UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Studies of metal-polymer interfaces Kono, Mari 2000

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

Item Metadata

Download

Media
831-ubc_2001-611280.pdf [ 8.52MB ]
Metadata
JSON: 831-1.0061347.json
JSON-LD: 831-1.0061347-ld.json
RDF/XML (Pretty): 831-1.0061347-rdf.xml
RDF/JSON: 831-1.0061347-rdf.json
Turtle: 831-1.0061347-turtle.txt
N-Triples: 831-1.0061347-rdf-ntriples.txt
Original Record: 831-1.0061347-source.json
Full Text
831-1.0061347-fulltext.txt
Citation
831-1.0061347.ris

Full Text

STUDIES OF METAL-POLYMER INTERFACES  by MARI  KONO  B . S c , Carleton University, 1994  A THESIS SUBMITTED IN P A R T I A L F U L F I L L M E N T OF THE REQUIREMENTS FOR THE D E G R E E OF DOCTOR OF PHILOSOPHY in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Chemistry)  We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH C O L U M B I A OCTOBER 2000 © Mari Kono, 2000  In  presenting  degree freely  at  this  the  available  copying  of  department publication  of  in  partial  fulfilment  of  University  of  British  Columbia,  I  for  this or  thesis  reference  thesis by  this  for  his thesis  and  scholarly  or for  her  Department  Date  DE-6  (2/88)  Columbia  purposes  gain  that  agree  may  shall  requirements  agree  I further  representatives.  financial  permission.  T h e U n i v e r s i t y o f British Vancouver, Canada  study.  the  It not  be  that  the  Library  permission  granted  is  by  understood be  for  allowed  an  advanced  shall for  the that  without  make  it  extensive  head  of  my  copying  or  my  written  Abstract  The chemistry of surfaces and interfaces formed by polymers and metals have been studied by surface science methods with the overall objective of gaining improved insights into interfacial adhesion. Three areas were emphasized: (i) interactions of thermally deposited metals with particular polymer surfaces; (ii) investigation of the possibility for using a remote hydrogen plasma for modifying polymer surfaces in controlled manners; and (iii) studies of the bonding of organosilanes to aluminum surfaces, in part with reference to developing nonchromating processes for corrosion protection. Interfaces formed between polyethyleneterephthalate (PET) and thin layers (<10 A) of thermally deposited aluminum or zirconium were characterized by X-ray photoelectron spectroscopy (XPS); it was concluded that the Zr bonds especially to O atoms in the PET surface, whereas the A l bonds both to carbonyl groups and to the aromatic rings.  The  conclusion for Al was suggested by new components in C ls at 283.9 eV and 282.6 eV, and by the O 1 s^component at 530.8 eV. Comparisons were also made with the Mg/PET interface, which shows M g bonding as both O-Mg-C and metallic-like clusters. The deposition of Al on the Mg/PET sample results in some intermetallic mixing, which is manifested by alloy formation and M g enrichment at the surface, but it also appears that some A l atoms penetrate the M g region to react with carbonyl groups in the PET to form Al-C bonding. The stability of this sample was studied on exposing to air and to water. Although the metal overlayers are strongly modified by such treatments, it is suggested that the A l may help passivate the Mg/PET interface, for example in relation to light emitting diodes. The deposition of Al on polyphenylvinylene (PPV) showed Al-O-C and Al-C interactions at the interface, although appreciable oxidation occurs on heating to 80°C apparently via oxygen from the PPV bulk.  ii  The effects of the remote hydrogen plasma process were studied on the three polymers polystyrene (PS), poly(butylmethacrylate) (PBMA) and PET.  The incorporation of new  functional groups involving oxygen was established by curve fitting the XPS spectra, and these changes arise from small amounts of O-containing molecules (e.g. H 0 ) in the background gas. 2  After a standard treatment, the oxygenated components in C ls spectra correspond to 7.8%, 3.0% and 6.5% of the total C ls envelope area for PS, P B M A and PET respectively, and this indicates that the extent of modification depends on the specific polymer structure. Surfaces of high-purity aluminum, which have been given different pre-treatments and subsequent exposures to air and water, were shown to differ markedly with regard to chemical o  o  composition (e.g. the oxide film thicknesses range from 29 A to 88 A , while the fractions of oxygen in O H form vary from 0.2 to 0.9) and topographical structure (e.g. the mean roughness value ranges from 37 nm to 77 nm). Characterizations of these surfaces after exposures to three  organosilanes,  y-glycidoxyl  propyltrimethoxyl  silane  (y-GPS),  Bis-1,2-  (triethoxysilyl)ethane (BTSE) and y-aminopropyl trimethoxyl silane (y-APS), indicate that the amount of silane adsorbed in each case shows a tendency to increase both with the number of OH groups detected at the oxidized aluminum surface and with the surface roughness.  The  XPS data are consistent with the adhesion of y-APS occurring by H-bonding through N H  + 3  groups.  in  Table of Contents  Abstract  u  Table of Contents  iv  List of Tables  vii  List of Figures  ix  List of Abbreviations  xiii  Acknowlegements  xiv  Chapter 1: Introduction  1  1.1 Background  1  1.2 Mechanism of adhesion  1  1.2.1 Mechanical interlocking theory  3  1.2.2 Electrostatic theory  3  1.2.3 Chemical bonding theory  5  1.3 Motivation for this work  5  1.3.1 Polymer light-emitting diodes  5  1.3.2 Silane-coupling agents  7  1.4 Objectives for this work  Chapter 2: X-ray photoelectron spectroscopy  13  14  2.1 Introduction  14  2.2 Qualitative analysis  15  2.3 Quantitative analysis  19  2.3.1 Inelastic scattering and sampling depth  19  2.3.2 Atomic concentration determination  22  2.3.3 Angular dependent measurements  24  2.4 Instrumentation  26  2.4.1 Ultrahigh vacuum (UHV)  28  2.4.2 X-ray sources  29  2.4.3 Energy analyzer (EA200)  33  2.4.4 Data processing  37  Chapter 3: Characterization of some metal-polymer systems  43  3.1 General introduction  43  3.2 XPS investigations of A l - P E T and Zr-PET interfaces  44  3.2.1 Introduction  44  3.2.2 Experimental  47  3.2.3 Results and discussion  49  3.2.4 Conclusions  68  3.3 XPS studies of the stability of Al/Mg/PET interfaces  71  3.3.1 Introduction  71  3.3.2 Experimental  72  3.3.3 Results and discussion  74  3.3.4 Conclusions  84  3.4 Characterization of Al/PPV interfaces by XPS and A F M  84  3.4.1 Introduction  84  3.4.2 Experimental  85  3.4.3 Results and discussion  91  3.4.4 Conclusions  100  Chapter 4: Modification of polymer surfaces by a remote plasma treatment  102  4.1 Introduction  102  4.2 Experiment  105  4.2.1 Sample preparation  '  105  4.2.2 Plasma treatment  105  4.2.3 Surface characterization  107  4.3 Results and discussion  107  4.3.1 Polystyrene (PS)  107  4.3.2 Poly(butylmethacrylate) (PBMA)  115  4.3.3 Polyethyleneterephthalate (PET)  118  4.4 Concluding remarks  121  Chapter 5: Characterization of aluminum surfaces after different pre-treatments and exposure to organosilane coupling agents 124  5.1 Introduction  124  5.2 Experimental  125  5.2.1 Scanning electron microscopy (SEM) 5.3 Results and discussion  127 128  5.3.1 Aluminum surfaces  128  5.3.2 Silane interactions  140  5.4 Conclusions  145  Chapter 6: Concluding remarks and future directions  147  6.1 Investigation of metal-polymer interfaces  147  6.2 Plasma modification of polymer surfaces  149  6.3 Silane treatment on Al surfaces  150  References  153  List of Tables Table 1.1  Summary of analytical techniques used for polymer-metal interface characterization.  Table 1.2  Some organosilanes of interest.  Table 3.1  C l s binding energies (in eV) and percentages of the different carbon components indicated for the initial PET film (measurements made for 90° exit angle), after deposition of A l (90° and 30° exit angles), and for Al-PET system after exposure to air (90° and 30° exit angles).  64  C l s and O ls binding energies (in eV) and percentages of different components indicated for Mg/PET and Al/Mg/PET including different exit angles 9.  75  Table 3.2  Table 3.3  M g 2p binding energies and percentages of oxide and metallic components indicated by XPS for different metallized PET samples (normal exit direction).  Table 3.4  A l 2p binding energies and percentages of oxide and metallic components indicated by XPS for different metallized PET samples (normal exit direction).  Table 4.1  Carbon and oxygen component binding energy values (eV) used in curve fitting analysis.  Table 4.2  Relative compositions (%) from curve fitted C ls spectra from PS film prior to and after plasma treatment. Results are given for two take off angles for plasma-treated PS.  Table 4.3  Relative compositions (%) from curve fitted C ls spectra from P B M A film prior to and after plasma treatment. Results are given for two take off angles for plasma-treated P B M A .  117  Relative compositions (%) from curve fitted O ls spectra from P B M A film prior to and after the plasma treatment. Results are given for two take off angles for plasma-treated P B M A . O/C ratios measured from XPS spectra are also included.  117  Table 4.4  108  Vll  Table 4.5  Relative compositions (%) from curve fitted C ls spectra from PET film prior to and after plasma treatment. Results are given for two take off angles for plasma-treated PET.  119  Relative compositions (%) from curve fitted O ls spectra from PET film prior to and after the plasma treatment. Results are given for two take off angles for plasma-treated PET. O/C ratios measured from XPS spectra are also included.  119  Table 5.1  Summary of pre-treatments and their codes used in this work.  126  Table 5.2  Percentage compositions, fraction of OH in aluminum oxide film (f H) and thickness of oxide film (t ) indicated by XPS after different pre-treatments.  134  Percentage compositions from XPS after different silane treatments.  141  Table 4.6  0  Table 5.3  ox  Vlll  List of Figures Figure 1.1  Figure 1.2  Schematic illustrations of proposed mechanisms for adhesion: (a) mechanical interlocking; (b) electrostatic attraction; (c) direct chemical bond formation.  4  Range and bond strength of typical interatomic and intermolecular interactions.  6  Figure 1.3  Schematic drawing of a polymeric light-emitting diode.  8  Figure 1.4  Progress in L E D efficiency.  9  Figure 1.5  Proposed mechanism for interaction between an organosilane and a metal surface. Production of a photoelectron by irradiation with soft X-rays and two processes for de-excitation of the atomic core hole produced.  Figure 2.1  Figure 2.2  12 16  XPS spectra from polyethyleneterephthalate (PET) film showing: (a) low-resolution survey spectrum, and (b) C ls spectrum at higher resolution with allocation to component peaks.  17  Extra structures in XPS spectra: (a) spin-orbit coupling for Zr 3d; and (b) plasmon loss structure (pl) associated with M g 2s and M g 2p spectra.  20  Figure 2.4  Inelastic mean free path of electrons as a function of kinetic energy.  21  Figure 2.5  Notation for: (a) photoelectron intensity from a narrow band {dx) in a semi-infinite homogeneous sample, and (b) a sample with an overlayer thickness t on top of a substrate.  23  Figure 2.3  Figure 2.6  Angle dependent measurements: (a) definition of take-off angle 0; and (b) A l 2p spectra measured at two different take-off angles.  25  Figure 2.7  A schematic diagram of the MAX200 system as viewed from above.  27  Figure 2.8  Pumping system for the MAX200 facility.  30  Figure 2.9  Sample transfer system in the MAX200 facility: (a) between different chambers; and (b) five degrees of movement possible for sample on the manipulator in analysis chamber.  31  Figure 2.10 Schematic diagrams of: (a) the Mg/Al dual anode X-ray source; and (b) energy distribution of photons produced, showing the K and Kp lines and the high-energy Bremsstrahlung radiation.  32  Figure 2.11 Schematic diagram for the concentric hemispherical analyzer and input lens system in the MAX200 system.  34  Figure 2.12 Energy levels for binding energy measurements.  38  Figure 2.13 Illustration of Shirley non-linear background subtraction: (a) applied to a raw C ls spectrum; and (b) result of curve fitting to two components after correcting for the background.  40  a  Figure 3.1  Monomer units for polymers studied in this work: (a) polyethyleneterephthalate (PET), (b) polystyrene (PS), (c) polyphenylvinylene (PPV), and (d) polybutylmethacryate (PBMA).  45  Schematic diagrams of evaporation sources: (a) Zr or A l , and (b) Mg.  48  Figure 3.3  Schematic diagram of steps for experiments in Section 3.2.  50  Figure 3.4  C ls spectra: (a) clean PET film with components; (b) comparison for PET film and Zr/PET; (c) difference spectrum Zr/PET minus clean PET as described in text; and (d) curve fitted spectrum for Zr/PET.  51  O ls spectra: (a) clean PET film; (b) comparison for PET film and Zr/PET; (c) difference spectrum Zr/PET minus clean PET as described in text; (d) curve fitted spectrum for Zr/PET.  53  Figure 3.6  Zr 3d spectra: (a) Zr/PET film, and (b) Air-treated - Zr/PET.  58  Figure 3.7  Spectra for Al/PS film: (a) O 1 s, and (b) A l 2p.  Figure 3.8  Spectra for Al/PET film: (a) comparison of C ls for clean PET and Al/PET; (b) curve fitted C ls spectrum for Al/PET; (c) comparison of O ls for clean PET and Al/PET; (d) curve fitted O ls spectrum for Al/PET.  62  Spectra for Oa-treated Al/PET sample: (a) comparison of A l 2p prior to treatment and after, and (b) C l s .  66  Figure 3.10 Schematic indications of metal-PET interaction models: (a) Zr/PET system; (b) Al/PET system, and (c) Al/PET after O2 treatment.  69  Figure 3.2  Figure 3.5  Figure 3.9  60  X  Figure 3.11 Schematic diagram of steps for experiments in Section 3.3.  73  Figure 3.12 Spectra from Al/Mg/PET sample: (a) M g 2p, and (b) A l 2p.  77  Figure 3.13 Spectra from Al/Mg/PET samples prior to (thin line) and after (thick line) the air exposure (see text): (a) C ls; (b) M g 2p, and (c) A l 2p.  80  Figure 3.14 Spectra from Al/Mg/PET samples prior to (thin line) and after (thick line) the water rinsing (see text): (a) M g 2p; (b) A l 2p, and (c) C ls.  82  Figure 3.15 Thermal conversion of polyjp-phenylene [l-(tetrahydrothiophen-lio)ethylene chloride]} (THT) to polyphenylvinylene (PPV).  86  Figure 3.16 Sample notations and procedures for experiments in Section 3.4.  87  Figure 3.17 Diagrams for A F M : (a) the operation of A F M ; and (b) cantilever oscillation amplitude in free air and during scanning for tapping mode A F M .  90  Figure 3.18 XPS spectra from annealed PPV film: (a) curve fitted C ls, and (b) O Is.  92  Figure 3.19 A l 2p spectra and A F M micrographs for PPV and Al/PPV samples at various A l coverages.  94  Figure 3.20 XPS spectra from Al/PPV films: (a) comparison of O Is spectra for AP and LAW samples; and (b) curve fitted C ls spectrum for sample MAIP.  Figure 3.21 Comparison of XPS spectra for MAIP and AMAIP sample: (a) A l 2p, and(b)0 1s. Figure 4.1  Figure 4.2  Figure 4.3  Figure 4.4  98  99  Schematic indication of processes involved in a plasma formed by molecular gas A B .  103  Schematic diagram of the system used for the remote hydrogen plasma treatment.  106  Survey spectra from (a) untreated PS film, and (b) plasma treated PS film.  109  C ls spectrum from untreated PS film.  1• •  Figure 4.5  C ls and O ls spectra from plasma treated PS film measured at different take off angles.  113  Curve fitted XPS spectra from P B M A film: (a) C ls and (b) O ls, both prior to plasma treatments; (c) C ls and (d) O ls, both after plasma treatment.  116  Curve fitted XPS spectra from PET film: (a) C ls and (b) O ls, both prior to plasma treatments; (c) C ls and (d) O ls, both after plasma treatment.  120  Figure 5.1  Illustration of a Hitachi S-4100 S E M .  129  Figure 5.2  Diagram of S E M main components.  130  Figure 5.3  Resolved A l 2p and curve-fitted O ls spectra measured for the pretreated samples designated Al, Ac, MandAW'm Table 5.1.  132  Values of oxide thickness (t ) and fraction of hydroxide (foH) measured from XPS for different samples considered in this work.  136  Figure 4.6  Figure 4.7  Figure 5.4  Figure 5.5  Figure 5.6  Figure 5.7  Figure 5.8  Figure 5.9  ox  A F M micrographs measured for differently pre-treated samples: (a) M, (b) Ac, and (c) Al.  138  S E M micrographs measured for aluminum samples given five different treatments (M, Ac, Al, AcW and AlW) and for the alkalineetched samples treated with y-GPS (GAl).  139  Curve-fitted N ls spectra measured from the sample designated AM (i.e. y-APS added after mechanical polishing) for different take-off angles: (a) 90°, and (b) 30°.  143  Plots of Si and N composition and NH2/NH3 ratio estimated by XPS (normal take-off angle) for different samples treated with y-APS.  '44  Illustration of 'up-side interaction' between y-APS and A l surface.  '46  +  Xll  List of Abbreviations  AFM  Atomic force microscopy  y-APS  y-aminopropyl trimethoxyl silane  BTSE  Bis-1,2-(triethoxysilyl)ethane  CHA  Concentric hemispherical analyzer  FWHM  Full width at half-maximum  y-GPS  y-glycidoxyl propyltrimethoxyl silane  HREELS  High resolution electron energy loss spectroscopy  PMFP  Inelastic mean free path  ISS  Ion scattering spectroscopy  ITO  Indium-tin oxide  LMWOM  Low-molecular-weight-oxidized-materials  MCP  Multi-channel plate (detector)  PBMA  Polybutylmethacrylate  PET  Polyethyleneterephthalate  PLED  Polymer-based light-emitting diode  PPV  Polyphenylenevinylene  PS  Polystyrene  SEM  Scanning electron microscopy  SPMS  Secondary ion mass spectrometry  SPM  Scanning probe microscopy  STM  Scanning tunneling microscope  TEM  Transmission electron spectroscopy  THT  Poly{p-phenylene[ 1 -(tetrahydrothiophen-1 -io)ethylene chloride]}  UHV  Ultra high vacuum  UPS  U V photoelectron spectroscopy  XPS  X-ray photoelectron spectroscopy  WBL  Weak boundary layer  xiii  Acknowledgements  I would like to thank my supervisor, Professor K.A.R. Mitchell, for introducing me to a challenging field of materials science, and for his persevering guidance and support throughout the period of the research. I also would like to thank Dr. P.C. Wong, Dr. K.C. Wong, Ms. R. L i , Dr. X . Sun, Ms. D. Susac, Dr. Y . L . Leung and Ms. L . Shi for their help on the XPS and S E M measurements, and for their useful discussions on numerous occasions. Additionally I am grateful for help received in the laboratories of Prof. T. Tiedje, for use of S E M and A F M equipment, and Prof. M . Wolf for preparation of PPV polymer. I would like to express special recognition to members of the Departmental Mechanical and Electrical shops; without their devotion and skill, many of my experiments would never have been achieved. To both my former and present colleagues, I would like to express my sincere gratitude for your constant encouragement. I feel privileged to have been able to share time with each one of you during the past years. Finally, I would like to thank my friends and family in Japan and Canada: Kay, Frank, Paddy, Dr Koga, Mrs. Koga, Mom, Dad and Yohei for their love. To them I dedicate this thesis.  XIV  Chapter 1 Introduction  1.1  Background An understanding of the interfacial regions between metals and organic polymers has  long been an important goal in materials chemistry. Metal-polymer interfaces have important roles in a variety of industrial products including food-packaging [1], microelectronics [2,3], semiconductor devices [4], corrosion protection [5] and biomedical practice (prosthetic devices) [6].  Measuring and understanding the characteristics of the interfacial regions  involved should enable new insight into the associated adhesion. In recent years, investigations of the growth modes [7,8], compositions and bonding arrangements [9-13] of metallized polymer surfaces have been made with various surface analysis techniques (Table 1.1). Among those techniques, X-ray photoelectron spectroscopy (XPS) is one of the most widely used methods for characterizing metal-polymer interfaces. Often, the original surfaces may not be suitable for forming strong adhesion because of the effects of surface contamination, the types of chemical functional groups present and unfavorable topological features. Accordingly some kind of pre-treatment may be required to overcome these interfering influences.  Such  treatments include mechanical ablation, chemical etching, plasma surface modification and application of coupling agents [14].  1.2  Mechanism of adhesion There has been much controversy concerning the factors controlling adhesion between  polymers and metals. Several mechanisms have been proposed [21,22] and these have been  Table 1.1 Summary of analytical techniques used for polymer-metal interface characterization.  Technique  Acronym  P h y s i c a l measurement  Information g a i n e d  A t o m i c force microscopy  AFM  M e a s u r e m e n t o f interatomic forces between fine cantilever tip and surface  Surface morphology  [15]  A u g e r electron spectroscopy  AES  K i n e t i c energies o f electrons emitted b y two electron deexcitation o f core hole  Elemental composition  [16]  H i g h resolution electron energy loss spectroscopy  HREELS  M e a s u r e s m a l l energy losses in a reflected l o w - e n e r g y electron beam  Vibrational modes  [17]  Ion scattering spectroscopy  ISS  M e a s u r e directions a n d energies o f scattered ions  Elemental c o m p o s i t i o n o f top most layer  [18]  Infrared spectroscopy  IR  A b s o r p t i o n o f IR radiation as function o f energy  Perpendicular vibration m o d e s  [19]  S e c o n d a r y ion m a s s spectrometry  SIMS  M e a s u r e ions sputtered from surface in mass spectrometer  Elemental composition  [20]  U V photoelectron spectroscopy  UPS  P h o t o e m i s s i o n f r o m valence electron shells  V a l e n c e structure  [16]  X - r a y photoelectron spectroscopy  XPS  M e a s u r e kinetic energies o f photoelectrons f r o m core levels  Elemental composition, c h e m i c a l states  [16]  Refs.  crystallized into three main categories: the mechanical interlocking theory, the electrostatic theory and the adsorption theory. These theories can be used, solely or in combination with each other, to describe many adhesion phenomena.  The mechanisms are schematically  illustrated in Fig. 1.1 and they are briefly described below. 1.2.1  Mechanical interlocking theory According to this theory, a major source of adhesion is from the physical interlocking or  mechanical keying that occurs when adherent species penetrate into irregularities or porous structures at a substrate surface.  Examples include the etching of polymer surfaces to  strengthen adhesion to deposited metals [23], and the repairing of teeth when a dentist drills out a cavity and fills it with mercury amalgam or a polymeric material. After it has been hardened and cured, the filling is held in place by mechanical interlocking. This theory does not have general applicability, which is evident from the fact that good adhesion has also been obtained with smooth surfaces such as glass [24]. But in either situation van der Waals forces operate for close approach of atoms from the two sides of the interface.  1.2.2  Electrostatic theory The electrostatic theory was developed by Derjaguin and co-workers [25], who  proposed that the adhesion between two solid materials could result from an electrical double layer formed by electron transfer from one material to the other. The theory originally arose from observations made during the fast peeling of dry pressure-sensitive tape, which resulted in light emission and a crackling noise. The fact that the broken parts of the adhesive specimens were sometimes charged was taken as evidence for this theory.  Indeed, it is this latter  statement that has caused the main criticism of the theory. It has been pointed out that the charging may have been produced during the peel-testing itself, but other results in support of  3  a o  a -a c o Xi  I | (U  CU  +jH »O 1) CCJ  C+H ,»H c/2 T 3  'S  o o cu  -5  e2 CU 1/1  O  03  ri .2  ca O . *->  P.  O  o t5 CO  o  <U CU  eoo  cu  9  'S <U  o cu  4  the electrostatic theory depend on a loss of adhesion during sample exposure to an electrical discharge [26]. 1.2.3  Chemical bonding theory The chemical bonding (or adsorption) theory is accepted as the most widely applicable  theory for the mechanisms of adhesion [14,21]. It should be expected that basic adhesion increases if the adhering phases undergo chemical reaction with each other to form primary bonds of the covalent and ionic types at the interface.  In the case of polymer-metal  interactions, metal-organic bonds are formed between reactive surface species and the adherent, and this type of interaction is the main subject of the current thesis. Other weaker interactions involve secondary bonds such as hydrogen bonds and van der Waals forces (non-polar dispersion forces, dipole-dipole interactions and dipole-induced dipole interactions). Figure 1.2 indicates bond strengths and ranges of these interactions [27].  1.3  Motivation for this work Two particular topics, which gave a motivation for this work from the practical point of  view, are introduced in this section; both are areas with large potential industrial application, and for both the metal-polymer interface plays a crucial role. One concerns developments in polymer-based light-emitting diodes (PLEDs), and the other has to do with the use of organosilane coupling agents for corrosion protection of metals.  1.3.1  Polymer light-emitting diodes The first polymer light-emitting diode (PLED) based on a thin film of conjugated  polymer was reported by Friend and his co-workers in 1990 [28]. The structure of the L E D consisted of poly(/?-phenylenevinylene), or PPV, sandwiched between two electrode materials,  5  0  0,1  0-2  0.3  OA  0.5  interroolecuiar d i s t a n c e r (nm)  Figure 1.2 Range and bond strength of typical interatomic and intermolecular interactions. Ref. [27].  namely indium tin oxide (ITO) and aluminum. Figure 1.3 illustrates the basic structure of a single-layer PLED.  As voltage is applied between the two electrodes, electrons from the  metallic cathode and holes from the ITO migrate and combine in the PPV layer. This results in photon production, and this electroluminescence can be emitted through the transparent ITO layer.  The principal interest for PLED technology, over the use of traditional inorganic  materials, lies in the opportunity for low-cost manufacturing, especially for large area display. An active layer of polymer film can be prepared rather simply by spin coating a polymer solution on to the substrate material in air [29].  Since the discovery of PLED devices, an  explosive growth of activity has occurred in both academia and industry, and this opens the challenge for materials science to understand issues associated with the synthesis and properties of the materials, as well as device properties [29-35].  Owing to numerous efforts, the  performances of today's P L E D devices have nearly reached the level of the inorganic LEDs [36] (Fig. 1.4), but one major obstacle remains on the way to commercialization, and this concerns the insufficient life-time [37]. In broken down devices, the degradation of the cathode and polymer interfaces have customarily been reported [29,38]. Savvate'ev et al. observed a delamination of the cathode material from the polymer layer [39], and several other workers [38,40,41] have reported the formation of non-emissive dark spots at cathodes.  Despite all  these studies, no conclusive theory has been established to explain phenomena observed in failed P L E D devices.  More complete and precise understanding of the metal-polymer  interfacial region is therefore required for devices of the future.  1.3.2  Silane-coupling agents  Organosilanes are traditionally used as adhesion promoters (i.e. coupling agents) between a mineral substrate and an organic coating, such as a paint, and more recently there  7  Figure 1.3  Schematic drawing of a polymeric light-emitting diode.  8  100  yelo*  ~f  1960  i  1965  1 1970  r—  1975  1 1 1980 1985 year  Figure 1.4 Progress in L E D efficiency. Ref. [36].  - r " 1990  1 1995  1— 2000  have been hopes that they can act as corrosion inhibitors, especially to replace the conventional and environmentally hazardous use of the chromating [42,43].  For the passivation of  aluminum surfaces, in particular, new procedures are needed to allow the phasing out of the chromating process, particularly given the health concerns associated with it [44]. The general formula of an organosilane is R-SiK'3, where R' is a hydrolyzable group on silicon and R is an organofunctional group chosen for compatibility with the organic coating to be subsequently added.  This bifunctional nature provides a hybrid linkage between organic and inorganic  materials.  Table 1.2 lists several commercially available silane-coupling agents with an  indication of their functional groups, chemical structure and the acronym used in each case. Even small applications of organosilane to mineral surfaces can significantly improve their adhesion to organic paints and protect against corrosion [42,45]. To this date, questions of how and why silanes give improvements in the adhesion remain challenging [46-48], although Fig. 1.5 illustrates a commonly accepted mechanism for organosilane adsorption on a metal [42]. The silane is initially believed to undergo a step-wise hydrolysis in aqueous solution to the silanetriol (reaction 1). Then hydrogen bonding may be established (reaction 2), and it is believed that during drying covalent M-O-Si linkages are formed in the interface after eliminating water molecules. The remaining silanol group may then condense with an adjacent silanol to form polysiloxane or remain partly uncondensed at the surface (reaction 3). Efforts have been made to show the presence of covalent metal-O-Si bond formation at metal-to-silane interfaces, and hence to relate to enhanced adhesion [49], but the evidence has generally been indirect. Various investigations have used vibrational spectroscopy to identify the metal-siloxane bond [50-54], and earlier evidence from our group [55,56] suggested that the presence of Si-O-Al linkages formed during the adsorption of y-GPS molecules on aluminum  10  Table 1.2 Some organosilanes of interest.  Acronym  Formula  VS  CH =CHSi(OCH )  CPS  ClCH CH2CH Si(OCH3)3  GPS  Functional group  2  3  2  Vinyl  3  Chloropropyl  2  A  Epoxyl  CH CHCH OCH CH2CH2Si(OCH )3 2  2  3  2  APS  H NCH CH CH Si(OCH )  MPS  CH =CCH COOCH CH CH Si(OCH )  MGPS  HSCH CH CH Si(OCH )  BTSE  (C H 0) SiCH CH Si(C H 0)  2  2  2  2  3  3  2  2  2  5  2  3  2  2  2  BTSE is a non-functional silane  2  3  2  Primary amine  3  2  3  Methacrylate Mercapto  3  2  3  5  3  *  C3 <U  cd  T3 C cd CD  l O C cd 6X) LH  O cd  c  CU CU CU  C  o o cd  c  cd  o <U  6  -a cu  i*>  O & O  CU S-i  12  surfaces could be identified by a new bias-potential technique in XPS [57]. Later Fang et al. provided support for this interpretation by identifying S i O A F fragments in the positive SIMS spectrum measured from this system [58].  This latter approach has been used earlier by  Getting and Kinlock to support the presence of direct Si-O-Fe linkages formed during the adsorption of y-GPS on steel surfaces [59].  1.4  Objectives for this work The general goal of this work is to explore the interactions between various metals and  polymers under different conditions, and in some cases to investigate the effect of pre-treatment on the original surfaces prior to the application of a second material. The thesis consists of six chapters. After this general introduction to metal-polymer interfaces, Chapter 2 introduces X ray photoelectron spectroscopy (XPS), the main analytical technique used throughout this research.  Chapters 3 and 4 report results from systems prepared in ultrahigh vacuum  environments, and they involve characterization of metal-polymer interfaces and plasma modified polymer surfaces respectively.  The former topic focuses mainly on the initial  interactions experienced by polymer surfaces after metals have been thermally deposited on them. The second of these chapters investigates the effects of applying a remote hydrogen plasma to various polymer surfaces, with a view to assessing the degree to which the surface modification is controlled in relation to the resulting functional groups. Chapter 5 focuses on a characterization of aluminum surfaces after different pre-treatments, and the effects of these pre-treatments on the adsorption of three silane coupling agents. A summary of the new results and possible future directions are given in Chapter 6. In parts of this work, atomic force microscopy (AFM) and scanning electron microscopy (SEM) are also employed as supplementary characterization techniques.  13  Chapter 2 X-ray photoelectron spectroscopy  2.1  Introduction X-ray photoelectron spectroscopy (XPS) is one of the most widely used techniques in  contemporary surface characterization. Its rich information content, its flexibility in addressing a wide variety of samples, and its sound theoretical basis contribute to the popularity of XPS. Throughout the current work, XPS was used as a main analytical technique, and this chapter will  introduce the XPS method and describe its theory, spectral inteipretation and  instrumentation. The history of XPS traces back to Hertz's discovery in 1887 of photoelectron emission from metals exposed to ultraviolet light [60]. A quarter of a century later Einstein explained the phenomenon by introducing the concept of the photon, and he received the Nobel Prize for physics in 1921 in part for this work. The development of XPS as a practical method started with studies by Robinson and Rawlinson in 1914 of the energy distribution of photoelectrons produced by X-ray irradiated gold, but the foundations of the modern analytical method were established by Kai Siegbahn during the 1950's and 1960's [61]. He was rewarded for his contribution with the Nobel Prize for physics in 1981; Siegbahn also coined the term ESCA (electron spectroscopy for chemical analysis) to emphasize the fact that both photo and Auger electron emission peaks appear together in an XPS spectrum [62]. The basic XPS experiment involves irradiation of a sample surface with soft X-ray radiation (commonly either M g K  a  or A l K ) . a  The atoms comprising the surface emit  14  photoelectrons, after direct transfer of energy from the photon to (usually) core-level electrons. The kinetic energies of the emitted electrons are subsequently measured and counted. To a first approximation, the kinetic energy (E ) of a photoelectron is k  ^E  k  = hv - E  (2.1)  b  where hv is the photon energy and E is the electron's binding energy inside the sample. An b  illustration of this process is shown in Fig. 2.1, which also indicates the mechanisms by which the excited system can relax. The processes of X-ray fluorescence and Auger electron emission compete with each other, and their relative probabilities vary with the atomic number Z and quantum shell associated with the original vacancy. When the binding energy of the vacancy is 2 keV or less, Auger production is always the dominant process [63]. The kinetic energy of a photoelectron relates to the environment from which it originates, and the number of electrons emitted depends on the concentration of the relevant atoms in the sample. Thus both qualitative and quantitative information can be obtained from XPS, as discussed in the following sections.  