecules or by the introduction

of an intermediate streptavidin adlayer followed by specific binding of

biotinylated capture molecules. Synthesis of other functionalized copoly¬mers such as vinylsulfone-terminated, PEG-grafted backbone polymers, are

potential candidates for further derivatization via covalent binding. Such

polymers are believed to provide protein-resistant adlayers for biosensor

applications and implant material modification for controlled binding of

biomolecules such as recognition elements and specifically biocompatibleunits (e.g. for selective cell growth), respectively.

The functionalization of alkyl phosphates with oligo(ethylene glycol) or

PEG at the ©-position is believed to have potential for producing protein-resistant surfaces based on the phosphate chemistry. The Van der Waals

interaction of the alkyl chains, responsible for two-dimensional ordered

SAMs, combined with the protein resistance of the PEG moiety and the

specific chemistry of terminal functions (e.g. biotin terminated PEG) is

believed to open a large field of applications due to the specific properties of

each components of such novel substances.

243


The detection limit of the used model system immunoassay (rabbitIgG/anti rabbit IgG) is in the temtomolar region, which is comparable to that

of well-established methods such as enzyme-linked immuno sorbent assay

(ELISA) or radio immunoassay (RIA). Further optimization of the system

(e.g. miniaturization) is expected to lower the detection limit to the attomolar

(1CT18M) or zeptomolar (10"21 M) region with a significant reduction of time-

to-result of 1 - 2 orders of magnitude.

The potential uses of biosensors cover a large spectrum, including:immunoassays, toxicology analysis, forensics, drug screening, genomics and

proteomics, agrodiagnostics, diagnostics and therametrics or patient

stratification. The worldwide financial impact of biochips has been

calculated to be about $ 40 million for the year 1998 and is estimated to

exceed the $ 500 million-per-year limit by the year 2005 [BioW,1999].

244

VI Literature

VI Literature

I Abel, 1995] A.P.Abel Faseroptischer hvancszenzfeld-Biosensor zum SequenzspezifischenNachweis von Oligonucleotide)!. Doktorarbeit, Basel, 1995.

[Ahlu, 19911 A.Aliluwalia, D.DcRossi. C.Ristori. A.Schirone, G.Serra Biosensors de

Bioelectronics 1991, 7 207.

[Ai/e,1998] J.Aizenberg. A.J.Black, G.M.Whitesides Nature 1998, 394, 868.

[AlarJ 990] J.P.Alarie. M.J.Sepaniak. T.Yo-Dinh. inal Chim. Acta 1990, 229, 169.

[Alst, 1999] J.G.Van Alston Langmuir. 1999, 75. 7605.

[Amer.2000] Amersham Life Science FluoreLink C> 5 Monofunctional Dye 5-Pack

Cat.No. PA25001. Amersham Rahn, Switzerland, 2000.

[Anza, 1993] J-I.Anzai, r.Hoshi. S.l ee. T.Osa Sensors and Actuators B 1993,13-14. 73.

1 Aron. 1997] Y.G.Aronoft B.Chen. G.I u. C.Seto. J.Schw art/. S.L.Bernasek J Am Chem

Soc. 1997, 119, 259.

[Bain, 1989] C.D.Bain, E.B.Throughton. Y.-T.Tao. J.F\ all. G.M.Whitesidcs. R.G.Nuzzo J.

Am. Chem. Soc. 1989, 111. 321.

[Ball. 1995] B.Ballou, G.W.Fisher. A.S.Waggoner. D.l .Parkas, J.M.Rciland, R.Jaffe.

R.B.Mujumdar. S.R.Mujumdar, T.R.Pfakala Cancer Immunology,

Immunotherapy 1995, 41, 257.

[Bart.2000] W.Barthlott, C.Neinhuis Planta 1997, 202, 1.

[BioW, 1999] Bio World 2/99 (Front Line, Foster Cit\, CA)

[Bier, 1991J K.Bier Ohcrflachenreinigung von Lithiumniohat und Ilerstellwig von

Aminosilanfilmen durch Gasphasenahscheidung Diplomarbeit, Heidelberg,1991.

[Bish,1996] A.R.Bishop, R.G.Yuz/o Curr Opin Colloid Interface Sei 1996,/, 127.

ß Ian. 1996] A.P.Blanchard. R.J.Kaiser. L.L.Ilood Biosensors & Bioelectronics 1996, /1,

687.

[Bous, 1991] L.J.Bousse J. Colloid Interface Sei. 1991, 142 22-32.

VI Literature

[Bram.l 997] Ch.Bram. Ch.Jtmg. M.Stratmann Fres. J. Anal Chem. 1997, 358, 108.

|Bram,1998] C.Bram Oherflachenanah tische l Untersuchungen zur Selhstorganisation von

aliphatischen Phosphonsauren auf Aluminium. PhD Dissertation. Erlangen,1998.

[Brov,1999] D.Brovelli, G.llähner, L.Rui/, R.Hofer. G. Kraus, A.Waldner, J.Schlösser,

P.Oroszlan, M.Ehrat N.D.Spencer Langmuir 1999,15, 4324.

[Busc, 1989] G.Busca, G.Rarais, V.Loren/elli. P.P.Rossi, A.L.Ginestra, P.Patrono

Langmuir 1989, 5, 911.

[Chid, 1991] C.R.D.Chidsey Science 1991, 251, 919.

[Cott.l988| F.A.Cotton, G.Wilkinson Advanced Inorganic Chemistry, 5th ed.; Wiley: "New

York, 1988.

fDann,1998] O.Dannenbcrger, J.J.Wolff. M.Buek Langmuir 1998,14, 4679.

|Desh, 1993] S.S.Deshpandc, B.P.Sharma: In P.Singh, B.P.Sharma, P.Tyle: Diagnostics in

the year 2000: VanNostratid Rcinhold. New York. 1993, 459.

[Desh, 1994] S.S.Deshpandc. R.M.Rocco Food technology 1994, June, 146.

[Duve, 1995] G.L.Duveneck. P.Oroszlan. A.P.Abel. B.Klee, V.Steiner, M.Ehrat. D.Gygax,Tl.M.Widmet* Proceedings of Medical and Fiber Optic Sensors and Delivery

Systems, SP1E. Vol. 2631, Barcelona. 1995, 14.

|Duve, 19961 G.L.Duveneck, M.Pawlak. D.Ncusehäler, W.Budach, M.Ehrat Proceedingsof Biomedical Systems und Technologies. SP1E, Vol. 2928, Vienna, 1996, 98.

|Elbe,1996] D.L.Elbert, J.A.llubbellJm?» Rev Mater Sei 1996,20", 365.

[Elbe, 19981 D.L.Elbert. J.A.I Iubbell J Biomed Mater Res 1998,72,55.

[Fcut,1993] P.Peuter. P.Eisenberger. K.S.l iang Fhys. Rev. Lett. 1993, "0. 2447.

[Fisc,1997| D.Fischer. A.Marti, G.Hähner,/; Vac. Sei. TeclmoT A 1997, 15, 2173.

[Folk,1995] J.P.Folkers, Ch.B.Gorman, P.P. Laibinis, S.Buchholz, G.M.Whitesides.

R.G.NU770 Langmuir 1995, //. 813.

[Fris,1994] C.D.Frisbie. LP.Ro/sm ai. A.No>, M.S.Wrightom C.M.Lieber Science 1994,

265, 2071.

LGabr.1999] CLGabriel J J'ac Sei Teelmol A 1999, /" 1494.

246

VI Literature

[Gao.1996] W.Gao, L.Dickinson. Ch.Grozinger. E.G.Morin, L.Rcven Lcmgmuir 1996, 12,

6429.

[Gao.1997] W.Gao. L.Dickinson. Ch.Gro/inger. F.G.Moria, L.Rcven Lcmgmuir 1997,13,

115.

LGate, 19641 B.M.Gatehouse, A.D.Wadsle\ Acta Cryst 1964, 17, 1545.

| Glas. 19961 J.E.Glass Hydrophilic Polymers, Performance with Environmental

Acceptance; Am. Chcm. Soc. Washington, DC 1996.

[Grcs. 19791 R.Gresch, W.Müller-Warmuth, H.Dutz J. Non-Cryst. Solids 1979, 34, 127.

[Hähn.19911 G.IIähner, M.Kinzler. Ch.Wöll. M.Grunze, M.K.Scheller, L.S.Cederbaum

Phys. Rev Lett, 1991, d", 851.

[Hatp,1992| A.Halperin, M.Tirrell. LP.Lodge Advances in Polymer Science 1992,100,

31.

[Hand, 19971 Handbook of Chemistry and Physics; CRC Press: Boca Raton, FL, 1997.

[Hard, 19981 P.I larder. M.Grunze, R.Dahint, G.M.Whitesides, P.R.Laibinis J. Phvs C hem

£1998, / 02. 426.

[Haze,19311 F.Ilazel, G.ILAyres J. Pin s Chem 1931, 35, 2930.

[Ileng,1987| A.Henglein Ultrasonics 1987, 25, 6.

[Hick. 19911 J.JJIickman, D.Ofer, P.E.Laibinis. G.M.Whitesides, M.S.Wrighton Science

1991, 252, 688.

[Higg,1987| I.L.lIiggins. C.R.LowcPhil frans R. Soc Land. 1987, B316, 3.

rilofe,2000a] R.Hofer, M.Textor. N.D.Spencer Langmuir 2000, (submitted).

[IIofe,2000bl R.Hoier. M.Textor, G.IIähner. N.D.Spencer Langmuir. 2000, (in prep.).

[Huan, 19961 S.-C.(P.)1 Luang. K.D.Caldwell. J.-N.Lin, H.-K.Wang, J.N.Hcrron Langmuir

1996, 12. 4292.

[Hubb,20001 J.A.Ilubbell. M.A.Textor Multifunctional Caiionic and Anionic Polymeric

Coatings for Surfaces in Analytic and Sensor Devices. Patent 2000,

(submitted).

|1 lula.l9911 A.lhilanicki. S.Glab, F.Fngman Pure & Appl Chem. 1991, 63,'9, 1247.