2.2  Qualitative analysis The low-resolution spectrum from polyethyleneterephthalate (PET) film in Fig. 2.2(a)  (also called a survey scan) shows a series of peaks on a step-like background. The peaks can be grouped into three types: (i) peaks due to photoemission from the core levels (O ls and C ls), (ii) structure from the valence band (binding energy in range 0 to 30 eV), and (iii) peaks due to Auger electron emission shown as C(A) and 0(A). In general the structure in an XPS spectrum can be identified from their binding energies using standard tabulations [64,65]. After each photoemission event, there is a cumulative background signal associated with  15  (a) photoelectron emission photoelectron  h  v  de-excitation -7  \ -  (b)  (c)  X-ray emission  Auger electron emission lon^-  2,3  J  A/V\A/VV<  h v  K  Figure 2.1 Production of a photoelectron by irradiation with soft X-rays and two processes for de-excitation of the atomic core hole produced.  (a) Survey scan 40000 - i  O ls  c  C ls  H2  O(A)  1400  1200  1000  Valence band region  800  600  400  200  0  284  282  280  Binding Energy (eV)  (b) C l s G 3 -Q s-  GO  a K —> Tt*  shake up peak  294  292  290  288  286  Binding Energy (eV)  Figure 2.2 X P S spectra from polyethyleneterephthalate (PET) film showing: (a) low-resolution survey spectrum, and (b) C ls spectrum at higher resolution with allocation to component peaks.  photoelectrons that have lost energy due to inelastic collisions in the solid, while still having sufficient energy to escape from the solid. In general there is a continuum of electron energies ranging from the kinetic energy corresponding to Equation 2.1 to zero kinetic energy. The sharp peaks correspond to photoelectrons which have left the solid without experiencing any inelastic scattering. Commonly XPS spectra are plotted as a function of binding energy, rather than measured kinetic energy, because the former relates more directly to the chemical structure of the surface and it is independent the photon source used. More information about a sample is available from the high-resolution (or narrow scan) spectra. Figure 2.2(b) shows such a spectrum measured for the C Is region from PET. From the spectral shape, it is apparent that this spectrum has a composite structure, that is it is composed of a number of individual components. These components occur at slightly different binding energies, and these chemical shifts are attributed to atoms and groups in different environments, as identified in the figure. Three major peaks correspond to phenyl carbons, methylene carbons (O-CH2CH2-O) and the ester carbon (0=C-0), and these are listed in order of increasing binding energy. This order simply correlates with the number of neighboring electron-withdrawing O atoms.  A minor broad peak is noted at around 291.6 eV which  corresponds to Tt—>7t* shake-up structure, and it is associated with a configuration interaction effect which depends on promotion of valence electrons from an occupied K level to an unoccupied higher 7t* level. This feature is generally apparent in XPS spectra for systems with aromatic ring structures, and its presence can be used as a fingerprint for aromaticity in a sample. A number of other spectral features are commonly observed in XPS spectra. Doublet structure due to spin-orbit coupling may be observed for photoemission from states with non-  18  zero orbital angular momentum (i.e. p, d and f core levels). An example is given in Fig. 2.3(a) for the Zr 3d spectrum. The different components are assigned by the j values (i.e. 1/2, 3/2 for p orbitals, 3/2, 5/2 for d orbitals, etc.), and the corresponding areas are determined by the respective (2j+l) number of states. Accordingly, the  3dy  2  and 3d  5/2  components have areas in  the ratio of 2:3; the energy separation between the peaks is 2.4 eV in this example. Sometimes, the exiting photoelectron can couple with plasmon excitations. These excitations are associated with the collective motions of valence electrons in the solid, and they can lead to characteristic energy losses. An example of plasmon loss peaks is shown in Fig. 2.3(b) for photoemission from magnesium; such features are usually well developed for metals, and for deposition on surfaces their presence may help identify the extensive (rather than locally clustered) nature of the material involved.  2.3  Quantitative analysis  2.3.1  Inelastic scattering and sampling depth  Although the penetration depth of incident photons into a solid may be in the micron range, or more, the photoelectrons excited by this irradiation have a high probability to lose energy via inelastic scattering with the matrix, and thereby contribute to the background signal. Since we are most interested in photoelectrons that are just elastically scattered by the solid, the sampling depth in an XPS measurement is determined by the inelastic mean free path (IMFP or X). This property is defined as the average distance traveled by an electron in the solid before it undergoes some inelastic scattering. Empirical values of the IMFP have been provided by Seah and Dench [66], and Fig. 2.4 shows its dependence on energy using measurements for different materials [67]. In XPS, the kinetic energies of the photoelectrons are usually in the 100 to 1000  19  Figure 2.3 Extra structures in XPS spectra: (a) spin-orbit coupling for Zr 3d; and (b) plasmon loss structure (pi) associated with M g 2s and M g 2p spectra.  Kinetic Energy (eV)  Figure 2.4 Inelastic mean free path of electrons as a function of kinetic energy. Redrawn from ref. [67].  eV range, where IMFP values of 6 to 30 A are typical [67]. The low values for the IMFP involved make XPS a surface sensitive technique. Conventionally, the sampling depth in XPS is taken as 3A,, and that corresponds to the depth from which 95% of the elastically scattered signal is contributed.  2.3.2  Atomic concentration determination  The area under a photoemission peak is generally taken as proportional to the amount of the appropriate element within the sampling depth. Accordingly, by measuring peak areas and correcting for appropriate instrumental factors, the percentage of each element in the surface region can be estimated. The number of photoelectrons collected by the analyzer per unit time (i.e. the intensity of the photoemission peak) depends on the X-ray flux (F), the area of the sample from which the photoelectrons are collected (A), the instrumental transmission function (T), the photoelectron cross section (a), the number of atoms per unit volume (n), and the inelastic mean free path (A,). The intensity of the photoelectrons from an incremental thickness dx in the sample as illustrated in Fig. 2.5(a) is  di = F A T o n e ~  x / x  dx  (2.2)  Integration from x = 0 to x = ~ for a semi-infinite homogeneous sample gives the simple expression  I = F A T an  X  (2.3)  Since the parameters F, A and T are associated with the particular instrumental setting, and the cross section G is determined by the specific element and orbital, it is a common practice to group them into a sensitivity factor S, which is defined as FATfJ. These factors, derived  22'  (a)  overlayer  \ t  (  substrate  Figure 2.5 Notation for: (a) photoelectron intensity from a narrow band (dx) in a semi-infinite homogeneous sample, and (b) a sample with an oxide overlayer thickness t on top of a metallic substrate. ox  relatively to the F ls peak, are available for the MAX200 spectrometer used in this work, where the transmission function is corrected for the particular instrumental conditions used for each spectral measurement. The composition ratio for two elements A and B in a sample can then be expressed as  (2.4)  Tabulated values of A . and A  AB  are available [64], but for semi-quantitative work the ratio  A /AB A  is commonly taken as constant and equal to unity. On this basis, measured peak intensities are combined with appropriate sensitivity factors to estimate the elemental ratios within the depth probed. 2.3.3  Angular dependent measurements Reducing the angle between the axis of the energy analyzer and the plane of the sample  surface enhances signals from surface species compared with those from deeper down into the sample. This angle is called the take-off angle, 8, and it is illustrated in Fig. 2.6(a). For normal take-off angle (0 equals 90°), the effective sampling depth is 3A, while at smaller angles of emergence, the effective sampling depth reduces to 3 A s i n 0 . This effect is illustrated in Fig. 2.6(b), where A l 2p spectra from an aluminum surface that is covered by a thin oxide layer are compared for measurements with 0 equal to 90° and 30°. At the lower take-off angle, the signal contributed from the aluminum oxide is seen to be more pronounced. For a sample with an uniform oxide overlayer (thickness t ) on top of a substrate metallic material, shown in Fig. ox  2.5(b), Equation 2.2 integrates after correcting for angular effects to give  24  3  rt rj  t/3  rt  o  rt  C  CU  W  i  ©  *d  CO +3  w  rt  O u  , rv »H W g  >  rt  OH CN —  tu  >.  CX  •< -a <S ° c o  r-  aa oo  03  g  cxT  CD  .s &  r--  ? £  •a ra*  (sjiuirqjB) A j i s i i a j u l  d  O  5U  —  o d o H <u rt  -rt — ^ - N  <U  ed  X  o  ra CN  CD — ;  d  H o  u c  SO U  -—' >-. bD  CD  r-  rt  >tu  ra (U S d  +J  c!  on s  si T3  ^  "Hb d a ( s j j u i r q j B ) v|isii.->)U|  ^  fe  CD  CN  "O  C+H  s* s cu  eso  O  °  -5  rt >  (2.5)  Iox = F o n A T A , ( l - e o x  o x  o x  mox sine  (2.6)  for the intensities from the oxide overlayer (I ) and the metallic substrate (I ). ox  equations, X  ox  and X  m  m  In these  are inelastic mean free paths for photoelectrons originating in, and  travelling in, the overlayer and substrate respectively, whereas A ,  mox  is the inelastic mean free  path of photoelectrons from the substrate traveling in the overlayer. The ratio l x/Im increases 0  as the take-off angle (9) decreases, and Equations 2.5 and 2.6 provide a basis to interpret orientation effects and composition gradients for different species in the surface region of a sample [68]. In addition, Equation 2.6 can be modified to compare I  m  values at various take-  off angles against that at 90°  1 n [ I (9)/I (90°)] = (-1 / X m  m  m o x  ) (1 / sin9) + (t / X  mox  )  (2.7)  This allows the overlayer thickness (t ) to be estimated from the slope of ln[I (0)/I (9O°)] ox  versus l/sin0, provided information is available for  2.4  m  X  mox  m  .  Instrumentation A schematic diagram of the Leybold MAX200 facility used in this work is shown in  Fig. 2.7. The system has four chambers that are all interconnected, and a sample can be  26  27  transferred between them without exposure to the atmosphere.  The analysis chamber  accommodates the main components of the spectrometer, as well as an ion gun for sample sputtering and for ion scattering spectroscopy (ISS), an electron gun for measuring Auger spectra, an electron flood gun to help compensate for any sample charging, and an X-ray monochromator. The sample transfer chamber is equipped with an automatic transfer rod and a sample magazine. In situ sample preparations are performed in the preparation chambers. That labeled as the metal deposition chamber is used for low-pressure surface treatments such as metal deposition, sample heating and gas dosing.  The background pressure during these  treatments is typically in the 10" to IO" torr range. A metal evaporation source, a variable leak 6  8  valve (for gas introduction) and a quadruple mass spectrometer ( R G A ) for residual gas detection are included in this chamber. Higher-pressure treatments (e.g. the plasma treatment used in Chapter 4) are performed in the plasma chamber.  2.4.1  Ultrahigh vacuum (UHV) There are several reasons why vacuum is required for the X P S experiment. First, the  emitted photoelectrons must be able to travel from the sample through the analyzer to the detector without colliding with gas molecules. Second, some components including the X-ray source and channel plates require vacuum conditions for effective operation. Third, the surface composition of the sample under investigation should not change during an X P S experiment. According to the kinetic theory of gases [70], at room temperature and a pressure of IO" torr, a 9  surface can be covered by an adsorbed monolayer, of for example C O , in about one hour assuming a sticking probability of unity. Only high vacuum (~10" torr) is required for the first s  two criteria, but better vacuum conditions are necessary to avoid surface contamination. In  28  general, vacuum conditions around the 10" torr range are required for very high level surface 10  cleanliness, although some types of sample are less sensitive to contamination than others. A schematic diagram of the pumping system for the MAX200 facility is shown in Fig. 2.8.  Ion pumps are used for the X-ray sources, while a combination of rotary pumps and  turbomolecular pumps are used for the main chambers. The base pressures are achieved after baking at around 120°C for 12 hours or more.  After the baking, typical pressures in the  analysis chamber and the metal deposition chamber are around 6x10"'° torr, and those of t h e transfer chamber and the plasma chamber are around 2xl0" torr. 8  Samples can be introduced to the analysis chamber either from the transfer chamber or from the metal deposition chamber. The sample transferring system in the MAX200 facility is shown in Fig. 2.9(a).  Samples that are prepared outside of the facility are mounted on a  standard sample holder and locked onto a sample magazine, which is then introduced to t h e transfer chamber prior to the initial pump down. Once the transfer chamber pressure is reduced into the 1(X torr range, one sample holder is transferred to the analysis chamber and locked on 8  the manipulator (PTM 60) dock. This manipulator can give five degrees of movement to t h e sample, three for linear motions (X, Y and Z) and two for rotational motions (T and R) as indicated in Fig. 2.9(b). The transfer system is computer controlled to enable proper position settings and to perform angular dependent measurements. Samples that are prepared in situ in the metal deposition chamber, or in the plasma chamber, can be manually transferred to t h e analysis chamber using transfer rods.  2.4.2  X-ray sources A schematic diagram of the dual anode X-ray source in the MAX200 system is shown  in Fig. 2.10(a). Films of M g and A l (~ 10 um thick) have been separately deposited on each  29  Figure 2.8 Pumping system for the MAX200 facility.  30  Transfer rod #1 Plasma chamber  Metal deposition chamber  Transfer rod #2  Pedestal  Transfer chamber Transfer rod  Analytical chamber  Sample magazine Manipulator Hindge door (a)  R  (b) r V  V y  Figure 2.9 Sample transfer system in the MAX200 facility: (a) between different chambers; and (b) five degrees of movement possible for sample on the manipulator in analysis chamber. Redrawn from ref. [69].  Figure 2.10 Schematic diagrams of: (a) the Mg/Al dual anode X-ray source; and (b) energy distribution of photons produced, showing the K and Ko lines and the high-energy Bremsstrahlung radiation. Ref. [69].  a  side of the source, and the external circuit can switch from one connecting filament to the other. The anode is operated with a potential of up to 15 kV, and electrons accelerated from the heated filament hit only one of the anode faces. The radiation generated passes out through an aperture covered with a thin A l (~2 urn) window. This window prevents stray electrons, radiation and contamination reaching the sample from the source, which provides either M g K  a  (1253.6 eV) or A l K (1486.6 eV) radiation according to the filament switch setting. Typically, a  the X-ray source is operated at 10 kV and 20 mA; the heat produced during operation is removed by an internal water circulation system and by having copper construction for the body of the source. Figure 2.10(b) shows the energy distribution of radiation produced by an Al source; the characteristic K  a  (K i a  i  and  K 3,4) H  and Kp lines are observed on a broad  background, which is called Bremsstrahlung radiation [71]. The best resolution is obtained by just using the K i , a  2  line, and for this an X-ray monochromator must be used.  The  monochromator consists of a bent quartz crystal (on the surface of a Rowland circle) through which the raw radiation is monochromatized by diffraction [16]. The resulting beam has a narrower energy spread and a lower background because the Bremsstrahlung radiation is removed as well as the  and Kp lines. The overall flux is reduced by several orders of  magnitude compared with that from a nonmonochomatized source, but the monochromatized source was not used in this work.  2.4.3  Energy analyzer (EA200) Figure 2.11 represents a schematic diagram for the EA200 energy analyzer provided in  the MAX200 system.  Its three main components are: (i) the input lens system, (ii) the  concentric hemispherical analyzer (CHA) and (iii) the multi-channel detector (MCP). These will now be introduced and discussed briefly in turn.  Concentric hemispherical analyzer (CHA)  Second lens stage  Multichannel detector (MCP)  A3  T T T T  I— A 2  A1  First lens stage  sample  Figure 2.11 Schematic diagram for the concentric hemispherical analyzer and input lens system in the MAX200 system. Redrawn from ref. [69].  The input lens system transfers an electron image of the analyzed area on the sample to the analyzer. Two transferring stages are involved. The first lens stage, including the variable angular aperture A l and image aperture A2, controls the analysis area (spot size) and acceptance angle (Q.) for the input electron image. The second stage accomplishes a focussing of the electron image onto the slit S1, a retardation of the electron energy to a particular pass energy, and A3 lens controls the entrance angle (a) for the electrons to enter the analyzer. The C H A is constructed from two concentric hemispherical electrodes (inner radius R i and outer radius R ) , to which a deflecting potential AV is applied [16]. Electrons on the 2  central circular trajectory reach the planar detector S2 at the nominal radius position R , where 0  R = (R,+R )/2. This requires the kinetic energy E , inside the analyzer (i.e. the pass energy) to 0  2  0  satisfy  eAV = E ( R / R , - R , / R ) 0  2  (2.8)  2  With entrance angle of a, the analyzer resolution A E i is given by ana  AE  a n a l  /E  = (S,+S )/4R + 0C /4 2  0  2  0  (2.9)  where Si and S are the slit widths for the entrance and exit of the analyzer respectively. Since 2  S|, S , Ro, and a are limited by the spectrometer construction, the resolution of the analyzer 2  necessarily varies with E . 0  The overall experimental resolution, as expressed by an observed peak width ( A E  peak  ),  defined as the full width at half-maximum (FWHM), has contributions from the inherent line width of the atomic level involved (AEi ), and from the natural line width of the X-ray source ine  (AE ourceX as well as the analyzer resolution. The observed peak width satisfies S  35  AEpeak = ( A E  anal  + AE ii e + A E 2  n  2  source  y  (2.io)  u  provided all contributions have the Gaussian shape [16]. The EA200 is operated in the constant resolution mode to ensure a constant analyzer resolution at all energies in a spectrum. The entering electrons are first retarded by the lens system to a fixed pass energy (E ), which is set for an entire spectral measurement. 0  The  deflecting potential on the analyzer (AV) is pre-set, according to Equation 2.8, for the particular pass energy. The ramping of the retarding field voltage during the pre-retardation process does the actual scanning of the kinetic energy to be measured. Equation 2.9 indicates that the lower the pass energy, the better the analyzer resolution, but the resulting signal intensity drops as well.  Therefore an optimal balance is required between resolution and intensity, and  accordingly an appropriate pass energy is chosen for each measurement. In the present work, a pass energy of 192 eV was used for the low-resolution survey scans, and the values 24, 48 or 96 eV were used at different times for higher-resolution measurements. Use of multi-channel plate (MCP) detection for the simultaneous recording of an energy band around E is possible because of the presence of electrons that enter slit SI at an angle 5a 0  to the tangential direction. These electrons are able to pass through the analyzer with slightly different (non-circular) trajectories, and accordingly they reach the detector away from S2. The EA200 detector is constructed from two MCPs assembled in a back-to-back configuration to form a chevron array. Each M C P is an array of 18 capillary-type microchannels, which act as individual electron multipliers. The plates are oriented so that the channel angles of the two plates are in opposition; this suppresses feedback by trapping ions at the interfaces between the two plates. The voltage across the plate is set to allow count rates at above It) s" . 7  1  36  Figure 2.12 illustrates that the kinetic energy of a photoelectron measured by the spectrometer (E' ) is referenced to the spectrometer's vacuum level, while the binding energy k  of the electron inside the sample (E ) is referenced to the Fermi energy of the sample. For a b  conducting sample in electrical contact with the spectrometer, so the Fermi energies are equal, the energy balance requires E'  k  = hv - E - W 5  (2.11)  s p  and this represents a modification of Equation 2.1. The spectrometer work function (W ) sp  remains constant while the analyzer is held under U H V . If the pressure of the spectrometer is raised above the U H V range different species can adsorb onto components in the analyzer, and that may change W . Then a re-calibration with a standard gold sample is necessary; the basic sp  reference for this work is to set the Au 4f  2.4.4  Data  7/2  measured binding energy to 84.0 eV.  processing  Prior to detailed spectral interpretation, some massaging of the raw data is required. The standard procedures of data processing in XPS are introduced in this section. For insulating samples, such as polymers and metallic oxides, a positive charge can build up during an XPS measurement due to the loss of electrons by photoemission. Electrons from the spectrometer can replace those electrons if there is sufficient electrical contact with the sample, and this is the case for conducting and semi-conducting samples. However, with insulating samples the exiting photoelectrons require some extra energy to overcome the barrier resulting from the net positive charge on the sample. Without correction this will result in a lower measured kinetic energy, and hence a higher than expected binding energy. When monochromatized X-ray sources are used, use of the electron flood gun is required during a  37  Vacuum level Vacuum level  E  \hv sp  Fermi level  Sample  hv s p  spectrometer work function  W  s  sample work function  k  Spectrometer  photon energy  W  E  Fermi level  photoelectron kinetic energy with respect to vacuum level of sample  E ,  kinetic energy of photoelectron measured by spectrometer  E  binding energy of electron in solid with respect to the Fermi level  k  b  Figure 2.12 Energy levels for binding energy measurements.  38  spectral measurement to compensate for the loss of electrons from the sample surface. When non-monochromatized X-ray sources are used, as for this work, and any charging is not too severe, there is some self-compensation resulting from the Bremsstrahlung radiation.  That  generally ensures sufficient electrons are present in the vacuum so that charges on the sample can be approximately neutralized, and the energy scale may then be calibrated to a known binding energy reference value.  This is often done by fixing the binding energy of the  adventitious hydrocarbon peak in the C ls spectrum to 285.0 e V , and that approach was used in this work. Raw spectra must be background subtracted prior to undertaking a curve fitting procedure, and there are several methods available for this [72,73]. In this work, the non-linear approach proposed by Shirley [72] is used, and that method is illustrated in F i g . 2.13(a). This approach is based on the assumption that the dominant contribution to the background comes from inelastically scattered photoelectrons such that, at any point in an uncorrected spectrum, the background signal is proportional to the number of electrons elastically scattered at higher kinetic energies.  The subtraction is applied to an energy range E i < E < E , for which the 2  operator chooses the lower and upper energies (Ei and E respectively). Shirley proposed the 2  following iterative algorithm to correct for the inelastic contributions E,  N'  k + 1  (E) = N(E) - N ( E ) - C J N ' k(E)dE  (2.12)  2  E  where N ( E ) is the measured count rate and the N'k(E) identify count rates after subtraction of background contributions (the latter have subscripts, for example k indicates the background corrected value according to the kth iteration). The reference background level is provided by N ( E ) ; C is a constant which is fixed by the requirement that N ' ( E | ) equals 0. The process 2  k  39  ?  Kinetic Energy —>  Figure 2.13 Illustration of Shirley non-linear background subtraction: (a) applied to a raw C ls spectrum; and (b) result of curve fitting to two components after correcting for the background. Measured data are shown as points and the solid line represents the synthesized data.  starts with N'i(E) = 0 and continues until N ' i ~ N ' ; it normally converges after three or four k +  k  iterations. After subtracting the background, the area associated with the elastically scattered contributions in a spectral feature can be determined by integration. As noted in Section 2.2, features in a high-resolution XPS spectrum for a particular element are often composed of a number of overlapping components which are chemically shifted due to the inequivalent atoms having different chemical environments.  A curve  synthesis is then required to define those individual components with regard to peak position, intensity and peak width (FWHM). The objective is to find the combination of component functions that sum to give a close representation of the background-corrected experimental spectrum. The most effective approach is to use the mixed Gaussian/Lorentzian form [16]  f(E)=  ^ , exp{(l-M)[3 [ln2(E-E ) ]} [l + M ( E - E ) / p ] h 6 i g h t 2  2  2  2  ()  (2.13)  0  where E is the individual component peak energy, the parameter [3 fixes the F W H M , and M is 0  a mixing ratio (1 for pure Lorentzian, 0 for pure Gaussian). In practice, a curve synthesis starts by making a reasonable estimation of these parameters based on chemical knowledge of the system, and an optimization procedure is then carried out to obtain the best fit. Figure 2.13(b) shows an example of such a curve synthesis for a C ls spectrum measured from a polymer sample. After the background subtraction, the major peak at 285.0 eV due to C - H and C-C bonding and the minor peak at 286.5 eV corresponding to C-O bonding are fitted. The curve synthesis was done iteratively by optimizing the fit between the measured curve (dotted line) and the optimized sum of the component functions (solid line). The quality of fit is evaluated by the least-square function  41  1 =  (Y,  N free  i = l  (2.14) Y,m.i  where the sum is over the N data points in the spectral region to be fitted, Y . j is the measured m  count rate at the /th data point, Y f j is the corresponding value of the fitting envelope, and N f equals N - N f , (Nf t  lt  is the number of parameter to be fitted).  r e e  Parameters for the various  components in the curve fitting are optimized by the condition that % is minimized, although a visual comparison is also made between a measured curve and the synthesized envelope to ensure there is visually a close agreement. If a given number of component functions do not converge to a reasonable correspondence, more functions can be added. However, this should only be done according to basic chemical and physical principles.  42  Chapter 3 Characterization of some metal-polymer systems  3.1  General introduction Metal-polymer interfaces formed by thermally depositing metals in vacuum onto  polymer surfaces have technological relevance, and yet there are many issues that are not well understood. While the metal may be characterized by an ordered atomic structure with close packing density, the polymer usually has a disordered, loose, and interwound molecular structure. The predictions of bonding characteristics for such a structural combination can be challenging, and depending on the combination of materials forming the interface the nature of the bonding can be significantly different. For example, contrasting results have been reported for the two separate interfaces formed by Cu and Cr with polyimide, and these systems have extensive applications in the field of microelectronic packaging. Observations obtained from XPS, UPS and transmission electron microscopy (TEM) all consistently suggest the presence of much weaker interactions in the interface formed by the former metal compared to the latter [74]. Indeed in general the nature of the bonding at the metal-polymer interface can vary from strong, directional covalent bonds at one extreme to the much weaker van der Waals interactions at the other extreme.  However, generally it is believed that knowledge of the  molecular-level composition and bonding arrangements at these interfaces should help understand their macroscopic adhesive, electrical and chemical properties. One particular example in which metal-polymer interfaces have essential roles is for the polymer-based light-emitting diodes (PLED) introduced in Section 1.3.  Although the  43  brightness of a working P L E D can reach a level almost comparable to that of the traditional inorganic semiconductor L E D , a major obstacle to the P L E D commercialization remains in the limited lifetime and some operational instability [75-77].  But even at this stage of  development, the principles associated with the metal-polymer interaction are still not well understood, and part of the work to be undertaken in Section 3.2 concerns comparisons between different metals on the particular polymer, polyethyleneterephthalate (PET).  The  structures of the polymers to be investigated in this thesis are shown in Fig. 3.1. During P L E D operation, the metallic cathode and polymer interfacial region may be damaged by ingress of water or oxygen from the atmosphere, and by local heating.  One  approach to prevent the former problem is to encapsulate the device by glass, and so improve device lifetime [78,79]. However the use of a glass cover unfavorably hinders the device's compactness and flexibility. The study presented in Section 3.3 concerns this issue, but takes a different approach by applying a second protective layer on a metal cathode. Although, both oxygen ingress and local heating are suspected as causes for cathode-polymer interface degradation, the amount of study associated with these issues is very limited compared with the reports  that have emphasized other  issues  such as material synthesis  and general  characterization. Section 3.4 initiates some consideration of the former issues in relation to the Al-PPV interface, which is used in an actual PLED device.  3.2  XPS investigations of Al-PET and Zr-PET interfaces  3.2.1  Introduction  The Al-PET system has appeared as a basic reference system for the metallization of polymers. PET (Fig. 3.1(a)) has a backbone composed of a regular arrangement of - C H 6  4  44  <">  CH  /  C  3  = 0  CH CH CH CH, 2  2  2  PBMA  Figure 3.1 Monomer units for polymers studied in this work: (a) polyethyleneterephthalate (PET), (b) polystyrene (PS), (c) polyphenylvinylene (PPV), and (d) polybutylmethacrylate (PBMA).  (disubstituted benzene), -COO- (ester link) and - C H C H - (dimethylene) groups. 2  2  Several  analytical techniques, including XPS, SIMS and HREELS, have been used for characterization of Al-PET interfaces, but interpretations by different authors have shown some differences [80]. For example, Bou et al. [81] using XPS concluded that A l atoms react first with carbonyl groups, to form local Al-O-C complexes, while later there is an interaction with the benzene rings and formation of A l - C bonds, although no signature for this was presented in their C Is 7t—>7t*  shake-up feature. By contrast, a HREELS study by Novis et al. [82] pointed to the A l  atoms reacting preferentially with ester groups to form COO A l bonding. More recently, +  Calderone et al. [83], using XPS and quantum-chemical calculations, concluded that the initial Al reaction is with the ester group to form Al-O and Al-C bonds, and that this may be followed by some interaction between A l and the benzene rings after the accessible ester groups have been consumed. A previous XPS study from this laboratory for the Mg/PET system showed no evidence for an interaction between M g and the benzene rings, although it was concluded that M g first binds to >C=0 in a "side-wise" orientation [84], somewhat analogously to the bonding reported between A l and polyvinyl alcohol by Akhter et al. [85] and Stoyanov et al. [86]. This issue is discussed further in the following sections by comparing metal-polymer interfaces formed by different combinations of metal and polymer. In an attempt to clarify some issues just mentioned, it seemed important to undertake a new XPS study for the Al/PET system, and we used the A l metallization of polystyrene (PS) as a reference since its molecular structure has benzene-type rings but not the ester groups (with O functionalities) present in PET. This work is presented on a comparative basis in order to highlight trends that may help to establish principles for interfacial metal-polymer bonding.  46  Specifically, our conclusions for the interface formed by PET with A l , a main group metal in Group III, are compared with new observations for the Zr/PET interface, as an example of metallization by a Group IV transition metal, and with those from the previous study for Mg/PET, where magnesium is taken as representative of a Group II (alkaline earth) main group metal. The analyses use conventional curve fitting although, especially for the Zr system, guidance is provided by difference spectra [87].  