247

VI Literature

[Jana, 1998] J.Janata, M.Josowic/, P.Vanysek, D.M.DeVancy Anal. Chem. 1998, 70,

179R.

[ Joha,19571 P.G.Johannsen. A.S.Buchanan Australian J. Chem. 1957,10, 398.

[JohnJ 991] B.Johnssou, S.Lotas, G.Lindqvist/J^/. Bioehem. 1991, 198, 268.

[Jung,1998] C.Jung AIDiphosphonsäuren als molekulare Haftvermitllerfür Aluminium

und Zinkwerkstoff. PhD Dissertation, brlangen, 1998.

[Kena,2000] G.Kenausis, J.Vörös, D.L.Libert. N.Huang, R.Hofer, L.Ruiz, M.Textor,

J.A.Hubbell. N.D.Spencer J. Phys. Chem. B 2000,104, 8298.

[King. 1999 ] P.Kingshot. HLJ.Griesser Currant Option in Solid State and Materials Science

4 1999, 403.

[Kinz,l 994) M.Kinzler, A.Schertel. G.Hähner, Ch.Wöll, M.Grunze, H.Albrecht,

G.Holzhüter, Th.Gerber./ Chem. Phys. 1994,100, 7722.

[Kric, 1993] L.J.Kricka Clinical Biochemistry 1993, 26, 325.

[Krög, 1998[ D.Kroger, A.Katerkam. R.Renneberg. K.Cammann Biosensors &

Bioactivalors 1998, 75, 1141.

[Kuma, 2000] C.V.Kumar. A.Chaudhari J. Am. Chem Soc. 2000, 122, 830.

[Kurr, 1997] R.Kurrat. M.Textor, J.J.Ramsten. P.Böni. N.D.Spencer Rev. Sei. Insirum.

1997,68, 2172,

[Kurr,1998] R.Kurrat Ph.D. Adsorption of Biomoleeules on Titanium Oxide Layers in

Biological Model Solutions PhD Dissertation. RTH Zürich, Switzerland,

1998.

[Laib,1989| P.KLaibinis. J.J.llickman. M.S.Wrighton, G.M.Whitesides Science 1989,

245, 845.

[Lec,2000] K.P.Lee. Il.Cho, R.K.Singh. S.J.Pearton. C.Ilobbs. P.Tobin J. Vac. Sei.

Teehnol. B 2000, 18. 293.

[Lin,1991] J.-N.Lin. I.-N.Chang. J.D.Andrade. J.N.Herron. D.A.ChristensenJ. of

ChromaiogrA99h542. 41.

[Löpe.1993] G.P.LÖpez, H.A.Biebuyek. C.D.Friesbi. G.MAVhitesides Science 1993, 260,

647.

[Löse! 9981 F.Löscher. T.Ruckstuhl. S.Seeger Adv. Mater 1998, 10. 1005.

248

VI Literature

[Lu. 19961 B.Lu, M.R.Smith, R.C/Kermedy Analyst 1996,121, 29R.

[Maeg,1997] l.Maegc, E.Jaehne, A.Henke, H.-J.P.Adler. C.Bram, C.Jung, M.Stratmann

Macromol. Symp. 1997, 126. 7.

[Macg, 1998] T.Maege, E.Jaehne, Ai lenke. H.-J.P. Adler. C.Bram, C.Jung, M.Stratmann

Progress in Organic Coatings. 1998, 34, 1.

[Maoz, 1984] R.Maoz, J.Sagiv Journal of Colloid and Interface Science 1984, 100, 465.

[Maoz. 1988] R.Maoz, L.Net/er. J.Gun, J.Sasgiv J Chim. Phys. 1988, 85, 1059.

[Matt, 1930] S.Mattson Soil Sei. 1930, 30. 459.

[Matt, 1934] S.Mattson, A.J.Pugh Soil Sei. 1934, 38. 229.

|McNe,19931 C.McNeill Biosem, Bioeleclron. 1993, S"

i-v.

[Méri, 1937] R.Mérigaux Rev. Opt. 1937, 9, 281.

[Mitt, 1993] K.L.Mittal, Contact angle, wettability and adhesion, VSP, 1993.

[Moul, 1992] J.F.Moulder. W.F.Stickle. P.K.Sobol, K.D.Bomben Handbook ofX-ray

photoe/eclron spectroscopy; Perkin-Flmer CorporalioiEPHI Division: Eden

Prairie, MN, 1992.

[Naka,1996] K.NakanisM, H.Muguruma. l.Karube Anal. Chem. 1996, 68, 1695.

[Nath,1997] N.Nath. S.R.Jain. S.Anand Biosensors & Bioactivaiors 1997, 12, 491.

[Nuzz,1983 ] R.G.Nuzzo, D.L.Allara J. Am. Chem. Soc. 1983, 105, 4481.

[Nyqu.2000] RM.Nyquist. A.S.Eberhardt. L.A.Silks 111, Z.Li. X.Yang, B.I.Swanson

Langmuir 2000, 16. 1793.

| Okam. 1985] Y.Okamoto Bull. Chem. Soc. Jpn. 1985, 58, 3393.

[Outk, 1988] D.A.Outka. J.P.Stöhr, J.D.Schwalen J. Chem. Phys. 1988, 88, 4076.

[Park, 1965 ] G.A.Parks Chemical Reviews 1965, 65, 177.

[Pate.1997] N.Patek M.C.Da\ies. M.Hartshorne. R.J.Heaton. C.J.Roberts, S.J.B.Tendler,

P.M.Williams Langmuir 1997,13, 6485.

[Poir, 1996] G.E.Poirier. E.D.P\ lant Science 1996, 272, 1145.

[Port.1987] M.D.Porter, L.B.Bright, D.L.Allara. C.E.D.Chidsey J. Am. Chem. Soc. 1987,109, 3559.

249

VI Literature

[Prim. 19931 K.L.Primc, G.M.Whitesides J. Am. Chem. Soc. 1993, 7/5, 10714.

[Rams, 1993J J.J.Ramsten.7. Stat. Phys. 1993, ~J, 853.

[Rams, 1995] J.J.Ramsten. D.J. Roush. D.S.Gill R.Kurrat. R.C.Willson J. Am. Chem. Soc.

1995, 7 7" 8511.

[Rayl, 19ll| Lord Rayleigh Nature 1911,66,416.

[Roge,1992] K.R.Rogers, J.N.Lin Biosens. Bioelcctron. 1992, 7, 317-321.

[Roth.1965] R.S.Roth, A.D.Wadsley Ada ( 'ryst. 1965, 18, 724.

[Saka,1998] N.Sakai. R.Wang, A.Fujishima. T.Watanabe. K.ïlashimoto Langmuir 1998,

14, 5918.

[Sala, 19951 M.T.Salama, T.Shido, H.Minagawa. M.lchikawa J. Catal 1995, 152, 322.

[Sehe, 19871 F.W. Serieller Stud. Biophys. 1987,119, 221.

LScof,1976] J.H.Scofield J. Electron Spectrosc. Re/at. Phenom. 1976, 8, 129.

LScah, 19791 M.P.Seah. W.A.Dench Surf. Interface ,inal. 1979, 1, I.

[Sell.19931 TT.Sellers, A.Ulman. Y.Sehnidman, J.E.Eilers J. Am. Chem. Soc. 1993,115,

9389.

LShri, 1997J L.C.Shriver, B.Donner, R.Fdelstein, K.Breslin, S.K.Bhatia, F.S.Liglcr

Biosensors & Bioactivators 1997, 12, 1101.

[Shum,2000] I.S.Shumaker-Parry. Ch. LCampbeil. G.D.Stormo. F.S.Silbaq, R.Il.Aebersold

Proc. of Medical and Fiber Optic Seniors and Delivery Systems, SPIE, San

José, CA, submitted, 2000.

[ Sigg.2000 J K. Sigg Micro Droplet Den s it\ } lessungen aufhydrophilen und hydrophoben

Oberflächen mit und ohne Strukturierung Semesterarbeit I, Lab. for Surface

Science and Technology. Dept. of Mat., ETH Zürich, Switzerland, 2000.

[Situ, 1998] M.Situmorang, J.J.Gooding, D.B.llibbert. D.Barnett Biosensors &

Bioactivaton \998,13, 953.

[Soft. 19981 S.J.Sofia. V.Premnath, E.W.Merrill Maeromolecules 1998, 31. 5059.

[Spie, 1998] U.E.Spichiger-Keller Chemical Sensors and Biosensors for Medical and

Biomedical Applications: Wiley-VCH. Weinheim, 1998.

[Spin,l 993J J.Spinke. M.Liley, E.-J.Sehmitt J.Chem.Phys. 1993, 99, 7012.

250

VI Literature

I StepJ 9711 N.C.Stephenson. R.S.Roth Acta Crys. 1971, B27, 1037.

[Stöh. 1992] J.Stöhr XEXAFS Spectroscopy; Springer: I leidelberg, 1992.

[Swal, 1987] J.D.Swalan, D.L.Allara. J.D.Andrade, EA.Chandross, S.Garoff,

J.Israelachvili, T.J.McCarthy. R.Murray, R.F.Pease, J.F.Rabolt, K.J.Wanne,

H.Yu Lungmuir 1987, 3, 932.

|"Text,2000"| M.Textor. L.Ruiz, R.Hofer, A.Rossi. K.Feldman, G.Häliner, N.D.Spencer

Langmuir 2000, 16, 3257.

[Thie, 1997] A.J.Thiel. A.G.Frutos. C.E.Jordan. R.M.Corn, L.M.Smith Anal. ( "hem. 1997,

69. 4948.

[Tosa.2000] S.G.P.Tosatti, M.Textor. N.D.Spencer, in prep.

[Ulma. 1990] A.Ulman Adv. Mater. 1990, 2, 573.

[Ulma,1991] A.Ulman/f/7 introduction to organic thin films: from Langmuir-Blodgett to

SelfAssembly Academic Press: Sanüiego.CA, 1991; p 279.

[Ulma.1996] A.Ulman Chem. Rev. 1996, 96. 1533.