Angle dependent measurements help  recognize orientation effects, and assessments are also made of the stability of these systems in oxygen, particularly given that the ingress and reactivity of this element at metal-polymer interfaces can have a negative effect on the performance of a PLED. 3.2.2  Experimental XPS analyses were carried out in the MAX200 facility described in Chapter 2. The  non-monochromatized M g K  a  source (1253.6 eV) was operated at 10 kV, 20 mA, and  measurements of survey scan spectra were made using a pass energy of 192 eV. Higher resolution measurements for the C ls, O ls, Zr 3d and A l 2p spectra used a 24 eV pass energy. Photoelectrons were collected in the direction normal to the sample, but sometimes additional measurements were made for the 30° exit direction; the analysis area was 2x4 mm . Binding 2  energies were referenced to the C ls binding energy associated with the phenyl carbons at 285.0 eV. Prior to use, each PET film sample (8 |im thick) was ultrasonically cleaned with methanol and dried under N for 0.5 h. The PS film was made by spin casting PS dissolved in toluene on to 2  an A l panel, and drying in N . After mounting on sample holders, the films were held in the entry 2  lock for several hours before transferring into the U H V system. The A l and Zr evaporation sources (Fig. 3.2(a)), enclosed in a quartz tube and covered by a shutter to protect other nearby  47  Copper feedthrough AI or Zr wire (a)  Sample  Tungsten wire  Quartz tube cover Power supply  (b) i  •  Mg turnings Quartz tube  Figure 3.2 Schematic diagrams of evaporation sources: (a) Zr or A l , and (b) M g .  48  components were installed in the metal deposition chamber. These sources were made using 0.5 mm diameter Al wire (99.999% purity) or 0.25 mm diameter Zr wire (99.7% purity) wrapped around tungsten coils (wire diameter 0.5 mm); the latter were connected to copper feed-throughs for supplying the heating current. To minimize contamination in the deposited films, the sources were outgassed for about 5 h at the operating temperatures (i.e. 500°C for A l , 1400°C for Zr) with the shutter in the closed position. The thicknesses of the metal overlayers on the Al/PET, Zr/PET and Al/PS samples studied in this work are estimated at 10, 4 and 5 A respectively from the attenuation of the C ls integrated intensities using Equation 2.5. These thickness values are deduced assuming that the mean free paths for electrons (energy just less than 1 keV) equal 25 A for the deposited layers [88], which must also be assumed homogeneous. Evidence below indicates that the initial depositions are more heterogeneous, and accordingly these thicknesses can give no more than a rough comparative guide. Figure 3.3 illustrates the experimental steps involved in this work. 3.2.3  Results and discussion  3.2.3.1 Zr/PET Four distinct components are now recognized in the C ls spectrum (Fig. 3.4(a)), and they are assigned as phenyl carbon atoms (binding energy 285.0 eV, component C l ) , methylene carbon atoms (286.6 eV, C2), ester carbon atoms (289.0 eV, C3), and the weak peak (below 292 eV) resulting from the 7t—>7t* shake-up transition. In all subsequent discussions the components C l , C2 and C3 are given constant widths (FWHM 1.4 eV). The O ls spectrum (Fig. 3.5(a)) was fitted with two components, one at 531.9 eV binding energy for the (doublybonded) carbonyl oxygen (referred to as component O l ) and the other at 533.5 eV for the (singly-bonded) ester oxygen (02). These peak areas are almost equal, as expected for the  49  PET  PET  Zr  PS  Al  Al  Al/PET  Al/PS  \(  Zr/PET  Air  >  Air-Zr/PET  Figure 3.3  0 - Al/PET 2  Schematic diagram of steps for experiments in Section 3.2.  294  292  ^290  1>88  ^286  ^284  ^82  280  B i n d i n g E n e r g y (eV)  Figure 3.4 C ls spectra: (a) clean PET film with components C l (phenyl), C2 (methylene), C3 (ester); (b) comparison for PET film (dotted) and Zr/PET (solid); (c) difference spectrum Zr/PET minus clean PET as described in text; and (d) curve fitted spectrum for Zr/PET.  Figure 3.4, continued.  es (  M  Zr/PET  A  ,  &  i  \  538 537~~536 535 534 533 532 531 530 529 528 B i n d i n g E n e r g y (eV)  Figure 3.5 O l s spectra: (a) clean PET film with components 01 (carbonyl), 02 (ester); (b) comparison for PET film (dotted) and Zr/PET (solid); (c) difference spectrum Zr/PET minus clean PET as described in text; (d) curve fitted spectrum for Zr/PET.  B i n d i n g E n e r g y (eV)  Figure 3.5, continued.  PET stoichiometry, and angle-dependent measurements for both C ls and O ls are essentially independent of the take-off angle. The latter observation that the surface composition appears constant with angle suggests that any contamination present is also not large; further, the essentially unchanged O/C ratio after the Zr deposition suggests that the addition of impurities to the sample, from the evaporation process, is relatively small. The Zr deposition gives small changes in the C ls spectra (Fig. 3.4(b)); specifically there is some line broadening with a general decrease in intensity (as expected from an overlayer attenuation effect), although the 7t—>TC*  shake-up transition is hardly affected. The latter observation is consistent, at least for  this level of Zr coverage, with there being no direct Zr-C bonding to the benzene ring. Changes in the O ls spectrum (Fig. 3.5(b)) are more marked after the Zr deposition, especially for the attenuation seen in the ester O peak. Although changes in the C ls and O ls spectra on adding Zr are small, an attempt is made at quantifying them by taking difference spectra, a technique that has been shown to be useful to enhance small spectral features [16].  Prior to undertaking the mathematical  subtraction of one spectrum from another, care is needed with regard to correct alignment and normalization of the spectra in order to avoid spurious results [89]. The approach taken in the present work follows an argument which notes that the metallization of an organic polymer surface can change photoemission spectra in two main ways. First, there will be attenuation from the added overlayer; second, there will be changes in the details of the spectra as a result of some metal-polymer interaction.  The basic philosophy is that these interactions will  especially modify spectra by introducing interfacial components at lower binding energies, compared with the corresponding components in the clean polymer situation, and this follows since the metal will preferentially donate electrons to the polymer. When taking a difference  55  spectrum, the aim is to have a common reference component which is especially associated with the substrate (i.e. PET in this case) and is mainly just subject to the attenuation effect. In the present context, this assumption can be satisfied by using as reference a well-defined component at the high binding energy side of a spectrum. For C ls, the main part of the reference component at 289.0 eV (ester C) is assumed to come from the PET substrate, and a difference spectrum is formed by normalizing this peak height in the PET spectrum to that in the Zr/PET spectrum, and subtracting the former spectrum from the latter.  Insofar as the reference peak is subject just to attenuation, the difference  spectrum formed in Fig. 3.4(c) should simply identify changes associated with C atoms at the Zr/PET interface. Indeed this difference spectrum suggests that all modified structure appears as a simple perturbation on the original spectrum for clean PET. Thus no structure is apparent which can be allocated to new bonding situations, but the phenyl C, the methylene C and the ester C are indicated to have new small components (identified in Fig. 3.4(c) as C l ' , C2' and C3' respectively), shifted by about -0.6 eV, -0.7 eV and -0.9 eV respectively from their values in clean PET, while these interfacial components still approximate the 3:1:1 ratio for the original PET stoichiometry. Tests on the re-synthesized direct spectra (see next paragraph) suggest that the uncertainties on these values are of the order of ±0.1 eV. In any event, these negative shifts are fully consistent with electron density being donated to the interfacial PET region from the metal. A similar effect is apparent when the corresponding approach is applied to the O ls spectra, using the singly-bonded O peak at 533.5 eV as the reference component, although in this case the differences between clean PET and Zr/PET turn out to be greater than for C ls. Again the changes following the addition of Zr can be seen as perturbations on the spectrum for clean PET, and the corresponding difference spectrum in Fig. 3.5(c) shows that  56  new interfacial components, identified as 0 1 ' and 02', associated with carbonyl and singlybonded O atoms are shifted by about -0.6 eV and -0.9 eV respectively from the situation for clean PET (these new components also approximately maintain the 1:1 ratio for PET alone). Synthesized C ls and O ls spectra for Zr/PET are shown in Figs. 3.4(d) and 3.5(d) respectively. These are obtained by optimizing the area of each reference peak, while keeping all main components in the same relative proportions they have for clean PET (with constant widths) and mixing in components at the positions indicated in the difference spectra. Through this optimization process for Zr/PET, the background was kept fixed as for clean PET. Compared with the PET substrate, the interfacial components in Zr/PET are indicated to have areas which correspond to 4% in C ls and 8% in O ls, and this fitting is consistent with the concept that the interaction with Zr occurs primarily through O atoms at the PET interface. The Zr 3d spectrum after deposition (Fig. 3.6(a)) shows a line shape which fits the 2:3 ratio expected for the 3/2, 5/2 doublet; this appears to correspond to a single component, and the binding energy for the 3d  5/2  component at 182.6 eV is much closer to the value (182.9 eV)  identified for Z r 0 rather than to that reported for zirconium carbide (180.6 eV) [90]. This 2  again can be taken to suggest that the Zr is especially interacting with O atoms, rather than C, at the PET interface. The Zr/PET film was subsequently exposed to air for 1.5 h and re-analyzed with XPS at the previous analysis position. The O ls spectrum shows a gain in the O component with binding energy 532.0 eV, while that for Zr 3d (Fig. 3.6(b)) shows the deposited Zr is oxidized further to establish a chemical state which, like Z r 0 , has the Zr 3ds/ component at a binding 2  2  energy of 182.9 eV. The C ls spectrum indicates no major differences from the situation prior to the air exposure, except that there is some gain in a component at around 285.0 eV  5 7  (a)  3D  3d  J  3/2  1  !  5/2  AT \—  190  r  \  1  1  1  |  '  i  1  \  1  l  I  189 188 187 186 185 184 183 182 181 180  Binding Energy (eV)  3.6 Zr 3d spectra: (a) Zr/PET film, and (b) Air-Zr/PET.  associated with air-borne carbon contamination. In summary, the air exposure does not appear to change the Zr/PET interfacial region significantly, although there is some further oxidation of Zr, plausibly for uppermost metal atoms that are only partially interacting with the polymer surface. 3.2.3.2 Al/PS  XPS observations were made for the Al/PS system in order to provide a reference for subsequent studies reported for Al/PET in the next section. Oxygen is not detectable from the PS film prior to the A l deposition, although afterwards an O ls spectrum (Fig. 3.7(a)) shows a main peak at 532.5 eV, which is consistent with the presence of some metallic oxide. In addition, a small higher-energy component (at around 534 eV) suggests the presence of Al(OH) [91], and hence the likelihood that the deposition occurs alongside some reactivity x  with ambient water.  The A l 2p spectrum (Fig. 3.7(b)) shows components associated with  metallic (binding energy 72.7 eV) and oxide (74.8 eV) bonding, although the former is much the larger. Two further points can be made in relation to the spectral changes occurring with the deposition: (i) comparison of C Is spectra before and after the addition of A l shows (aside from some attenuation) no change in structure which could be associated with either C-O or CAl bond formation, nor indeed are any changes detected for the n—>7t* shake-up peak; (ii) survey spectra taken in the A l 2s and A l 2p regions still show the presence of the metallic A l plasmon loss structure. It appears for this system that any direct chemical interaction between the A l and the PS surface is small, although it is clear that A l is partially oxidized during the deposition process.  59  Figure 3.7 Spectra for Al/PS film: (a) 0 ls showing a component at 532.5 eV due to A l oxide and a component at 534.0 eV consistent with Al(OH) ; (b) A l 2p showing metallic component at 72.7 eV and oxide component at 74.8 eV. x  60  3.2.3.3 Al/PET The C ls and O ls spectra before and after the A l deposition are compared in Fig. 3.8((a),(c)) respectively.  A normalization has been applied to each spectrum after the  deposition in order to visualize better the changes induced by the metal. It is clear for the conditions used for Al/PET that the spectra are much more strongly modified than is the case for Zr/PET reported above. Figure 3.8(b) shows the curve-fitted C ls spectrum for the PET. surface after depositing the A l , and Table 3.1 reports the relative peak areas for different C ls components in the PET and Al/PET films.  The changes in the C ls spectrum after the A l  deposition are summarized as follows: (i) the total intensity is attenuated by about 7%; (ii) the component associated with the ester group (C3) is reduced relatively more than those for the other components; (iii) the  Tt—>7t*  shake-up peak shifts to higher binding energy (by ~1 eV)  with an increase in intensity; and (iv) two new components are formed at about 283.9 eV (designated C4) and 282.6 eV (C5). The components C4 and C5 at the lower binding energies result from the Al-PET interfacial interaction, although no comparable bonding structure was seen for the Zr/PET system. In any event, following assignments reported in other comparable studies [81,92,93], the first of these new components for Al/PET is interpreted as arising from the Al-ester group interaction (i.e. C4), while the second results from interaction of the metal with the phenyl group (C5). The results in Table 3.1 suggest that the A l is deposited somewhat unevenly around the different groups on the PET surface. Following the interpretations just given, and relating peak areas to relative compositions, the photoelectrons emerging at 90° from the surface appear to be detecting relatively less phenyl and methylene groups in the Al/PET sample, compared with clean PET (the relative loss is close to 20% for both groups). By contrast, the ester groups,  61  294  292  290  288  286  284  282  280  B i n d i n g E n e r g y (eV)  Figure 3.8 Spectra for Al/PET film: (a) comparison of C ls for clean PET (dotted) and Al/PET (solid); (b) curve fitted C ls spectrum for Al/PET showing new components C4 (283.9 eV) and C5 (282.6 eV); (c) comparison of O ls for clean PET (dotted) and Al/PET (solid); (d) curve fitted O ls spectrum for Al/PET showing three components at 531.9, 533.5 and 530.8 eV.  I — I — I — ' — I — ' — 1 — ' — 1 — ' — I — ' — I — '  I  '  I  1  I  1  I  538 537 536 535 534 533 532 531 530 529 528  B i n d i n g E n e r g y (eV)  Figure 3.8, continued.  63  Table 3.1 C l s binding energies (in eV) and percentages of the different carbon components indicated for the initial PET film (measurements made for 90° exit angle), after deposition of A l (90° and 30° exit angles), and for Al-PET system after exposure to oxygen (90° and 30° exit angles).  Cl  C2  C3  C4  C5  C6  285.0  286.6  289.0  283.9  282.6  284.3  Initial PET  63.0  19.2  17.8  -  -  -  Al/PET (90°)  46.1  15.1  10.8  22.8  5.2  -  Al/PET (30°)  31.3  10.7  4.5  48.7  4.8  -  0 - A l / P E T (90°)  51.0  11.6  5.9  -  -  .31.5  O2-AI/PET (30°)  49.9  9.3  3.2  -  -  37.6  Notation Binding energy (± 0.2 eV)  2  * These components are interpreted: C l phenyl C; C2 methylene C; C3 ester C; C4 ester C bonded to A l ; C5 phenyl C bonded to A l ; C6 oxidized Al-PET complex.  including those that have reacted with Al (i.e. contribute to C4) appear increased by about 90%. The reacted ester groups therefore appear to be oriented more to the surface, compared with methylene and phenyl which presumably are covered more effectively by the A l , and this trend is emphasized by changing the photoelectron collection to the 30° exit angle. This gives a marked increase in the relative involvement by the C4 component, although the small C5 component has only a relatively weak angular dependence.  Of the carbon components  associated with the original PET, that for C3 appears to reduce most at the lower exit angle, and that is consistent with Al interacting selectively with ester groups to form new chemical bonds. This situation contrasts with that for the Al/PS system where the A l appears to cover the polymer surface rather evenly, and without evidence for new chemical bond formation. Figure 3.8(d) shows the curve-fitted O ls spectrum after deposition of Al on to the PET. In contrast to the situation in Fig. 3.5(b) for Zr/PET, the component corresponding to the single-bonded O in the ester group decreases markedly, while a new component is formed at 530.8 eV, and the component at 531.9 eV is increased. These changes are similar to those observed for the Mg/PET system [84]. An interpretation is that Al interacts with the carbonyl group by breaking the TC bond and forming direct bonds to both C and O atoms.  This is  expected to shift both the carbonyl O and ester O to lower binding energies. Plausibly the first shift generates the new component at 530.8 eV, while the second results in an overlap with the substrate carbonyl at 531.9 eV. Unlike the observation for the Al/PS system, the survey scan spectrum for Al/PET gives no evidence for the A l plasmon loss structure associated with the A l 2s and Al 2p peaks. The second case has a higher A l coverage, but M . Bou et al. [81] also reported a lack of plasmon loss peaks for the initial stage of deposition. The higher-resolution Al 2p spectrum (Fig. 3.9(a);  65  Binding Energy(eV)  Figure 3.9 Spectra for 02-treated Al/PET sample: (a) comparison of A l 2p prior to treatment (dotted) and after (solid); (b) C ls showing components C l , C2, C3 and the new broad component C6.  66  dotted line) shows two resolved components. One is centered at around 72.0 eV (referred to as the A l component), a value less than expected for the pure metal form (e.g. 72.7 eV reported m  above for the Al/PS system). This difference may result from some cluster formation at the PET surface or from bonding to C. The second component (Al ) is located at 74.1 eV and ox  appears related to an Al-O species. The Al/PET sample just described was exposed to oxygen (at ~ l x l 0" torr) for 4 min in 7  the, preparation chamber, and then re-introduced to the analytical chamber and re-analyzed at the previous analysis position. A l 2p spectra measured before and after the 0 exposure are 2  shown in Fig. 3.9(a). This exposure resulted in one dominant peak at 74.9 eV in the spectrum, and it appears that essentially all the A l is converted to oxide.  The C ls spectrum after  oxidation is shown in Fig. 3.9(b), and it has been fitted to the components C l , C2 and C3 (from the original PET) and to one more component (C6) with a peak at 284.3 eV. After the oxidation there is no clear evidence for the presence of the Al-reacted components C4 and C5, although it was necessary to make C6 somewhat broader compared with the situation for C l to C3 (i.e. F W H M = 1.7 eV instead of 1.4 eV as used for the latter three components). This broader form may indicate that C6 strictly has a composite structure, although here it is just taken as a signature of oxidation for that part of the Al-PET interface that involves metal interaction with the original ester groups. Data in Table 3.1 show that C6 is increased at the 30° take-off angle, and this indicates that the oxidized Al-PET complex is oriented more toward the uppermost surface (rather as seen above for the Al-ester interaction in the unoxidized sample).  67  3.2.4  Conclusions Various previous studies have indicated different interfacial bonding behaviors when  thin metallic layers are deposited on PET, and the study reported here adds to the catalog of examples.  The XPS investigation for the Zr deposition showed no evidence for either the  presence of direct Zr-C bonding, or of the breaking of major bonds in the PET structure. The situation appears similar to that reported by Gerenser for Ag/PET [92]. The latter author concluded that the metal interacts especially with carbonyl O atoms, but our evidence from the difference spectra suggests that the Zr interaction may actually perturb the singly-bonded O atoms more than those of the carbonyl type. In any event, the larger binding energy shift seen for the ester C in the difference spectra for C ls, compared with the shifts for phenyl and methylene C atoms, is fully consistent with the ester C being the closest to the center of bonding interactions at the Zr/PET interface. The model in Fig. 3.10(a) appears consistent with the various observations, and this has PET groups wrapping around individual Zr atoms so that electron donation occurs from the metal to unoccupied molecular orbitals in the PET. Each Zr atom is then in an environment of O atoms from the PET chains, and this local situation approaches a coordination complex with Z r - 0 ~ bonding involving both types of O atoms. 8+  8  Consistently the Zr chemical state does not change markedly on exposing to oxygen, the metal atom just becomes a little more "ZrOa-like". The spectral changes on depositing metal in this work are clearly much greater for the Al/PET system. Evidence is presented for the formation of direct bonds of both Al-C and A l - 0 types, and this is similar to conclusions reached in previous XPS studies [81,83,85,86]. Interestingly, the report by Novis et al. [82] of a HREELS investigation on this system indicated that the interactions were especially of the -COO" A l type, and hence more in line +  68  Figure 3.10 Schematic indications of metal-PET interaction models: (a) Zr/PET system; (b) Al/PET system, and (c) Al/PET after O2 treatment.  Al  (b)  Al  A  ,  'C  H  Q  v  o '  C  v  A 2  c '  0  v  ^0 c -  Al Oxidation  Al  / O  (c)  Al — \ /  O  .0 H 2  Al Al  O  A U 0' .Al  Figure 3.10, continued.  70  with conclusions reached in this work for the Zr/PET system.  Chemical bond formation  between A l and the aromatic.ring was supported by changes in the C ls spectrum, including new structure, and a shift to higher binding energy and enhancement in the shake-up peak. The latter changes are consistent with a more localized 7t-electron system [80]. A schematic view of the oxidized Al-PET interface, after exposure to 0 , is indicated in Fig. 3.10((b),(c)). 2  Although different behaviors are reported here, and in other work, for metal-PET systems, it must still remain an open question whether the varying metal coverage represents a key factor in the different observations. A challenge with the very lowest metal coverage is that the spectral changes are small, but it is felt that a conservative use of difference spectra may be helpful for interpreting XPS spectra in such a context.  3.3  XPS studies of the stability of Al/Mg/PET interfaces  3.3.1  Introduction Recent work done by Broms et al. [94] showed the ability of M g to act as a cathode  electrode in a P L E D , but in general performances of such devices depend strongly on the natures of the interfaces between the polymer film and the metal electrodes [95-97]. Device failure may result from the ingress of water or oxygen from the atmosphere to the metal/polymer interface, but encapsulation by glass is one approach to mitigate against such factors, and so improve device lifetime [98,99]. In previous work from this group [84], it was found that the interface formed by PET and thermally-deposited M g is relatively stable in air, but not with water. Indeed, it was shown that a water rinsing could remove the M g component completely from the film, and leave behind a modified PET surface.  The present study  reinvestigates the Mg/PET system, after the thermal deposition of A l , in order to assess whether  71  the added A l can form a protective layer to passivate the Mg/PET system from the destructive attack by water. 3.3.2  Experimental The PET film was prepared as noted in Section 3.2.2, and the in situ preparation and  characterizations were made in the MAX200 facility. The non-monochromatized Al K source a  (1486.6 eV) was used for this project at 10 kV x 20 mA power. The analyzer was operated at a 24 eV pass energy for measuring C ls, O ls and Al 2p spectra and at 96 eV for measuring M g 2p spectra; survey spectra were obtained with a pass energy of 192 eV. The A l evaporation source used was described in Section 3.2.2. The M g evaporation source is illustrated as Fig. 3.2(b); it is composed of a small quartz tube containing M g turnings (Johnson-Matthey, paratonic grade 99.98%) wrapped by tungsten wire. The steps for this study are summarized in Fig. 3.11. XPS spectra were first measured for the cleaned PET surface, which was indicated to be stable under the X-ray exposure. The sample was transferred to the preparation chamber for the deposition of Mg, and back to the analytical chamber for study by XPS. Next AI was deposited on to the Mg/PET sample. After a characterization by XPS, the Al/Mg/PET sample was exposed to air (1 atm) for 2 min, and it was then reintroduced to the U H V system for further analysis. Finally the sample was taken back to atmosphere, rinsed with deionized water for 2 min, dried in air for 30 min and returned to the analytical chamber. All XPS analyses were made from the same lateral position on the sample. The metallic thicknesses were estimated from the attenuation in the C ls integrated intensity to be 8 A and 20 A for M g and A l respectively; this assumed the depositions were even in each case, and mean free paths of 34 A and 25 A for the respective metals [88].  72  PET Mg >  i Mg/PET  PIIT Al  • •"• *.*- •.** *.** *.*»  * *• *.** *.*' *.*? *i,? *i * *i * *i,? *» .* *• * < ? i .* • .*»•*.  Al/Mg/PET  PET Water rinse  PET  Figure 3.11  Jfater rwiserf - Al/Mg/PET  Schematic diagram o f steps for experiments in Section 3.3.  73  3.3.3  Results and discussion  3.3.3.1  Mg/PET  C ls and O ls spectra for the cleaned PET film and for the Mg/PET sample, with the addition of M g 2p spectra for the latter, agree closely with previous observations [84]. In summary, the C ls spectrum for the initial PET film shows components for phenyl carbon atoms (binding energy 285.0 eV), methylene carbon atoms (286.6 eV), ester carbon atoms (289.0 eV), and a weak peak (below 292 eV) resulting from the n—>n* shake-up transition; the O ls spectrum was fitted with two components, one at 531.9 eV for the carbonyl oxygen and the other at 533.5 eV for the ester oxygen. The relative contributions from the different components are shown in Table 3.2 and the situation for clean PET closely matches expectation from the stoichiometry. Also no significant changes are seen when the take-off angle (0) is reduced to 30°, and that strongly suggests that the sample is homogeneous within the X P S probing depth and that the surface contamination is minimal. Observations for the C ls spectrum after the M g deposition are summarized as follows: (i) the ester carbon peak shows a marked attenuation, (ii) a new component is formed at 284.0 eV, and (iii) the n—>K* shake-up peak is essentially unchanged and there is no indication of structure associated with Mg-phenyl bonding analogously to the situation shown by the corresponding Al/PET system in Section 3.2.3.3. The O ls spectrum shows a large attenuation for the singly-bonded O component at 533.5 eV, as well as the presence of a new component at 530.8 eV [100]. The new components in both C ls and O ls spectra after the M g deposition are taken to indicate that the metal attacks the C-to-0 Tt bonds on the ester groups, with formation of O-Mg-C rj-type bonding. Angular-dependent studies (Table 3.2) show that these new components are emphasized at lower take-off angles (i.e. this signal in both C ls and O ls  74  Table 3.2  C ls and O ls binding energies (in eV) and percentages of different  components indicated for Mg/PET and Al/Mg/PET including different exit angles 9.  Carbon components Binding Energy (+ 0.2 eV)  Oxygen components  284.0  285.0  286.6  289.0  530.8  531.9  533.5  -  61.8  20.0  18.2  -  49.5  50.5  Mg/PET (90°)  10.0  62.1  18.2  9.7  24.2  47.8  28.0  Mg/PET (30°)  15.2  56.8  21.1  6.9  32.1  49.1  18.8  Al/Mg/PET (90°)  20.4  49.9  21.6  8.1  42.0  39.7  18.3  Al/Mg/PET (30°)  33.2  53.8  13.0  -  52.9  37.8  9.3  PET (90°)  75  is increased by about 50% on reducing 9 from 90° to 30°), and hence that these components are closer to the surface than the other contributions from PET. The M g 2p spectrum was fitted with two components at 48.5 eV and at 50.6 eV with F W H M values of 1.3 eV and 1.8 eV respectively. The former binding energy is less than values (49.3 to 49.9 eV) reported for bulk Mg metal [101,102], and this discrepancy may be accounted for by M g cluster formation on the PET film [103]. A further indication that the M g is not deposited as a flat layer is provided by the non-observance of plasmon loss features in the M g 2s and M g 2p spectra. The higher M g 2p binding energy is close to the value for MgO (50.8 eV) [104], and this appears consistent with M g - 0 bonding within the interfacial O-Mg-C complex. The angle-dependent spectra for Mg 2p do not distinguish the relative positions of the two M g components, but it is clear that two types of M g exist on the PET surface, namely M g bonded to O in the O-Mg-C complex (supported also by the C ls and O ls spectra) and metallic Mg, assumed to be in cluster form. 3.3.3.2 Al/Mg/PET The M g 2p spectrum after the A l deposition still indicates two components (Fig. 3.12(a)), which are interpreted as a metallic form at 49.1 eV and an O-bonded form at 50.6 eV. The first is now shifted by +0.6 eV from the corresponding value found for the Mg/PET film, while the O-bonded form is unchanged. The shift in the metallic peak may be associated with A l - M g alloy formation. Core binding energy shifts of this sort of magnitude are typical [105107], and in this case an increase appears consistent with trends in electronegativity values. Complementary information is available from the A l 2p spectrum (Fig. 3.12(b)), which shows the presence of a large metallic component at 72.0 eV and a small oxidized component at 74.1 eV. The binding energy of the metallic A l peak is 0.7 eV lower than that reported for the pure metal form (e.g. 72.7 eV reported above for the Al/PS system in Section 3.2.3.2 and by Taylor  76  (a) M 2 p g  Figure 3.12 Spectra from Al/Mg/PET sample: (a) M g 2p showing metallic component at 49.1 eV and an oxidized component at 50.6 eV due to a O-Mg-C complex; and (b) A l 2p showing a metallic component at 72.0 eV and an oxidized component at 74.1 eV.  77  et al. [108]). This shift may represent some combination of effects due to Mg-Al alloying and perhaps some A l clustering (no features associated with plasmon losses are detectable). The binding energy for the oxidized component is also less than that (74.8 eV) observed for the Al/PS system after some oxidation, but it matches to the value associated with O-Al-C bonding in our earlier interpretation for the Al/PET interface in Section 3.2.2.2. So although most A l remains in metallic form, some may be interacting directly with the PET to form the O-Al-C bonding arrangement, although we have no evidence for a direct interaction between A l atoms and phenyl carbon atoms (i.e. no new component is apparent at 282.6 eV, and no change is seen in the  7t—>7t*  shake-up structure).  The O-metal-C components seen in C ls (284.0 eV) and O ls (530.8 eV) occur at the same binding energies as for the deposition of M g and A l separately, and these components appear more prominent (Table 3.2) after the A l deposition (e.g., the change in 0 from 90° to 30° shows that the C and O components increase from 20.4 to 33.2% and from 42.0 to 52.9% respectively). In part this results from the pure PET components now being deeper, and hence less within the XPS probe depth, but it also seems clear, for the conditions of the deposition, that some intermetallic diffusion occurs. After the A l deposition, elemental analysis shows that the M g 2p component is still more prominent at the 30° take-off angle than at 90°, and the relative proportion of M g appears to be larger than for A l (31.1% and 21.2% respectively at 30°). These observations strongly indicate that some M g has migrated to the surface. This is consistent with observations on commercially available A l - M g alloys after heating and anodizing treatments [109-110]. In our situation, it appears that there is sufficient thermal energy to surmount the activation barrier for diffusion so that the component of lowest surface tension (i.e. Mg) can concentrate at the surface.  78  3.3.3.3 Air-Al/Mg/PET After the Al/Mg/PET film was exposed to air for 2 min, the sample was reintroduced to the U H V system and re-analyzed at the previous XPS analysis position. Comparisons of the C ls, M g 2p and A l 2p spectra from the surface, prior to exposure and after, are shown in Fig. 3.13.  The C ls spectrum shows an increase in intensity at 285.0 eV due to some air-borne  carbon contamination, however no change in shape is exhibited, including in the region at lower binding energy where any component associated with the O-Metal-C complex would be expected (spectrum not shown).  After the air exposure, the magnitudes of the oxidized  components (Tables 3.3 and 3.4) are increased in both M g 2p and A l 2p (from 47% to 74% and from 21% to 72% respectively), although there is no evidence from the C ls spectrum that the metal-polymer interface is appreciably changed. Equally no change could be detected in the position of the oxidized peak in M g 2p, although that in A l 2p is shifted to higher binding energy and is broadened (FWHM -2.0 eV). The latter peak can be decomposed into two components at 74.1 eV and 74.8 eV, with relative contributions of 11.9% and 60.1%, assuming that the major component at 74.8 eV corresponds to the actual oxide, whereas the minor component arises from the interfacial O-Al-C complex (74.1 eV). 3.3.3.4 Water-Al/Mg/PET After the air exposure, and subsequent XPS analysis, the Al/Mg/PET sample was again removed from the U H V chamber and this time rinsed with de-ionized water for 2 min. In contrast to the unprotected Mg/PET system studied previously in this group, which manifests a complete removal of M g from the PET surface [84], M g signals were still detected after returning the water-rinsed Al/Mg/PET sample to the analysis chamber, as shown in Fig. 3.14(a), although the M g 2p spectrum assumes the form of a single oxide peak centered at 50.4  79  Figure 3.13 Spectra from Al/Mg/PET samples prior to (thin line) and after (thick line) the air exposure (see text): (a) C ls; (b) M g 2p, and (c) A l 2p.  Table 3.3 M g 2p binding energies and percentages of oxide and metallic components* indicated by X P S for different metallized PET samples (normal exit direction). Oxide component  Metallic component  Sample  B.E. (eV)  Area (%)  B.E. (eV)  Area (%)  Mg/PET  50.6  34.2  48.5  65.8  Al/Mg/PET  50.6  46.9  49.1  53.1  Air-Al/Mg/PET  50.6  74.2  49.1  25.8  Water-Al/Mg/PET  50.4  100.0  -  -  * F W H M values ~1.8 eV for oxide components and ~1.3 eV for metallic components  Table 3.4  A l 2p binding energies and percentages of oxide and metallic  components* indicated by XPS for different metallized PET samples (normal exit direction). Oxide component Sample  Metallic component  B.E. (eV)  Area(%)  B.E. (eV)  Area(%)  Al/Mg/PET  74.1  21.0  72.0  79.0  Air-Al/Mg/PET  74.8  60.1  72.0  28.0  74.1  11.9  -  -  74.2  100.0  -  -  Water-Al/Mg/PET  * F W H M values -1.8 eV for oxide components and -1.3 eV for metallic components  81  1  54  79  '—s—<  53  78  \ M M  \—<—\—'—\—'—\—'—\  52  77  1  51  50  49  7 6 75  74  73  —\—'—I  48  72  47  71  46  70 69  IM MM ' \' \ M M I M M (  293 292 291 2 9 0 2 8 9 288 2 8 7 2 8 6 285 2 8 4 283 2 8 2 281 2 8 0  Binding Energy (eV)  Figure 3.14 Spectra from Al/Mg/PET samples prior to (thin line) and after (thick line) the water rinsing (see text): (a) M g 2p; (b) A l 2p, and (c) C ls.  eV.  