[VanG. 1991] J.VanGent, P.V.Lamheck. R.J.Bakker, Th.J.A.Popma, L.J.R.Sudhölter,

D.N.Rcinhoudl&won and Actuator* A 1991,25-27. 449.

[Verw,19411 L.J.W.Verwey Rec. Trav. Chim. 1941, 60. 625.

|Vikh,1998] LVikholm, W.M.Albers Langmuir 1998,14. 3865.

[Wang,1998] R.Wang, K.Hashimoto. A.Fujishima, M.Chikuni, E.kojima. A.Kitanuira,

M.Shimohigoshk T.Watanabe.L/v. Mater. 1998, 10. 135.

[Wcll.1984] A.F.Wells Structural Inorganic Chemistry. Clarendon Press, Oxford, 1984.

[Well, 1991] A.F.Wells Structural Inorganic Chemistry. Clarendon Press, Oxford, 1991.

[Wert,1999] C.F.Wert, M.M.Santore Langmuir 1999, 15, 8884.

[Will, 19941 R.A.Williams, H.W.Blanch Biosensors & Bioelectronics 1992, 7, 405.

|Wood,1996a] J.T.Woodward, A.Ulman. D.K.Schwartz Langmuir 1996, 12, 3626.

[Wood,1996bl J.T.Woodward, D.K.Schwartz J, Am. Chem. Soc. 1996, 118. 7861.

[Xiao,1999] S.-J.Xiao Tailored Organic Thin Films on Gold and Titanium. PhD

Dissertation, K FH Zürich, Switzerland. 1999.

| Yu. 1994] H.Yu Nucleic Acids Research 1994, 22. 3226.

VII Appendix

VII Appendix

1 ToF-SIMS

1.1 Negative ToF-SIMS Spectra of phosph(on)ate SAMs

App. 1: ToF-SIMS data (neg. ions) from ODPO4 self-assembled on TiÖ2

from a 0.5 mM ODPO4 solution in n-heptane + 0.4 % 2-propanol.

TixPO%H,-Fragmente

TiO, no, TiPO, Il PO, TiPO«, TiP207 TiP20, Ti,0„

79.9705 95.9624 142.9028 221.8105 237.809

H 80.9778 96.9695 159.8874 175.8903

H2

H,

H7 198.8605

CJI, -Fragmente OCNH,-Fragmente

C C; C, c,

12.0134 24.0163 36.0155

H 13.0213 25.0229 37.0232 49.0205

II2 14.0292 26.0273 38.0298

H, 39.0368 51.0356

O 16.0115

0; 32.0006

011 17.0180

OC?H 41.0160

ll,POv-Fragmente

PO PO; PO, PO,

46.9686 62.9732 78.9654

11 95 9624

H, 96 9695

Cil, 109.9725

C\H, 122.9760

CJI,, 180.8768

CigHui 349.1861

VII Appendix

App.2: ToF-SIMS fragments (neg. ions) of DDPO3 self-assembled on

Ta205 from a 0.5 mM DDPO3 sol. in ethyl acetate/methanol 10:1.

raP,0,C,H„-Fraementc

TaO TaO, TaO, TaO, FaPO, TaPQ TaPQ, Ta,0, Ta,PQ, Ta,Ofi Ta,PO,

1969453 212.9«" 228.934C 244.926,' 275.9112 291 924<3 441 849c 504.81«

H 197 952F 213.9515 229.9481 245.937-; 276.925*; 292.93 34 308.9391 457.8433 458.875.'

H, 198 9681 214.9721 230.968: 246 955?

CH 288.9 m

CH, 280 4335 3059458 518 849!

CH, 290 938(1 306 9654

CH 307 960C

CH, 3 18 964t

CH, 236 937"

CH, 237.9373

CH„ 238 9581

HLPO.-Fraemente OH,-Fraemente OCT 1,-Frasmente

PO PO, PO, PO,

62 964 78 9.582

H 95 954!

H, 96 963e;

CH, 93 9821

CH, 94 985!

CH, 89 9976

C,H, 91004 e 106 989^

CiHs 108 003;

CJL 86 999C

CT I, 120 007!

CÄ 121 026(

C-;H^ 122 926-

CH, 92 030e

CI I, 135 020-

CH,„ 149 036!

QH„ 1630531

G,H,V 229 1 05*

C,,H,j 213 185(1

GTk 249 1 84-

G 41„ 250 188*

C C

12 0003

H 13 007! 25.007!

H, 14015;

O 15 9933

OH 17 0011

OGT1,,, 155 149:

253

VII Appendix

App. 3: ToF-SIMS data (neg. ions) from DDPO3 self-assembled on Ti02

from a 0.5 mM DDPO3 solution in ethyl acetate/methanol 10:1.

T ixPO, 11,-Fragmentc

n TiO Ti02 TiO, TiPO TiPO, TiPOj TiP04 TiPO,

79.9700 142.8991 158.8943

11 64.9729 80.9714 96.9519 95.9563 127.9154 143.9049 159.9008

H, 97.96 160.9064

IT,

H7

CII, 140.9111 156.9033

GI3 141.9100 157.9079 173.9139

C2II2 136.9353

CJI4 122.9340 138.9292 186.9157

C2H, 139 9414

C2H7 189.8644

C,H4 198.9049

C4H7 102.9580

TiP06 TiP206 TiP20- TiP20, n2p:o, Ti204 TiA

205.8589 221.8524 236.8491

II 175.9012 206.8658 222.8524 237.8420 238.8371

H2 176.9004

H, 196.8948

H7 198.9049

CII,

CTT

QII2 185.9010

C2H4

C2TT,

QH,

CJI4

C,ll7

254

VII Appendix

QH, - Fragmente OCxH, -Fragmente

C Q c, c4 c,

11.9999 24.0009

H 13.008 25.0079 49.0109

H2 14.0154 38.0159 50.0182

H3 51.0270

H. 64.0359

II. 65.0426

O 15.9957

o2 31.9865

OH 17.0025

OC2H 41.0028

OC4H7 71.0160

OC4ri9 73.0117

OQII,, 87.0038

I1VP(X-Fragmente

PO PO, PO,

46.9686 62.9666 78.9606

H 47.9807 63.9742 79.9700

H3 48.9881

CH 59.9712

CH2 60.9814

CH, 62.0187 93.9844

CH4 94.9858

C2H3 89.9896

QII4 90.9964 106.9912

C2H, 107.9989

C,H5 119.9994

QH8 135.0202

C5H10 149.0342

C6H12 163.0476

C„HM 231.1467 247.1480

C12H26 233.1591 249.1631

O52H27 250.1662

255

VII Appendix

App. 4; ToF-SIMS data (neg. ions) from OH-DDPO4 self-asserabled on

Ta2Ü5 from a 0.5 mM OH-DDPO4 solution in ethyl acetate/ethanol

10:1.

TaJ\07C0HL-Fragmente

Ta TaO TaO, TaO, TaO, TaPO TaPO, TaP04 TaPO,

180.9454 196.9493 212.9402 228.928 244.9193 275.8834 291.8780

H 197.9573 213.9501 229.9390 245.9275 276.9003 292.8899

H2 198.9628 214.9583 230.9502 246.9457

CH 272.9132 288.8795

CH2

C2H,

C2H, 254.9414

C2H5 256.9426

C,H,

C12H22

TaPO„ TaPA TaP207 TaPA Ta:0, Ta2P04 Ta206 Ta2P07

307.8888 354.8752 370.8792 441.8700 457.8660 504.8408

H 308.9016 458.8733

H2

CH

CFI2 368.8794

C2H2 348.8761

C,H,

C2H,

C,H4 820.8773

^12^ '22 670.0019

256

VII Appendix

CXH%-Fragmente

C C; c3 c,

12.0008 24.0008

H, 13.0085 25.0088 49.0112

H2 14.0166 26.0052 38.0169 50 0091

II, 51.0264

HxPOy-Fragmente OxQ H,-Fragmente

o 15.9952

o2 31.9894

H02 32.9974

OH 17.003

C2FIO 41.0050

C2H20 42.0023

C2H30 43.0217

CH02 45.0017

CH202 45.9963

C,H30 55.0329

C2H20, 58.0083

C2H302 59.0165

C7H,oO 110.0732

CuHhO 163.1057

P PO po2 PO, P04

30.9750 46.9686 62.9641 78 9582 94.9520

H 47.9738 63.9671 79.9621 95.9566

H2 64.9697 80.9655 96.9637

CH 59.9672 75.9595

CH2 60.9748 76,9680

CH3 61.9884 77.9638

72.9998

C2[I.» 74.9887 122.9844

CJI, 123.9929

C3H, 71.0148

C3H6 137.0002

C5H 139.9726

C5H4 142.9878 158.9687

C5H, 143.9934

C,H7 146.0103

C5H14 137.0718

C6H4 138.9890

CnHis 229 0881

257

VII Appendix

App. 5: ToF-SIMS data (neg. ions) from OH-DDPO4 self-assembled on

Ti02 from a 0.5 mM OH-DDPO4 sol. in ethyl acetate/ethanol 10:1.

ld„l>vO,C U Ir.iEmcnlt

li liO hO no, MO, I iPO IlPO liPO, IiPO, 3 lPO, llPOf li O,

47 975 5 1 1 1 9288 1T» 9089 112 9021 158 8973 174 8965

II 1I29J57 127 9150 143 9071 159 9018 175 8988

11 81 9560 97 9523 1 13 9409 144 «02 "• 160 9038 176 9005

II, 98 9590 1 15 9079 178 9002

til 1399187

CI I 1 56 9048

CII, 125 9247 15-9064 173 9031 1899013

c 119 9357

CII 104 9484 12» 9418 H5 9261 1510139 167 9023

c 11 101 9t4b 136 9^8 l-i; 91 12 168 9025 1 8 1 9065

111, PI 9307 I1]Wb 185 9104

C 11, 138 92^j 1869195

CII 179 9008

CII 1 6 f 9109 180 9083

CMI4 131 95S-,

C,H, 264 9389

li Of ll O, 111' O, IlPO III» O, IlP O I iP O, 1 1 PÜ 11 PO, ll PO, 11 PO, ll PO,„

189 8700 205 8749 '21 S "'62 2i7r,ii

11 "1 8610 206 880j 22' 8"M '38 8-14

II HO 8697

II, 240 86 12

II, 196 9016

11, 197 7844

CII 202 8869 218 8841 2,4 875^

CII. 203 8762 210 8815 23 5 8600

CII, 204 8802 220 875T 2 56 11,00 301 8421

CII, 25 t 884',

en« 255 8808

(11, 2 56 9029

en 198 89-8 214 8802 230 869 1

C II, 199 8845

CII, 200 8957 216 8012

CII, 201 9048 217 8911

C2II< 316 8687

CJI, 300 8551 ,17 8707

CI1„ 2 84 9436

CII„ 280 9490

CII,, 208 0619

CII, 28.5 9287

Cll, ,18 8666

C, II 364 9181

CII,, 3 80 0464

C, II,, ,60 048 8

C, 11, ,~s 9572

C, II, 308 9600

c 11115 0178 461 0105

258

VII Appendix

OCJI,-Fragmente 1 1\P0\ -Fragmente

P PO PO, PO, PO„

30.97X2 46.9720 62.9630 78.9598 94.9568

H 63.9635 79.9613 95.9605

H: 64.9682 80.9665 96.9657

C 58.9691 106.9455

CH 59.9664 75.9639

CLL 60.9742 76.9713

CH, 77.9703 93.9835 109.9778

Cil, 110.9800

C; 102.9587 118.9506

C2FL 74.9915 106.9862 122.9831

C;IL 123.9925

CH, 136.9992

C,H, 148.9971

CIL 151.0163

CIL 158.9863

CR 163.0146

CIL 165.0275

O o2

15.9958 31.9913

11 17.0034 32.9991

IL 34.0053

C 43.9908

CH 29.0030 44.9989

CH, 31.0187 45.9966

C 39.9967

Cl! 41.0045

CIL 42.0000 58.0031

C2FT, 43.0198 59.0119

CIL 45.0393

C, 51.9923

CH 52.9998 68.9974

Cil, 55.0168 71.0139

C,H 65.0022

CIL 66.0001

C„H, 67.0182

CH, 101.0630

CnIIm-Fragmente

C C C, C C C

12.0007 24.0012 36.0019

H, 13.0087 25.0086 37.0095 49.0091 73.0096

IL 14.0159 26.0035 38.0167 50.0069 62.0144

H, 15.0233 27.0243 39.0234 51.0209

IL 41.0370

259

VII Appendix

App. 6: ToF-SIMS data (neg. ions) from NEt-DDPCb self-assembled on

Ta205 from a 0.5 mM NEt-DDP03 sol. in ethyl acetate/water 1:1.

FaPOCH-Fragmente

TaO TaO, TaO, TaO, 1 aPO IdPO, TaP05 TaPOs Ta,0« Ta;0,

] 96 9175 212 9289 228 9281 244 9215 275 898S 291 9023 307 8916 441 S487 457 8647

II 197 9544 213 9538 229 9422 245 9344 276 9096 292 9060 308 8980 458 8865

II, 198 960] 214 9684 230 9598 246 9494 309 9064

CH 272 9281 288 8961

on. 2899018

CH, 290 9011 306 9184

c. 236 931 5

CJI 237 9318

CH, 238 9391 254 9574 270 9500

C,ll, 239 9526 255 9617 271 9427 118 9487

CJI, 256 9716

C,H, 348 8916

C„Hm-rragmente C„N„-I ragmcntc

c C3 C, C,

12 0003 23 9996

iL 13 0077 25 0085 37 0077 49 0079

H2 140164 380151 50 0089

N NO NO,

45 9950

H 15 0103

C 26 0051 42 0013

CJIf 80 0538 96 0337

OCJF,-Fragmente

HxPOy-I'iagmente

o 15 9932

OH 17 0011

o2 31 9862

0,11 32 9838

CO 39 9963

C,OH 41 0054

C;OH, 43 0217

cry i 45 0004

C;0;H; 58 0101

C:0;H, 59 0195

C,0,H, 71 0192

PO PO, PO, PO, PO, PO

46 9637 62 9665 78 963

11 47 9637 63 9646 79 9622 95 9M6

H, SO 9689 96 9644

en 59 9679 75 9631

au 60 9773 76 9723

CH, 61 9914 919852

C,H, 136 9584 152 9481

CJI, 107 0042 122 995

CJI.c, 121 0398

260

VII Appendix

App. 7: ToF-SIMS data (neg. ions) from NEt-DDPO, self-assembled on

Ti02 from a 0.5 mM NEt-DDPO-, sol. in ethyl acetate/water 1:1.

TaPOCH-rtagmente

11 IiO TiO, IiO, IiO, IiON 1lO,N liPO 1 iPO

47 9683 1 1 1 9349

H 112 93-6

H 65 9560 81 951 1 97 9 »62 113 9 »29

H, 98 9510 114 9481

C 75 9598 105 9383

err 76 9690

Cil 77 9682

c n 104 9438 120 9390 118 9411 135 915]

en 136 9339

C 11, 121 9339 137 9383

CH, 123 9531

CII 124 9758

CH 121 0129 135 0150

TiPO, 1 iPO., TiPO, liPO TiPO V 1 iPO,N 1 iP06N TiPjOj liP C\

1 12 9015 158 8979 174 906-, 140 9071

H 1279106 143 9035 1 59 9024 175 9053 141 9086

H 144 9012 160 9062 1 76 9070

Hi 161 9124 177 9138 256 8372

C 1 5 1 9071

Cll 155 9134

Cll, 172 9212 1 18 9335 202 8999 219 8930

CH, 173 9264 139 9363 204 8867 220 8807

Cll 167 8976 1 83 9098 214 8870 230 8716

C 11 152 9135 168 9075 1849157 200 90 -. 2 215 8930 231 8835

C 11, 1 53 9290 169 9100 (85 9165 201 9070 216 8945 232 8970

C 11, 1 70 9 1 42 186 9308

C H, 171 9171 2 34 8992

C,H 179 9069

Cll, 1809162 196 9249

c,ir, 181 9221 2 14 8898

Cll, 19891 00

CH, 199c> 190

liP o7 "liP Oi 1 iP 0,N Tl PO,i ll P O 1 l,PO, Ti.PO, Tl,PO,o

221 8736 237 8641 301 801 t 302 7768 318 7781

H 222 8669 238 8613 303 777 1

II, 239 8603 205 8818 104 81 1 320 7836 316 7806

CH2 217 8923 316 7888

Cll, 218 8947

CH5 254 8821

Cll, 264 8792

261

VII Appendix

CnHnrrragmcntc OCJi-Fragmente

C c. C c C L C C O 15 9950

12 0004 24 0000 35 9994 47 9981 OH 17 0028

H, 13 0081 25 0077 37 0073 49 0060 61 0052 710010 02 319899

TI2 140156 38 0142 62 0129 S6 0264 98 0182 0;H 32 9966

H, 27 0224 51 0227 75 0223 CO 39 9942

H, 100 0330 COH 41 0018

H, 65 0160 77 0357 1010363 COH, 43 0166

CO,H 44 9961

C„N,„(\-riagmente

CO 5 t 9938

COH 53 0009

C3OII, 55 0168

CNO CNO CNO CNO CNO CNO CNO; CjO;H, 56 9961

41 9972 65 9957 COH, 57 0315

H, 44 OHO 68 0107 72 0070 C202H; 58 0036

II, 105(1224 C02H, 59 0120

11. 58 0286 94 0280 COH 65 0019

M„ 72 0430 120 0387 COH, 67 0158

C,02H 68 9965

C„Nm-Fragmente

COH, 69 0318

CO;H, 71 0117

COJH, 83 0123

NC NC; NC, NC NC NC C02H, 85 0290

26 0026 50 0021

11 27 0026 39 0111 630121

H, 400181 52 0172 64 0147

II, 113 0312 125 0268

H, 66 0129

262

VII Appendix

IKPOy-Fragmente

P PO PO, PO- PO, PO, PN PON P02N PO,N PO^N

46 9688 62 9615 78 9559 94 9537

Tl 639607 79 9550 95 9482 109 9678

H2 64 9665 80 9620 96 9564 110 9733

C 58.9693

CH 59 9657

cn2 60 9743 122 9725

CIL 93 9777

C; 102 9564 100 9663

QU, 74 0010

C,ll, 106 9847

C2II7 110 0203

c2ns 1110131

C-,11, 87 0056 1029915

CIL 107 9889 122 0067

C.H, 99 0094

c,n„ 133 0060

c,u8 149 0255

C311, 112 0057

C,H,„ 165 0306

C,H,2 167 0432

QH« 109 0150

QU,, 163 0506 179 04S6

C6H„ 1810691

c,ti,6 207 078i

C8His 193 1094 209 1009

GH2„ 223 1194

C„,H22 221 1252

c„n2, 251 1491

C,2H;j 247 1547

c,2n26 249 1765 263 1792

CI2H27 250 1720

c12n28 265 1893

CnH|, 219 127?

C„H2„ 237 1345

CnH„, 279 2064

C„H22 235 1546

Cjiliio 291 1969 307 2609

263

VII Appendix

App. 8: ToF-SIMS data (neg. ions) from BÏPH self-assembled on Ta2Os

from a 0.25 raM BIPH solution in ethyl acetate/methanol/acetic

acid 10:l:traces.