The A l 2p spectrum (Fig. 3.14(b)) is also consistent with the removal of the metallic  component, and some loosely bound  AI7O3  from the air exposure, but an oxidized contribution  remains at the new binding energy of 74.2 eV. This value is just 0.1 eV greater than that reported above for the O-Al-C complex in the initial Al/Mg/PET sample. The C ls spectrum shown in Fig. 3.14(c) supports the picture of a protected metal/polymer interfacial region insofar as it exhibits no change in its line shape after the rinsing; the only significant change is the increase in intensity for the peak at 285.0 eV apparently due to the presence of air-borne carbon contamination introduced by the rinsing operation. In summary, it is clear that the water removes appreciable amounts of the metallic A l and the AI2O3 components from the oxidized Al/Mg/PET sample, although there is still sufficient coverage to protect the interfacial complexes formed by A l and Mg from being washed away in the presence of the A l overlayer. This contrasts with the situation in the previous study on the Mg/PET system where the loss of M g after the water treatment left a modified PET surface. 3.3.3.5 Reproducibility of measurements Because of the relatively high cost of XPS measurements, generally the various experiments can only be done once. Nevertheless, it is important to have a sense of the level of correspondence that can be expected for independent measurements on nominally the same sample. Assessments have been made through comparison with other work, and within this study some examples of independent measurements are reported. For example, Tables 3.1 and 3.2 reports compositions of the C ls components determined for PET films by two completely independent measurements.  83  It is seen that the quoted values are consistent within ± 1% in this case; uncertainties are likely to be higher for smaller components, but this gives some guidance for the more favorable situations encountered in this research. Of course for PET the molecular stoichiometry requires contributions of 60%, 20% and 20% for the phenyl, methylene and ester C atoms respectively. But XPS measurements may in general give slightly different values because of the possibility of the presence of some hydrocarbon contamination (which is expected to boost the signal for the 285.0 eV binding energy), and perhaps also the packing of the groups may lead to very slight orientation effects at the PET surface. 3.3.4  Conclusions For the current work, it is concluded that the deposited Al has a protective influence on  the Mg/PET interface, insofar as this interface does not appear to change its character after exposure to water.  But a determination of the mechanism by which this occurs would  inevitably require a more complete understanding of the inter-metallic mixing and oxidation that is indicated. Finally, it is noted that this study was designed to follow behaviors using XPS and accordingly the metallic films were much thinner, and the water treatment more severe, than would occur in practice in a P L E D device.  3.4  Characterization of Al/PPV interfaces by XPS and AFM  3.4.1  Introduction The degradation of interfaces between metallic cathodes and polymer is believed to be a  leading cause for P L E D failure [111]. In PPV, the electron mobility is smaller than that of a hole [112], and consequently their recombination is more likely to take place in the vicinity of the cathode/PPV interface rather than at the center of the PPV or close to the PPV/ITO  84  interface. The heat generated by the recombination process, and the current flow, may cause interfacial degradation. The formation of "dark or black spots" [113] and cathode delamination from the polymer layer are reported in failed PLED devices [114], and it is claimed that they are caused by the reactive interaction between the cathode metal and the polymer layer. A recent study by Friend and co-workers, using short electrical pulse measurements, quantified the temperature rise due to current heating in an active region of a P L E D device to be at least 60°C, and this temperature rise was accompanied by a decline of the device efficiency [115]. No studies have been made to date to establish how the heating effect degrades the cathodepolymer interface. The present work aims to investigate aspects of this problem with the goals of establishing: (i) the bonding mode for A l metallization on PPV, and (ii) the link between temperature change and any degradation at the Al/PPV interface region.  3.4.2  Experimental The PPV sample was prepared by thermal conversion from a soluble precursor polymer  (THT) [116] following the process illustrated in Fig. 3.15. The THT in methanol was spin coated on to an indium-tin oxide (ITO) coated glass substrate, and this was heated under high vacuum (~1(T torr) conditions at 230°C for 12 h as summarized in Fig. 3.15. The PPV sample 3  formed was then treated following the steps illustrated in Fig. 3.16. First, the PPV sample (abbreviated by P) was introduced to the metal deposition chamber (Fig. 2.7) for a heating treatment (100°C, 20 min to remove oxygen) and this defined the annealed sample (AP). Three samples with different A l overlayer coverages (classified as low, medium and high coverages) were prepared in the metal deposition chamber from samples AP to give samples LAIP, MAIP o  o  o  and HAIP (the A l thicknesses are estimated at 4A, 1 IA and greater than 75A respectively). Two of the metallized samples, namely MAIP and HAIP, were heated at 80°C for 10 min in the 85  Heat at 230°, for 12 h under l(r torr pressure 3  Figure 3.15 Thermal conversion of poly{p-phenylene [l-(tetrahydrothiophen-lio)ethylene chloride]} (THT) to polyphenylvinylene (PPV).  86  P (PPV) >3l  PPV ITO  H e a t i n g at 100 °C for 20 m i n  A P (heated PPV) PPV  ITO A l deposition  LAW (low coverage)  MAIP (medium coverage)  HAIP (high coverage) >75A  4A PPV  ITO  ITO  A MAIP (heated MAIP)  H e a t i n g at 80 °C for 10 m i n A HAIP (heated HAIP)  PPV  PPV  ITO  ITO  Figure 3.16 Sample notations and procedures for Section 3.4.  metal deposition chamber to give the samples designated as AMAW and AHAW.  These  samples were then transferred to the analysis chamber, without exposure to the atmosphere, in order to study the effect of temperature change on the Al/PPV interface. XPS spectra were measured in the MAX200 facility using the non-monochromatic A l K a source (1486.6 eV) operated with 10 kV x 20 mA power. The analyzer was operated at 24 eV pass energy for measuring the C 1 s and O 1 s spectra, at 48 eV for the Al 2p spectra, and at 192 eV for the survey spectra. Binding energies were referenced to the hydrocarbon peak in the C ls spectra at 285.0 eV. 3.4.2.1 Atomic force microscopy (AFM) The recent development of new methods in scanning probe microscopy (SPM), including the scanning tunneling microscope (STM) and the atomic force microscope (AFM), has proved to be one of the most exciting developments in surface science since their discoveries by Binning and his co-workers [117,118]. Unlike more traditional topological imaging techniques, such as optical microscopy and S E M , the SPMs can process information not only parallel to the surface (i.e. in the x and y directions) but also along the normal (i.e. in the z direction) to produce three dimensional images.  The S T M and the A F M are close  relatives insofar as both employ stylus-type instruments, in which a sharp probe is scanned raster fashion across a sample in order to detect changes in surface structure at the atomic (or near-atomic) scale. These techniques are separated by what is measured between the tip and the sample surface to produce the topographical imaging. S T M measures a tunneling current produced by application of a bias voltage, while the A F M measures the interaction force between the probe and the sample surface. Indeed, it is this difference that gives the wider applicability to the A F M over the STM; both non-conductive and conductive samples can be  88  measured under ambient conditions by A F M , although some conductivity is required for a sample to be studied by S T M . Continuum theories have been developed for the A F M process [119,120]. Figure 3.17(a) provides a schematic illustration of an A F M instrument. consists of a cantilever with a tip that moves laterally very close to the surface.  The probe The force  between tip and surface varies, and the deflection on the cantilever changes, as the tip is moved across the surface in a raster pattern. The cantilever deflections are detected by a laser system that applies a feedback to a piezoelectric ceramic that controls the distance between the sensor and the sample, for example to provide constant force across the surface. B y this means, the movement of the sample can build up a three-dimensional view of the surface topography. There are several operational modes available for A F M measurements:  (i) contact  mode, (ii) non-contact mode, and (iii) tapping mode (TappingMode™) [121]. The first two have been considered conventional. The contact mode operates in the repulsive force region by the tip being in actual contact with the sample as the scanning occurs, while the non-contact mode operates in the attractive force region without direct tip-surface contact.  In the latter  mode the cantilever curves towards the sample; the tip is oscillated above the surface and the change in frequency of the cantilever oscillation is measured in response to the force gradient from the sample. Although both the contact and the non-contact modes have been used for a range of materials, each has limitations. These include damage of sample and probe by the dragging motion associated with the contact mode, and the relatively weak van der Waals forces detected by the non-contact mode, which results in lower resolution [122]. The tapping mode approach is illustrated in F i g . 3.17(b) [123], and it was developed recently to overcome the limitations just mentioned. The tip is alternatively in contact with the  89  (a)  Laser  Piezoelectric scanner  (b)  Amplitude reduced  L  Figure 3.17 Diagrams for A F M : (a) the operation of A F M ; and (b) cantilever oscillation amplitude in free air and during scanning for tapping mode A F M .  90  surface to provide high resolution, and it is then lifted off to avoid dragging across the surface. A stiff silicon cantilever oscillates with large amplitude (greater than 20 nm) near its resonance frequency when the tip is not in contact with the surface. The oscillating tip is then moved toward the surface until it begins to lightly touch, or "tap" the surface. During scanning, the vertically oscillating tip alternately contacts the surface and lifts off.  As the oscillating  cantilever begins to intermittently contact the surface, the cantilever oscillation is necessarily reduced due to energy loss caused by the tip contacting the surface. The laser focused on the cantilever detects the change in the oscillation amplitude. The digital feedback loop then adjusts the tip-sample separation to maintain constant amplitude and force on the sample. In the present work, the tapping mode of operation was applied using a Digital Instruments Nanoscope® III with a silicon cantilever/tip (cantilever resonance frequency 280320 kHz, spring constant less than 1 Nm' ); this equipment is in Prof. T. Tiedje's laboratory in 1  AMPEL.  3.4.3  Results and discussion  3.4.3.1 PPV surface XPS survey scan spectra confirm that a freshly prepared PPV film is free from impurities, except for some oxygen (8%), to the level detectable by XPS. Signals from the ITO substrate were not detected, thus the thickness of the PPV film must exceed the probing depth of XPS (i.e. the minimum film thickness is 3X, where X equals about 24 A for polymeric materials and electron energies around 1 keV [65]). Higher resolution scans for the C ls and the O ls spectra are in Fig. 3.18((a),(b)). The asymmetric C ls envelope is fitted with one main peak at 285.0 eV and a much smaller peak (10%) at 286.6 eV. The former peak is assigned to both phenylene (Q,H ) and vinylene (-HC=CH-) carbons of the PPV structure, while the latter 4  91  Binding Energy (eV)  Figure 3.18 XPS spectra from annealed PPV film: (a) curve fitted C ls, and(b)0 ls.  peak appears due to carbon single bonded to oxygen either in hydroxyl (C-OH) or methoxyl (C-O-C) bonding. Unlike the cases of PET and PS, the  7t—>7C*  shake up feature, which is  expected for the structure of PPV, is not detected. The reasons for this are not fully clear, but the PPV surface may modify upon exposure to light and oxygen in air to lose its conjugated character [124]. The O ls spectrum exhibits a broad peak (FWHM 2.1 eV) centered at 532.7 eV, and this may indicate the presence of organic C-O species. Several possible origins of the oxygen impurity have been suggested previously. These include hydroxyl replacement of the sulfonium-leaving group during the polymerization [125], formation of carbonyl groups during the thermal conversion process [126], as well as adsorption of water and oxygen from the atmosphere [127]. Upon annealing the sample at 100°C for 20 min under U H V , the oxygen concentration was reduced by half (i.e. to about 4% total content) while the C ls spectrum exhibits no change in its shape, including the C-O region. Given the latter observation for the C ls spectrum, the oxygen removed from PPV by the annealing is more likely in the form of trapped water or oxygen molecules than oxygen chemically bonded to PPV (e.g. as C-O). The reduction of the oxygen content after heating PPV films has previously been reported by others [127,128]. 3.4.3.2 Al/PPV Figure 3.19 shows Al 2p spectra and corresponding A F M images from AP and Al/PPV samples with different Al coverages: LAIP, MAIP and HAIP. For the last sample, no C ls or O ls signals could be detected from the PPV layer, and thus the Al layer thickness must exceed the XPS probe depth. At the initial stage of deposition, to form the sample designated LAIP, the Al 2p spectrum (Fig. 3.19(b)) shows a single peak around 74.1 eV, which is consistent with the presence of Al bonded to oxygen. As the coverage increases, to form the medium coverage  93  sample MAIP, a new peak arises at lower binding energy, 72.0 eV, and this is suggestive of a metallic A l form (Fig. 3.19(d)). However, this A l 2p binding energy is somewhat lower than that measured for the Al/PS system (72.7 eV), but the discrepancy may be accounted for by the metallic growth occurring in cluster form for this initial stage of metal deposition. Similar observations have been made for the Al/PET and Mg/PET interfaces presented in Sections 3.2 and 3.3, and justifications have been given in terms of the electronic levels involved [129,130]. Also, comparable results have been obtained for Al/PPV interfaces by Konstandinidis et al. [131] using XPS. Additionally LeGoues et al. using T E M identified the island formation of Al at the PPV surface [132]. The A F M images measured in this work (Fig. 3.19(e)) show sharpprotruding features for the MAIP sample, although these morphological structures were absent in the low-coverage sample (Fig. 3.19(c)). The development of these features in A F M images coincides with the appearance of the Al metal-like peak in the A l 2p XPS spectrum. Figure 3.19((f),(g)) shows the Al 2p spectrum and A F M image measured from sample HAIP. In contrast to the sample with low Al coverage (Fig. 3.19(b)), a single Al metallic peak is now observed at 72.6 eV binding energy, and this is closer to the regular metallic value (72.7 eV). The A F M image indicates that the surface is much smoother at the higher Al coverage compared with surfaces that had smaller Al deposits (note the change in vertical scale in the A F M image for HAIP compared with those for the MAIP and LAIP samples). All evidence supports the view that for the sample with the thick A l deposit, the nature of the surface becomes similar to that occurring with bulk A l . It is also concluded that Al oxide species exist at the Al/PPV interface region, but as the deposition continues the metallic Al layer covers the surface.  96  Figure 3.20(a) compares O ls spectra prior to and after metal deposition to form sample LAIP. The addition of A l shifts the main peak to lower binding energy, that is to the region where contributions due to A l - O bonding are expected [131].  This change parallels the  appearance of an Al-O component in the A l 2p spectrum. In contrast with O ls, the C ls spectrum measured for the low-coverage sample (i.e. LAIP) shows no change from that of sample AP except for some attenuation. But changes occur at the medium coverage of A l , when the C ls spectrum (Fig. 3.20(b)) shows a new component at 283.6 eV. This is consistent with the A l starting to interact with carbons atoms in the PPV structure to form Al-C bonding [133,134]. The interaction of Al with the PPV surface appears coverage dependent given that A F M and XPS results indicate cluster formation at the initial stage of the A l deposition, although a more uniform A l layer forms at the higher coverage. Plausibly the initial interaction of Al is with oxygen atoms at the PPV surface. The small C-O peak in C ls at 286.6 eV binding energy may be overlapped by a possible Al-O-C contribution [131]. Also from the Al/PET study in Section 3.2, a component at 530.9 eV in the O ls spectrum was interpreted as arising from A l O-C bonding, and a similar contribution may be present in Fig. 3.20(a) at or near the binding energy marked by the corresponding arrow. 