TaJ\0,CoH„-Fragmente

TaO, TaO, TaO, TaPO, TaPO, TaPO, TaP06 TaP,0,

2)2.9367 228.9269 244.9229 275.8949 291.8926 307.8946

H 213.9491 229.9393 245.9337 276.9006 292.9022 308.8995

H; 214.9469 230.9537 246.9476

H,

Cil 272.9129 288.8950

CH, 289.9033

eu, 306.9176

cyu 237.9353

C2H2 238.9361 254.9379 270.9508 348.8492

C2H,

C,IIf,

C4H2 262.9394

C,H7 399.0088

C,2H9 461.0121

TaP20„ TaP,0- TaP208 Ta20, Ta20, Fa20, Ta206

354.8301 370.8536 441.8649 457.8842

II 410.8453 426.8564 458.8991

H2

H, 444.9794

Cil

CHj, 352.8643 368.8548

CH,

cjr

C2H2 364.8792 380.9603

C2H, 381.9668

C,H0 396.9723

C4H2

C,H,

C,2H,

264

VIT Appendix

HxPOy-Fragmente OxCvH,-Fragmente

PO PO, PO, PO, P:Os (?) P20( (?)

62.9641 78.9582

II 63.9671 79.9621 95.9509 138.9247

H2 64.9697 80.9649 96 9637

CH 75.9595

CH2 76.9680

C2H 118.9400

C'2Hj 72.9998

C2H4 122.9733

C,H() 117.0248

O 15.9938

OH 17.0016

02 31.9878

C20 39.9954

C2HO 41.0054

C2II20 42.0015

CH02 45.0003

CH20, 45.9959

C2Pi202 58.0083

C2H,02 59.0165

0,11,0, 71.0165

C]2IT90 169.1098

C12H,oO 170.1130

CxHv-Fragmento

C c2 C, c, c,,

12.0003 24.0000 35.9998 47.9994

FI, 13.0076 25.0081 37.0084 49 0086

FI2 14.0158 26.0045 38.0167 50.0091

H, 51.0264

H, 141.0560

265

VII Appendix

App. 9: ToF-SIMS data (neg. ions) from BIPH self-assembled on TiC>2

from a 0.25 mM biph solution in ethyl acetate/methanol/acetic

acid 10:l:traces.

'laJ'ÄCHu-Fragmente

Ti02 IiO, TiQ, TiP02 TiPO I iPO, TiPCX I'lPOo TiP20, 1 iP20„

126 9054 142 9013 158 8991 174 8946 205 8776

II 112 9414 1119194 127 9130 143 9061 159 9050 175 8994 206 8892

Hi 1139474 128 9489 144 9000 160 9044 176 8988

H, 114 9520 129 9464 161 9067 177 8951

eu 139 9181 I56 9070 202 8914

Cil, 140 9058 156 9096 203 8919 219 8873

Cil, 157 9059 220 8739

en, 221 8793

cyi 104 9456 135 9275 230 8706

02H2 105 9428 136 9^2 200 9065

C2lh 137 93 39 169 9058 201 9152 216 8948

C2Il, 1389307 217 8944

C2I13 218 8952

C,H2 196 9086

Cil, 134 9565

C,2H8 310 9942

^ ! 2'

' " 3119937

TiP,0 1 iP2Os IiP.O, TiAPO, Ti.O PO, Ti2P20, Ti2P 0„

237 8658

H 222 8667 238 8634

H2 223 8702 239 8639 398 7705

H, 240 8576 256 852

CH 234 8697

CH; 235 8716 300 8499

CH, 236 8673 316 835

CI14 317 843

Cil, 318 8292

c2n, 264 8695 280 8732

CnH„ 390 9808

266

VII Appendix

HxPOy-Fragmentc 0NC, H7-Fragmente

0 15.9957

on 17.0033

o2 31.9922

c2o 39.9970

C2HO 41.0049

C2FI20 42.0004

C2H,0 43.0214

CH02 44.9999

CHA 45.9954

CIIO 53.0023

C,II202 58.0065

C2II,0, 59.0138

C.1IO 65.0027

C,H;0 65.9992

C,H02 68.9984

C,H,0, 71.0140

C6H50 93.0335

C7H,0 103.0156

C7H,0 105.0279

QJI,02 108.0211

C8H,0 116.0247

C,H,0 117.0332

C,H60 118.0357

C,1I,0 119.0506

C-IIA 121.0246

Clf,RO 141.0365

C„,H,0 142.0379

C„,H,0 143.0434

C„H80 144.0458

C,:H,0 167.0488

C:H„0 168.0588

CJW 169.0688

CpHK0 170.0711

C1:H„0 171.0808

PO PO: PO, PO,

46.9723 62.9645 78.9579 94.9497

H 63.9627 79.9581 95.9518

H2 80.9639 96.9596

C 106.9512

CH 59.9679 75.9624 107.9508

CH, 60.9751 76.9696

Cil, 110.9734

c2 102.9565 118.9440

C2II 103.9481

C2H2 72.9897

C;H6 109.0085

C2H7 110.0162

C2II» 112.0287

C,IIS 135.0252

CiHio 137.0330

C,H,2 183.0410

C,H„ 184.0515

C,TT„ 185.0618

C,;H, 247.0392

^~ 121 i ï 0 249.0438

C,2H„ 250.0472

267

VII Appendix

CJlv-riagraente

C C, C, c4 C Q cfc C, *~io c„

12 0005 24 0012 36 0019 48 0030 se) 9997 72 0010

H, 13 0086 25 0085 37 0099 49 0100 61 0082 73 0074 97 0070

11 140157 26 0035 38 0169 •>0 0084 62 0155

H, 27 0243 39 0221 M 0240 63 0211

H, 100 0278 112 0287

H, 1010389 113 0375 125 0370

H, 102 0404 114 043<i 126 0425 138 0322

H7 1150531 139 0512

H, 140 0574

268

VII Appendix

1.2 Positive ToF-SIMS Spectra of Phosph(on)ate SAMs

App. 10: ToF-SIMS data (pos. ions) from DDPO3 self-assembled on ÜO2


TaxPON H7-Fragmente OCJHy-Fragmente

OCH 29.002

OCH3 31.0183

OC2H 40.9998

OC2H3 43.0181

OC2H5 45.0332

OC3H3 55.0182

OCsHj 57.0342

OC3H7 59.049

OC7H7 107.0493

Ta TaO Ta02 TaP04

180.9480 196.9425 212.9398

H 181.9544 197.9509 213.9449 276.919

CH2 210.9527

CH, 211.966

C,H 205.9628 221.9513

C,H2 206.9668 222.9574

C2H3 207.9675 223.9639

C2H4 240.9789

C3II 217.9546

C3H2 218.9642 234.9605

C3IT3 219.965 235.9687

C3H4 220.9749 236.9763 252.9747

C3H5 237.9858

C4H2 230.9513 246.9632

C4H3 231.9686

C4H4 232.9794 248.9765

C4H, 249.9846

C4H6 250.9957

PC) PO,

46.9687 62.9852

269

Vil Appendix

CH2-Fragmente

c c2 C, C, C, C, C- C, c9 C,o c„

12.0019 23.985

H 13.0089 24.9909

It 14.0162 26.0152 38.0149 50.0141

H, 15.0235 27.0234 39.0229 51.0223 63.0211

H4 28.031 40.0303 52.0299

Ils 29.0392 41.0389 53.0381 65.0378 77.039

Ü6 30.0421 42.0456 54.0455 66.0452 78.0434 102.0431

Tl7 43.0547 55.0544 67.0532 79.0546 91.0539 103.0517 115.0546 127.0492

H, 44.0584 56.061 68.0607 80.0602 104.0637 116.066 128.0588

H„ 57.0702 69 0703 81.0703 93.0718 105.0675 117.0742 129.0663 141.0712

H« 58.0712 70 077 106.0727 130.0736

II„ 71 0883 83.0856 95.0856 107.083 119.0827 131.0891

Un 85.1041 109.0997

270

VII Appendix

App. 11: ToF-SIMS data (pos. ions) from DDP03 self-assembled on Ti02


TivPO,H,-Fragmentc

Ti TiO Ti02 TiP TiPO TiP02 TiP03 TiP04

47.9581 63.9439 110.9115** 126.9072

H 48.9603 64.9526 80.9459 111.9188 127.9136

H2 49.9571 65.9472 81.9525 128.9052 144.9113

H3 50.9597

H4 83.9680

CH 60.9691

CH2 61.9698 108.9435 124.9333 140.9125

CH, 62.9735 125.9385

CH, 79.9742

CH, 96.9740

C2H 72.9625

QH2 73.9650

QU, 74.9721 90.9670 137.9401

C2H4 75.9775 106.9530

C2U, 76.9878 92.9817

C3U2 85.9699

C3H3 86.9712

C3H, 87.9798

C,I1, 88.9861

caif, 152.9588

C3H7 153.9442

C3H, 154.9231

c,T14 172.9242

C,H, 112.9877

C7H„ 205.9980

C7H„ 208.0039

271

Vil Appendix

HJ"(X - Fragm ctite OCxH, - Fragmen te

OCH, 31.0191

OC2H5 45.0358

OC,H7 59.0453

OC7H, 105.0357

p PO PO, PO, P04

46.9676 62.9763

H 79.9785

H-, 81.9811 97.9770

H, 98.9774

Cil 59.9711 91.9721

Cl-h 77.9772 93.9885

CH5 95.9966

cyi 71.9676 87.9714 103.9724

C2IIi 89.9774 105,9861

C2I1, 60.0100

C2H6 109.0045

C\H2 84.9726 100.9872

C,H, 85.9921 101.9936 117.9843

C2H4 71.0056 102.9935 118.9965

C,H, 72.0078 119.0028

dk 73.0200

CA 113.9917 129.9914

C4H4 130.9923

C,II, 99.9822 116.0059

CA 85.0232 117.0021

C,H„, 149.0235

câ-k 157.0041

CöHh 165.0684

C7H, 152.0026

»—B-ti 1 g 177.0942

C,H2„ 191.1227

C„IIa 217.1399

C,2H14 189.0867

Cn^n 193.1222

CI2H26 233.1741

C12H,8 251.1796

272

VII Appendix

CH-Fragmentc

C c, C, c4 C, Q C, c8 c,„

23.9885

H 13.0114 37.0131

Hj 14.0179 26.0184 38.0198 50.0150

H, 15.0237 27.0233 39.0254 51.0225 63.0226 75.0235

II, 28.0315 40.0321 52.0298 64.0281

H, 29.0383 41.0388 53.0400 65.0378 77.0369

H6 30.0429 42.0473 54.0475 66.0449 78.0447

H, 43.0553 55.0547* 67.0520 79.0515 91.0532 103.0520

Hs 44.0588 56.0626 68.0598 80.0589 92.0568 128.0490

H, 57.0706* 69.0705* 81.0680 93.0678 105.0681

H,„ 58.0744 70.0739 82.0732 106.0727

H„ 71.0833 83.0842 95.0846 107.083

H,2 72.0864 84.0866

H„ 85.1008 97.0995 109.0997

273

VII Appendix

App. 12: ToF-SIMS data (pos. ions) from NEt-DDPC>3 self-assembled on

ÜO2 from a 0.5 mMNEt-DDPOs sol. in ethyl acetate/water 1:1.

TaPOCH-Fragmente

Ta TaN TaO 1 aO\ IaO. Ta02N TaO, Ia04

180 9531 196 9389 212 9409

II 181 9617 195 9535 197 9452 211 9529 213 9402

IU 182 9616 198 9479 214 9448 230 9467

H, 215 9494 231 9584

II, 232 9700 248 9660

CH 193 9572 207 9591 209 9*12

CH, 194 9614 210 9580

CH, 241 9655

CI I 205 9577 221 9524 253 9525 269 9254

CH, 206 9601 222 9542 238 9590 254 9471 270 9487

CH, 22)9610 255 9516 271 9638

CH, 218 9569 234 9*S1

CJ I, 219 9665 235 964 S

CH, 236 9744 2*09610

C,H, 249 9857

TaPO; TaPO,N IaPO, 1 aPO \ IaPO, TaP(W IaPO,

II 276 8765 292 93 36

H. 260 9418

H, 261 9*64

CH 272 943*

CU2 257 9392

CH, 258 955) 274 9400 290 8986

CI I, 259 9467 27* 9*7*

CI I, 286 9101

CH, 287 9*4*.

CH, 332 9450

C-H- 366 96*8

C12H-, 426 0542

CplK 44U)457

C,;H^ 461 0718

274

VII Appendix

C, Hm-rragmente

c C2 C, C, C C C C, c C,„ c„

12 0005

H, 1 ? 0084 37 0068

TU 140158 26 0155 38 0148 50 0149 62 0123

H, 15 0237 27 023 ? 39 0233 51 0228 63 0218 75 0243

H, 28 0250 40 0306 52 0292 64 Ü266 76 0282

H, 29 0396 41 0391 53 0383 65 03S6 77 0383

H, 42 0445 54 0445 66 0447 "8 0441 102 0470 138 0569

II7 43 0554 55 0549 67 0542 79 0536 91 0533 103 0532 1 15 0563 127 0523 139 0450

Hg 68 0562 104 0569 116 0622 128 0643

H, 57 0701 69 07^ 8 1 0699 9? 0639 105 0643 117 0705 129 0695 141 0686

II„ 71 0842 83 0854 95 0848 1190807

Un 85 0979 97 0943

I KPOy-rragmente OCJ 1 -1 ragrnente

P PO PO, PO, PN PO\ PON PO,N CH,0 31 0187

46 9689 Cl 1,0 43 0188

H2 64 9759 80 9725 CH,0 44 0226

H, 81 9780 CH.O 45 0341

C 102 9249 ci 1,0 55 0179

c2n2 56 9619 CHO 109 0563

C2TI, 75 0241

C4I1, 149 0280

C,H,NOCll„ 150 0570

Cl 1,2 151 0603

CfH„ 179 0745 CNOH, 46 0302

CxHn 160 1049 158 H55 CNOH, 58 0308

C,H„ 158 8989 CNOH« 60 0456

ChiHîs 169 1044 CNOH, 73 0592

C,,H„ 257 2058 CNOH, 74 0601

CmHij 276 1640 CNOH, 87 0748

275

VTI Appendix

CnNm-Fragmente

N NC NC2 NC NC NC, NQ

Iii 17.0270

114 18.0353 30.0350

II, 31.0421

H6 44.0512 56.0536 80.0552 92.0548

H7 45.0585

Hs 46 0674 58 0657 70 0697 82.0678 94.0661

[], 59 0720

»,0 60.0761 72.0815 84.0812 96.0800

H,2 86.0961 98.0941

Hi. 100.1130

H„ 101.1155

NC7 NC, NC, NC,U NC,, NC,2

H« 152.0557

H7 153.0604 165.0602

H8 106.0688 118.0682 154.0657

H., 107.0740 155.0741 167.0820

H,o 108.0824 120.0845

H„ 121.0880 157.0788

H,2 110.0977

H,4 112.1106

Jiu, 114.1289 126.1289 138 1246 162.1265

His 128.1444

H;o 142 1619 154.1547

H22 156.1730 168.1720

Tin 157.1736

H24 170.1859

HJ6 184.2003

276

VII Appendix

App. 13: ToF-SIMS data (pos. ions) from NEt-DDPC^ self-assembled on

TiOi from a 0.5 mM NEt-DDPO.3 sol. in ethyl acetate/water 1:1.

Ta^PACH«-Fragmente

Ti TiO Ti02 I iPO TiPO, TiPO, TiPO,

47.9501 63.8925 110.9291 126.9117

H 48.9566 80.9474 111.9380 127.9157

H2 81.9547

cn2 61.9700

CI 1, 159.9417

QH, 136.9136 168.9181

C2H, 90.9713 137.9235

C2H4 138.9239

Cft 148.9433

C-,11, 134.9613

C,H„ 135.9752

CI I, 136.9835

C|H4 146.9354

C4II7 149.9919 165.9764

C7H7 170.9891

Clin 208.003

C.II„ 206.0358

TiP02N TiPO-,N TiP04N I iP20„ TiP20,N Ti?P03 Ti2P20,

170.8716

H 206.8325

CH 169.9087 218.8705

C3H3 213.8898

CH7 207.9548

Cmo 206.9921

ä f*> r~i

VII Appendix

CnHm-Fiagmcnte

C c C c, C, C» C, c, C, CV c„

12 0030

II, r» 0112 17 0059

IL 14 0167 26 0147 18 Olli 50 0101

Hi 15 0225 27 0210 19 0197 51 0244 61 0212

\li 28 0287 40 0261 S2 0121

H; 29 0159 (1 0141 51 0199 65 0194 77 0192

Hr 54 0411 66 0464 7S0158 126 0554 138 0431

41 0495 55 0546 67 05^0 79 05 t: 91 0547 103 0741 115 0661 127 0561 119 0494

Ilj 104 0719 116 0678 128 0625

5n 0707 ö9 0~04 81 0709 105 0890 117 0720 129 0671

Um 106 0849

Hu 71 OSli 95 08(1 107 0882 119 0841

II,, 108 0929 120 0907

Hu 110 1074

Hi« 100 1161 112 1278

Hl! 114 1415

II,, 155 1859

II2I 156 1859

278

VII Appendix

C„Nm-Fragmente

N NC NC, NC, NC, NC, NCt

11 15.0095 26.9981

H2 28.0160

H, 17.0252

11, 18.0323 30.0314 42.0366

H, 31.0376

lk 32.0499 44.0459 56.0521 68.0540 80.0530

Il7 45.0527 93.0628

H, 46.0609 58.0663 70.0676 82.0673 94.0654

II9 59.0702 83.0741

H,o 72.0824 84.0827 96.0813

H12 86.0983 98.0972

HI4 100.1007

H„ 101.1377

NC7 NC, NC, NC,,, NC,, NC,, NC„

H, 105.0562

H, 118.0740 130.0662

Hm 132.0808

H„ 121.0796 133.0805

H,2 122.0986 134.0905

H,3 111.1066

H„ 124.1176 136.1089 160.1166

11h 113.1233 125.1164

Hi6 114.1415 126.1314 138.1208 150.1321

11,8 128.1456 140.1345

H20 142.1505

HM 198.2240

H29 199.2286

H,„ 212.1810

279

VII Appendix

HxPOy-Fi agmente OQH-I"! agmente

Oil, 180177

COU, 30 0059

COU, 11 0146

COH 40 9993

COU, 42 0044

COH, »0114

COHs 4"; 029:>

C,OH, «0135

COH, 57 0290

COH, 111 0598

C„,OH6 1310767

COH,, 142 0397

C,„OH, 144 0660

COH, 145 0655

C„OH,, 146 0621

C„,OH,, 1571615

C„OH, 155 0512

P PO PO, PO, PO, P\ PON PO,N

30 9897 62 9655 78 9793

C 74 9717

CH 59 9718

CH, 76 9745

C 86 9618

CH, 105 9999

CH, 56 9874 106 9884

CH, 116 98S4

CH, 117 9943

CH, 102 9901 119 0007

CH, 103 9944

CH« 104 9999

C,H4 130 9915

CH, 84 0125 1000119

C,HS 101 0118

C,H, 102 0162 114 0102

C,1I,, 1510512

CH, 125 9999

CH, 110 9914 12^0012

CH, 113 003''

CH» 100 0546

QH„ 102 0715

QH, 170 9891

C,H4 109 0500 i2,ono

QU, 156 0810

QH,, 147 0526

QH„ 148 0654

CH, 149 0094

QU, 1 56 0281

CH„ 1550512

QH„ 156 1010

C,„H, 206 0018

C,„II, 207 03 r

Q„H„ 214 07-.9

QiH„ 212 0960

C,,H,, 214 168-.

CNO' CNO

H, tb 0233

Hf, 72 0457

H, 73 0498

H» 74 0607

280

VII Appendix

2 XPS Data

2.1SAMs from Aqueous Solutions

App. 14: XPS data of DDP04 self-assembled from 0.5 mM DDP04(NH4)2solution on different oxide substrate.

Substrate Element Take-off

angle 0

Eb

(eV)

Sensitivity

Factor

Atomic Cone.

%

Ta205 C(ls) 75° 290.2 5.165 27.27

15° 298.6 67.75

O(ls) 75° 535.8 12.903 53.71

15° 535.5 23.18

Ta(4f) 75° 33.4/31.6 52.351 17.41

15° 31.1/33.0 6.71

P(2p) 75° 139.4 8.322 1.61

15° 138.6 2.36

Al203a) C(ls) 75° 298.4 5.165 25.26

15° 298.3 67.58

O(ls) 75° 535.7 12.903 47.13

15° 536.5 21.05

Al(2s) 75° 123.8 4.930 25.51

15° 123.8 8.85

P(2p) 75° 138.8 8.322 2.10

15° 138.8 2.53

a) natural oxygen layer on anodized (25 V) Al

VII Appendix


angle 0

En

(eV)

Sensitivity

Factor

Atomic Cone.

%

Ti02b) C(ls) 75° 286.1 5.165 35.49

15° 286.2 70.70

0(ls) 75° 531.3 12.903 46.75

15° 532.9 22.40

Ti(2p) 75° 459.8/465.6 35.793 15.48

15° 460.0/465.8 3.98

P(2p) 75° 135.1 8.322 À.Zo

15° 135.2 &*ä * V A/

Ti02c) C(ls) 75° 289.5 5.165 31.22

15° 288.9 68.60

0(ls) 75° 534.8 12.903 49.35

15° 534.3 23.04

Ti(2p) 75° 463.2/469.0 35.793 17.59

15° 462.7/468.4 6.13

P(2p) 75° 138.5 8.322 1.83

15° 137.8 2.23

b) 20 nm Ti02 sputter coated on glass

c) 20 nm Ti02 sputter coated on Si wafer

282

VII Appendix


angle 0

EB

(eV)

Sensitivity

Factor

Atomic Cone.

%

Nb205 C(ls) 288.7 5.165 27.00

15° 288.6 63.12

0(ls) 75° 534.2 12.903 53.51

15° 534.6 28.08

Nb(3d) 75° 211.15/213.9 50.437 18.52

15° 211.1/213.9 7.13

P(2p) 75° 138.0 8.322 0.96

15° 137.7 1.67

Zr02 C(ls) 75° 288.6 5.165 35.90

15° 288.25 72.21

0(ls) 75° 533.8d)

12.903 48.97

15° 535.0 e)21.96

Zr(3p) 75° 336.7 31.911 12.86

15° 336.6 3.06

P(2p) 75° 137.8 8.322 2.28

15° 137.2 2.77

highest peak correspond to ZrCh

e) highest peak corresponds to ROPO3

283

VII Appendix


angle 0

EB

(eV)

Sensitivity

Factor

Atomic Cone.

%

Fe203 C(ls) 75° 289.15 5.165 28.93

15° 288.6 66.13

0(ls) 75° 535.5 12.903 56.53

15° 535.35 27.51

Fe(2p) 75° 715.9/729.5 54.865 10.16

15° 714.7/728.1 3.09

P(2p) 75° 137.8 8.322 1.93

15° 137.5 2.86

Ti(2p) 75° 463.3/469.1 35.793 2.46

15° 462.8/468.0 0.41

284

VII Appendix


angle ö

EB

(eV)

Sensitivity

Factor

Atomic Cone.

%

Tio 4Sio 6O2 C(ls) 75° 288.5 5.165 14.90

15° 288.1 26.20

O(ls) 75° 535.6 12.903 60.24

15° 535.8 54.50

Ti(2p) 75° 462.6/468.4 35.793 9.97

15° 462.3/468.2 3.04

P(2p) 75° 137.6 8.322 0.85

15° 137.4 1.13

Si(2p) 75° 106.3 5.792 14.05

1 ro 106.4 15.13

Si02 C(ls) 15° 289.3 5.165 6.20

O(ls) 15° 537.0 12.903 70.83

Si(2p) 15° 107.7 5.792 22.87

P(2p) 15° / 8.322 0.09

285

VII Appendix

App. 15: Curve-fit of OH-DDPO4 SAM on Ta205 produced by immersion

of clean metal oxide sample in an aqueous 0.5 mM OH-

DDP04(NH4)2 solution.

286

VII Appendix

2.2Angle Depending XPS of OH-DDP04(NH4)2

App. 16: XPS Ois detail spectra of OH-DDP04 self-assembled on faÔs

(top) and on NbÔs (bottom), recorded at different take-off angles.

Spectra are compared to a spectrum of a DDP04 SAM on the

corresponding metal oxide, measured at J 5° take-off angle with

respect to the surface plane

5 to M5 534 m M2 5 M 5*0 529 528 ^27

Binding Fnetgy [eV

536 515 514 Ml M2 S ^ I MO 529 528 *527

Binding Fneigy [eV]

JLC

VII Appendix

App. 17: XPS data of C]s (squares) and P2p (dots) of DDP04 self-assembled

on Ta205. The samples were stored in a PBS buffer solution.

Atomic concentrations are plotted as a function of immersion

time.

9500 *a

8500

H 7500

u

~£ 6500 Î

^5500

4500

3500 Î-

0 1 10 100

immersion time [h]

1000

288

\ II Vppendix

3 Contact-Angle and Microdroplet Density Data

3J SAM stability in water

App 18 Contact-angle data from DDPCX DDP()4(Nïl4)2, ODPCX and

ODPO4 self-assembled on la^CX lioni 0 5 inM solutions 111 n-

hcptaiic + 0 8 % 2-propanol, watei. n-heptane ^ 1 5 % 2-propanoland n-heptane + 04% 2-propanol, lespectively The samples were

measmed immediately after the sell-assembly process, after an

additional nnsmg step with puie water and after storage m water

for 3 days

DDPCH ODPOï ODP04 1)DP04(\H4)2

ID before nnsmg with water B alter nnsmg with watet D after stoiage 111 watei ( Id)

289

\ II Appendix

App. 19 Microdroplet density data from DDPCX DDPÖ4(NlIfh ODPCh and

ODPO4 self-assembled on laÔs from 0 5 mlM solutions in n-

heptane + 08 % 2-propanol, water, n-heptane -+ 1 5 % 2-propanoland n-heptane + 0.4 % 2-propanol, respectively The samples were

measured immediately after the êlf-assembly proems, after an

additional rinsing step with pure water and after storage in "water for

3 days

<a

DDPCH ODPCH ODP04 DDPOt(\H4)2

\n before nnsing with watei H aftet iinsmg with watet D aftet stoiage in water (3d)

290

Vil Appendix

4 Optical Sensor Data

4.1 Specific Binding (SB) and \on-Specific Binding (NSB)

App 20 FOBIA data from a ÜDPO, coated 1 a->(\ planar waveguide chip

Non-specific binding (MSB) of Cv5-RlgG is determined on the

BSA-biotm-activated surface (light blue) and after

functionali/ation with streptavidin (deep blue) The specific

binding (SB) after immobilization of captme molecule, aRlgG, is

indicated by the brown line Base line stability of the assay buffer

(green) as well as self fluorescence of streptavidin (pink) and

aRlgG (black) are plotted m parallel

baseline

•2 5e-lÜMO\5-RIsG

stieptavidin

2 5e-IOMC\5-RIgü,

biotm-aRlgG

2Se-H)\fCvS-RlgG

O •*•") \o O^ *^> v 00rH ?»H rH î~H rv| ^î rs)(

data points (time)

291

VII Appendix

App. 21

d

.3<x>

o

s4»o

V!

F0B1A data from a DDP04(NH4)2 coated 'Ya2®5 planar waveguidechip: Non-specific binding (NSB) of Cy5-RIgG is determined on

the BSA-biolm-activatcd surface (light blue) and after

functionali/ation with streptavidin (deep blue). The specificbinding (SB) after immobilization of capture molecule, <xRIgG, is

indicated by the brown line Self fluorescence of streptavidin(pink) and ctRTgG (black) are plotted in parallel.

— 2 5 e-10 M RIgG-Cy5

Stteptavtdni

—2ê-I0 M R1gG-Cy5

—- aRIgG- Alvbiotin

2 5e-10MR]gG-Oy5

G

S iN iN

data points (time)

-f r S

292

VII Appendix

p 22 rOBIA data from a PI 1 (20>g[3 5"|-PL G(2)/PDGhiotin(3 4)10%

coated 1^0-, planar waveguide chip Non-specific binding (NSB)of C\5-RlgG is determined on the BSA-biotin-activated sur race

(light blue) and after functionali/ation with btteptavidm (deepblue) The specific binding (SB) after immobilization of capture

molecule, aRIgG, is indicated by the brown line Base line

stability of the assay buffer (green) as well as sell fluorescence of

streptavidm (pink) and aRIgG (black) are plotted m parallel

1>

4»

3

PBS nnse

— 2 Se-10 Vf RIgG-CyS

Strepta\ldin

— 2 5 e-10 VI RlgG-Cy*

—aRlgG-Ab-biotin

2Se~10V1RTgG-t\5

«Mum*

vo o\

T 1 1

rs) *r 00 s»H "^r^ r\| r-$ f*^ fv*5

I o i"0 ON

data points (tunc)

293

VII Appendix

App. 23.: FOBIA results from biotin-aRlgG/Cy5-RlgG immunoassay on

streptavidin activated ODPO, sensor surface (Application of

different Cy5-RIgG concentrations: 0 - 250 pM). Data point N° 18

(end of precondition) is used to define the zero point of the y-scale. Figure b) shows the zoomed lower standard concentration

region.

littmtlftMliiM»Milr~ Tf !**<

measuring points

measuring points

294

VTI Appendix

App. 24.: FOBIA results from biotin-«RlgG/Cy5-RIgG immunoassay on

streptavidin activated DDPOt. sensor surface (Application of

different Cy5-RIgG concentrations: 0 -- 250 pM). Data point N° 1 8

(end ofprecondition) is used to define the zero point of the scale.

Figure b) shows the zoomed lower standard concentration region.

c0)

mo

c

0)o

w

O)1_

o

KB M Ml Ml Ml HBLiim m ^m

iiiiiiliititm IMHMMMËWMMJiMMllHHH

-r-T-T-c^tMC^rOOcOo fO CO cr>

<t Tj- -«fr •*

measuring points

OfM

50 fM

250 fM

500 fM

-«-2 5 pM

-a-5 0 pM

-*-25pM

--50 pM

^s-250pM

__0fM

-©-50 fM

-•-250 fM

-*-500fM

~*-2 5pM

-&~5 0 pM

~*-25pM

50 pM

250 pM

measuring points

295

VU Appendix

App. 25.: FOBIA results from biotin-cxRïgG/C\5-RlgG immunoassay on

streptavidin activated PI I (20)-g[3.5]-PFG(2)/PEGbiotin(3.4) 10%

sensor surface (Application of different C}5-RIgG concentrations:

0 - 250 pM). Data point N° 18 (end of precondition) is used to

define the zero point of the \-scale. Figure b) shows the zoomed

lower standard concentration reeion.

MlllllMIllllUlllllMMMft,_ ^j. fOt0CDCNLDCOT-^r».OCOtOO>

'r-T-T-tNCNICMrOfOMTt'tJ-TfTt

measuring points

-blank

-50 fM

-250 fM

-500fM

-2 5pM

-5 0pM

-25 pM

-50 pM

-250 pM

c

o

c<uÜCDCDL-

o

üfil§ttttMlfl liai rillMllMiiilillllMllIMiMaCOtDO>CMlDCOT-Tj-h»-OCOC£>0>

measuring points

blank

50 fM

250 fM

500 fM

2 5pM

5 0pM

25 pM

50 pM

250 pM

296

VII Appendix

4.2Immunoassay performance in serum samples

App. 26: Dose-response curve for a RIgG immunoassay on BSA-biotin -

streptavidin activated ODPO4 surface. The different serum concentrations

reflect the amount of newborn calf serum the RIgG dilutions, followed by1:1 mixture with Cy5-ocRIgG-tracer solution of each sample.

0 10 20 30 40 50 60

concentration (pM)

297

Curriculum Vitae

Curriculum Vitae

Personal

Adress: Rolfllofcr

Fleideweg 40

CH-2503 Biel

>1.:-M1 365 39 22

Date of birth:

Martial status:

Children:

Nationality:

Languages:

Februan 3. 1961

married

Nathalie, 20.12.1990

Patrick, 14.4.1994

Swiss, Citizen of Langnau i.F.

Native language is German.

Fluent Fnglish and French and basic knowledge of Italian.

298

Curriculum Vitae

Education and Working Experience

2000 PhD.

Department of Materials, Swiss Federal Institute of Technology, ETH

Zürich, Switzerland. Thesis on 'Surface Modification for Optical

Biosensor Applications'.

1998-2000 PhD. student

Department of Materials, Swiss Federal Institute of Technology, ETH

Zürich, Switzerland / Novartis Pharma AG, CH-Basel

1997 Diploma in Chemistry

Institute of Chemistry, University of Fribourg, CH-Fribourg.

Diploma theses: Inorg. Chem.: 'Synthese neuer Pt(II)-Komplexe mit

duralen Bipyridinderivaten als Liganden.

Org. Chem.: "Isolation und Derivatisierung des Chloro-

phyll-Kataboliten von Cercidiphyllum japonicuin'

1995 Teacher at the professional school

Introduction course for laboratory assistants, CII-Bern.

1993 -1997 Student

Institute of Chemistry, University of Fribourg, CH-Fribourg.

1987-1993 Head of the group

Toxicological Laboratory of the Institute for Forensic Medicine,

University ofBern, CH-Bern.

1982 - 1987 Laboratory assistant

Toxicological Laboratory of the Institute for Forensic Medicine,

University of Bern, CH-Bern.

1980 - 1982 Laboratory assistant

Med. Chem. Laboratory. CH-Biel.

1977 - 1980 Apprenticeship for laboratory assistants

Gebr. Schnyder AG, CH-Biel.

299

Curriculum Vitae

Publications and Patents

2000 Multifunctional Calionic ami Anionic Polymeric Coalings for Surfaces in

Analytic and Sensor Devices

Patent submitted, USA, 2000.

2000 Verfahren zur Abscheidung von Monoschichten aus Organophosphaten und/

oder -phosphonaten und deren Verwendung ah Teil von Sensorplattformen,

Implantaten und medizinischen Hilfsgcraten

Patent submitted. OH, 2000.

2000 Alkyl Phosphate Monolayers, Self, i s scmbled from Aqueous Solution onto Metal

Oxide Surfaces

R.Hofer, M.l cvtor. N.D.Spencer langmuir in preparation.

2000 Self-Assembly of Organic Phosphoric and Phosphonic. \cids onto Tantalum

Oxide and Titanium Oxide Surfaces

R.Hofer, M. 1 extor, N.D.Spencer. E.Jaehne. H.-J,Adler Langmuir in preparation.

2000 Biotin-derivalised poly(L-lysine) graftedpoh (ethylene glycol) as a model inter¬

face for the specific récognition of strcptavidin via planar optical waveguides

J.Vörös, R.Hofer, N.P.Huang. M.Textor. N.D.Spencer Biomaterials in

preparation.

2000 Imaging ofSurface Heterogeneity In the \ficrodroplet Condensation Technique

R.Hofer, M.Textor. N.D.Spencer langmuir submitted, 2000.

2000 Poly(L-lysine)-g-poly(ethylcne glycol) layers on Metal Oxide Surfaces- Surface-

analytical C 'haraclcrization and Resistance to Serum and Fibrinogen Adsorption

N.P.Huang, R.Michel. J.Vörös. M.l extor. R.Hofer, A.Rossi, D.L.Elbert.

J.A.Hubbell. N.D.Spencer langmuir in press. 2000.

2000 The limpact of Xanomter-Scale Roughness on Contact Angle Hysteresis and

Globulin Adsorption

B.Müllcr, M.Riedel. RAlichol. S.DePaul, R.Hofer. D.Heger. D.Grützmacher./

Vac. Sei. Technol. submitted. 2000.

300

Curriculum Vitae

2000 Poly(L-lysine)-g-Poly(ethylene glycol) Layers on Metal Oxide Surfaces:

Attachment Mechanism and Effects ofPolymer Architecture on Resistance to

Protein Adsorption

G.L.Kenausis, J.Vörös, D.L.Elbert, N.Huang, R.Hofer, L.Ruiz-Taylor,

M.Textor, J. A.Hubbell, N.D.Spencer J, Phys. Chem. 5 2000,104, 3298.

2000 Structural Chemistry ofSelf-Assembled Monolayers ofOctadecylphosphoric

Acid on Tantalum Oxide Surfaces

M.Textor, L.Ruiz, R.Hofer, A.Rossi, K.Feldman, G.Hähner, N.D.Spencer

Langmuir 2000, 16, 3257.

1999 Highly Oriented, Self-Assembled Alkanephosphate Monolayers on Tantalum(V)

Oxide Surfaces

D.Brovelli, G.Hähner, L.Ruiz, R.Hofer, G.Kraus, A.Waldner, J.Schlösser,

P.Oroszlan, M.Ehrat, N.D.Spencer Langmuir 1999, 15, 4324.

1993 Bestimmung von Morphin in Spermafhissigkeit

R.Hofer Toxichem und Krimtech 1993, 4, 121.

1993 Glaskörperflüssigkeit als Untersuchungsmaterial


1993 Immunochemisehe Tests (EMIT) als Teil eines Screeningverfahrens


301

Acknowledgements

Acknowledgements

Many people have contributed, through their knowledge and friendship,to make my time at the ETH-Zürich and Novartis Pharma AG Basel such a

fruitful period of my life. In particular, T would like to express my thanks to

Prof. Dr. N.D. Spencer, ETH Zürich, for providing the opportunity to

do this interesting work at the Laboratory for Surface Science and

Technology, and for his support and confidence;

Dr. M. Textor, ETH Zürich, my advisor, for stimulating discussion,

permanent support and continued advice, for the familial atmosphere, for

performing ToF-SlMS measurements, and for teaching me thoroughness in

scientific thinking;

Dr. M. Ehrat, Zeptosens AG, Witterswil, for making it possible to have

access to high-performance apparatus and for supporting me through the

whole thesis;

People from Laboratory for Surface Science and Technology (LSST),ETHZ for the pleasant cooperation, especially V. Frauchiger for his help in

computer-trouble-shooting;

People of Zeptosens AG, Witterswil, especially Dipl. Ing. Chem. E.

Schürmann for helping me in the immunoassay measurements and

evaluations, Dr. A. Abel and Dr. M. Pawlak for their interesting inputs and

theoretical support, and Dr. G. Duveneck for helping in writing a patent;

Dr. L. Ruiz-Taylor, ETH Zürich, supervisor for the first year of the

thesis, for her help and introduction in the use of X-ray pliotoelectron

spectroscopy;

Dr. G. Kenausis and Dr. J. Vörös, LSST, Schlieren for executingOWLS measurements;

302

Acknowledgements

Dr. G. Hähner, Dr. D. Brovelli and I. Pfund-Klingenfuss for performingNEXAFS analyses;

Prof. A. Rossi, ETHZ/LSST, for supporting me in XPS measurements

and for the work-out concerning the application of the 3-layer model;

Dr. S. De Paul, ETHZ/LSST, for the careful proof reedings of the

dissertation;

F. Bangerter, ETHZ, for measuring NMR spectra;

People from Novartis Pharma AG, Basel, especially Dr. D. Neuschäfer,Dr. W. Budach, Dr. G. Kraus and Dr. P. Oroszlan, for introducing me in the

theme of immunoassays and the planar-waveguide technique;

Dr. D. Anselmetti, Novartis Pharma AG, Basel, and K. Feldman, ETH

Zürich, for AFM measurements;

Dr. E. Jahne and Prof. Dr. H.-J. Adler, Institut fur Makromolekulare

Chemie und Textilchemie, Dresden, for the synthesis of alkyl phosph(on)atederivatives;

Dr. Hendrich, University of Würzburg, for executing toxicology tests

on alkyl phosph(on)ate SAMs;

The Commission for Technology and Innovation (CT1) and the

MedTech Initiative for financial support.

Zürich, 2000

303

Documents

Surface modification for optical biosensor applications