3.4.3.3  Heated Al/PPV The thermal stability of the Al/PPV interface is investigated by heating the samples  MAIP and HAIP to 80°C for 20 min under U H V conditions. This was done in the metal deposition chamber of the MAX200 facility with the pressure in the 10" torr range. Al 2p and 8  O ls spectra measured prior to and after these treatments are shown in Fig. 3.21. The changes observed for the AMAIP sample are summarized as follows: (i) the Al 2p spectrum shows an  97  C-O  (a) O ls  I AI-O-C  Figure 3.20 XPS spectra from Al/PPV films: (a) comparison of O ls spectra for A P (thick line) and LAIP samples (thin line); and (b) curve fitted C ls spectrum for sample MAIP.  Figure 3.21 Comparison of XPS spectra for MAIP (thin line) and AMAIP (thick line) samples: (a) A l 2p, and (b) O ls.  increase in the oxide component and a decrease in the metallic component without change in binding energy values; (ii) the O ls spectrum shows an overall increase in the surface oxygen concentration after the heating; and (iii) the C ls spectra (not shown) exhibit no change in intensity or shape.  These observations clearly show that the heated A l / P P V interface is  oxidized, especially by the conversion of metal to oxide.  But unlike the chemical changes  detected by X P S , the A F M image from the A l / P P V surface after heating did not exhibit any morphological change compared with the image taken prior to the heating. A t this point it is not established whether the source of the oxygen was the vacuum system, or whether it came from the sample itself. This was investigated by giving the high A l coverage sample (i.e. HAIP) the same heat treatment as applied to the MAIP sample. The A l 2p spectrum for sample HAIP did not exhibit any changes after the heating to form sample AHAIP. It is assumed that the oxidation is indeed taking place in the HAIP sample, as previously noted for the MAIP sample, and that this occurs at the metal-polymer interfacial region (i.e. below the X P S probe depth for HAIP).  Such evidence supports the view that the source of oxygen must  be the P P V itself, and not the background in the vacuum chamber which would lead to oxidation of the topmost surface, and hence be easily detected by X P S for HAIP.  We cannot  distinguish between oxygen originating at the I T O layer [135] or from within the P P V itself [201,202]. Either way, this species at the A l / P P V interface may be expected to affect operation of a P L E D device.  3.4.4 Conclusions The deposition of A l on a P P V surface under U H V starts with oxidation from oxygen in the polymer, and X P S and A F M measurements are consistent with cluster growth for the initial A l deposition. The changes observed in the X P S spectra mostly agree well with conclusions  100  drawn by Konstandinidis et al. [131] and by Tran et al. [136] for the formation of Al-O-C species as well as Al-C species at the Al/PPV interface. An increase in temperature to 80°C results in oxidation of the metal cathode at this interface region, and this was confirmed by an increase in the Alox component in the A l 2p spectrum, as well as by an increase in oxygen detected according to the O ls spectrum. Evidence was provided to support the conclusion that the origin of oxygen is provided by the PPV bulk rather than by any background oxygen in the vacuum chamber. About half of the oxygen in the PPV surface was removed when the PPV film was heated prior to the metal deposition, and it is an indication that significant amounts of oxygen can be trapped in the PPV structure. Apparently this source can be activated to diffuse to the surface, or to the A l cathode layer, by heating. This is consistent with other observations. Rasmusson et al. used A F M to observe the presence of large pinholes in a surface of PPV film [137]. These structures may represent sites for trapping oxygen and water molecules from atmosphere prior to the metal deposition, although such pinholes were not detected in the present work. The oxidation of the Al-PPV interface may provide one factor that initiates the degradation of the P L E D device performance during operation, insofar as the heating effect is detrimental. This suggests extensive oxygen removal may be needed from the PPV prior to diode fabrication.  101  Chapter 4 Modification of polymer surfaces by a remote plasma treatment  4.1  Introduction The work presented in Chapter 3 focussed on how the natures of metal-polymer  interactions vary depending on the particular materials involved. In other words, the chemical structures of both the polymer and the metal play a role in determining the nature of the interface.  For example, direct chemical bonds were believed to be formed in the Al-PET  interfacial region, via A l interaction with the ester and phenyl groups of the PET structure, while no such bond formations were detected in the Al-PS interface due to a lack of reactive functional groups in the PS structure.  For achieving greater adhesion in the metal-polymer  interface, the former type of interaction is preferred. Thus, some means of modification to a polymer are often considered in order to bring its surface into a favorable state for the metal adhesion. Common methods used include chemical etching, ozonolysis, mechanical abrasion, U V exposure and a corona discharge process [138,139]. More recently, plasma treatments are being considered for the modification of polymer surfaces. In this technique, the polymer surface is exposed to plasma produced from gas (the type chosen depending on the purpose of the modification), but a complex interaction can occur between the polymer surface and various components of the plasma, including electrons, ionic species, radicals and photons [140-142]. Some processes are illustrated schematically in Fig. 4.1. In this approach, the polymer surface can be modified, with short processing times, while preserving the favorable bulk structure. Also unlike the corona discharge method, which uses  102  Plasma e " + A —> A * + e A * —> A + hv e " + A —> A  e"  A  +  + 2e "  +  t * Energy Transfer Photon absorption  Adsorbed A , B  polymer  Figure 4.1 Schematic indication of processes involved in a plasma formed by molecular gas A B .  103  air as the source gas at atmospheric pressure, a plasma treatment has more potential for controlled modification of the surface by using different types of gases, and unlike the use of wet chemical etching processes, a dry plasma technique has no waste disposal issues to deal with.  Common by-products from plasma treatments are C 0 and water vapor, although 2  quantities are generally small. Despite these advantages, the plasma modification process can produce low-molecular-weight-oxidized-materials (LMWOM) as a result of bond breakage in the original carbon backbone structure due to the vigorous interaction of plasma and surface [143]. Often the production of L M W O M leads to formation of a weak boundary layer (WBL), on which subsequently deposited material exhibits poor adhesion, with potential fracture occurring at the WBL/bulk interface. Many efforts have been made to understand the interactions between various types of plasmas and polymers [144-149], however the complexities of the plasma composition and the strong dependency of the modification pattern on both the polymer structure and parameters of the plasma system, such as gas, system configuration, treatment time, pressure and temperature, have made it a challenging task to understand the interactions involved [150-152]. Improvements in the plasma modification process require the formation of the W B L to be minimized, and in turn that depends on relatively small modifications being made to the polymer structure. This principle guided the design for the present work, which uses a remote hydrogen plasma. This is different from the direct plasma method in which the sample is placed directly in the region where the plasma is generated. For the present work, the distance between the plasma region and the sample surface is about 30 cm, and this is expected to result in a gentler and more controlled interaction with the polymer surface. Also hydrogen is used as primary plasma source, although its operation inevitably gives desorbed water from the  104  system's walls. But H is the main component, and its effects should be gentle. This is both 2  from a chemical point of view, and its small mass should help minimize the effect of sputter etching of polymer surfaces caused by physical bombardments of ions and radicals that may result in the formation of unwanted W B L . In the present work, the effects of the remote hydrogen plasma process are studied on the three polymers PS, P B M A and PET (Fig. 3.1). These were chosen because they have different functional groups. Thus PS has phenyl groups but no oxygen functionalities, P B M A has ester groups while PET has both phenyl and ester groups.  4.2  Experiment  4.2.1  Sample preparation  The PS and P B M A films were prepared by spin-coating the PS and P B M A solutions in toluene onto freshly polished aluminum panels. After spin coating, each film was dried for 24 h in air before being introduced to the analysis chamber of the M A X 2 0 0 facility. These spin coated films were sufficiently thick that no XPS signals could be detected from the A l substrate. Commercially available PET film (8 (im thick) was cleaned with methanol and deionized water, and then dried in air for 24 h before placing in the vacuum system for XPS analysis.  4.2.2  Plasma treatment  Figure 4.2 represents a schematic diagram of the remote plasma system which is located in the plasma chamber in the MAX200 facility as indicated in Fig. 2.7.  The plasma was  generated by a microwave discharge operating at 2.45 GHz with a total power of 60 W. During use, a stream of hydrogen gas at 2 torr pressure was continuously flowed through the quartz  105  Figure 4.2 Schematic diagram of the system used for the remote hydrogen plasma treatment.  tube, and each sample was treated by the remote plasma for 5 min before transferring in situ to the analysis chamber for XPS measurement. 4.2.3  Surface characterization XPS spectra were obtained in the MAX200 facility with the M g K source operating at a  10 kV x 20 mA power. The analyzer was operated at 192 eV pass energy for measuring survey scans and 48 eV pass energy for the higher resolution spectra; photoelectrons were collected from a sample area 2 x 4 mm either in the direction of the surface normal (take off angle 90°), 2  or from the exit direction corresponding to a take off angle of 30°. During XPS measurement, a positive charge builds up at the non-conductive polymer surfaces, so causing apparent shifts in energy scale. This was corrected by referencing the C ls hydrocarbon peak to a binding energy of 285.0 eV for all spectra studied in this project. The spectra were curve fitted using the procedure discussed in Section 2.4. Those from the untreated polymer samples were curve fitted with components at the reference binding energies [65,153] summarized in Table 4.1. Spectra from the plasma-treated samples were fitted with additional components as discussed below while keeping the parameters for the original components constant.  4.3  Results and discussion  4.3.1  Polystyrene (PS) Figure 4.3 shows XPS survey spectra from the untreated and the plasma-treated PS  samples. Prior to the plasma treatment, the spectrum indicates no contamination in the PS surface including oxygen species. There is a small peak at about 610 eV. This is a ghost peak, which arises as a C ls photoelectron peak excited by C u L radiation. As the anode material in a  the X-ray source wears down, some copper can be exposed with the production of radiation  107  Table 4.1 Carbon and oxygen component binding energy values (eV) used in curve fitting analysis. Values from Beamson and Briggs [65].  Carbon components  ,. Binding energy  Oxygen _ components  C-FLC-C  285.0  0*=C-0 or  5  C-O  286.5  O-C  532.8  0-C-OorC=0  288.0  O-C-O  532.5  0=C-0  289.0  0=C-0*  533.5  J  b  . ,. Binding energy D  3  )  9  108  Cls (a)  I  C(A)  1 1000  '  I 800  1  I 600  1  I — ' I 400  1  200  I 0  Binding Energy (eV)  Figure 4.3 Survey spectra from (a) untreated PS film, and (b) plasma treated PS film.  109  characteristic of this element on the electron impact [64]. A binding energy shift of 323.9 eV is expected for C u L compared with M g K , which is consistent with this interpretation (i.e. C Is a  at 285.0 eV for M g K  a  a  and 608.9 eV for C u L ) . a  After the plasma exposure, the survey  spectrum from the PS surface shows the presence of oxygen photoelectron and Auger peaks at about 530 eV and 730 eV respectively. It is clear that some incorporation of oxygen has occurred into the polymeric surface. Similar observations have been made previously for other plasmas generated by non-oxygen gases; for example, Lub et al. [154] observed some incorporation of oxygen species into a PS surface treated with a N plasma and Callen et al. 2  [155] also detected oxygen in the surface of PS after an Ar plasma treatment. As the chamber pressure increases to 2 torr during the plasma treatment, the displacement of gases such as water and air from the chamber walls is expected, and such gases are likely to be the source of oxygen detected on the PS surface after the remote hydrogen plasma treatment. The C ls spectrum from the untreated PS sample (Fig. 4.4) shows a dominant symmetrical peak at 285.0 eV corresponding to carbons in the phenyl ring and the backbone of the PS structure. The small Tt—>%* shake up feature below 292 eV is commonly detected for an aromatic system, as noted in Section 2.2. After the plasma treatment, the C ls structure became unsymmetrical and the shake up peak disappeared. The loss of the shake up peak strongly suggests that phenyl rings in the surface region become modified during the interaction with the plasma to lose their aromatic character. Callen et al. [155] using a remote Ar plasma treatment and a longer treatment time (i.e. 60 min), and Occhiello et al. [156] using a direct 0 plasma 2  treatment also reported the loss of the K^K* feature. Although the interaction of the plasma with the PS surface is considered to be gentler in the present work compared with these other studies (shorter treatment time, longer distance between the sample to plasma generated  1 10  1 < \ 294  292  < I 290  « I < I 288  286  <-I 284  < I 282  5  1 280  Binding Energy (eV) Figure 4.4 C l s spectrum from untreated PS film.  111  region), the modification of the phenyl ring was still indicated by the loss of the shake up feature. This is consistent with the phenyl ring structure showing a high reactivity to H atoms and other species produced by the remote hydrogen plasma treatment. Other researchers have reached similar conclusions about the high reactivity of the phenyl ring towards the plasma treatment by using other techniques, including time-of-flight secondary ion mass spectrometry, and also by studying changes in the PS valence band using XPS. The initiation of the oxidation is very likely the result of C-H bond rupture on the carbon atom bonded to the aromatic ring by singlet state molecular oxygen [157,158]. An indication of the chemical structure of the incorporated oxygen in the PS surface can be deduced by curve fitting the higher-resolution spectra. A new small component was fitted in the C ls spectrum at the binding energy of 286.5 eV, that is 1.5 eV higher than the main hydrocarbon peak at 285.0 eV as shown in Fig. 4.5(a). The O ls spectrum exhibits a single symmetrical peak at 532.8 eV for a 90° exit angle in Fig. 4.5(b). These new binding energies are consistent with ether (C-O-C) or alcohol (C-O-H) bonding [155,159]. Furthermore, the ratio between the O ls component and the new C-O component in the C ls spectrum obtained using Equation 2.4 is 0.9. This value is closer to the value expected for each O atom bonding to only one C atom (i.e. C-OH and a ratio of 1) than for each O atom bonding to two carbon atoms (i.e. C-O-C and a ratio of 0.5); accordingly it seems more likely that the hydrogen plasma as used in this work results in some C-OH bonding at the surface. The C ls and O Is spectra from the plasma-treated PS surface measured at the smaller take off angle of 30° are shown as Fig. 4.5((c),(d)) respectively. New small components are apparent in both spectra: at 288.0 eV in C ls and at 531.8 eV in O ls. These new components are attributed to C=0 species and they are only detected at the small take off angle, where the surface sensitivity is  1 12  enhanced.  Table 4.2 summarizes percentages of the different components deduced for the  curve fitted C ls and O ls spectra. This evidence strongly supports a model where the oxygen species are located in the topmost surface region. 4.3.2  Poly(butylmethacrylate) (PBMA)  Curve fitted C ls and O ls spectra from the untreated P B M A sample are shown in Fig. 4.6((a),(b)) respectively. The C ls envelope is fitted with three components, a hydrocarbon (CFf and C-C) component at 285.0 eV binding energy, a carbon singly bonded to oxygen (C-O) component at 286.5 eV and an ester carbon (0-C=0) component at 288.9 eV. The O ls spectrum is fitted with two components; one at 531.9 eV for oxygen doubly bonded to carbon (*0=C-0), and the other at 533.5 eV for oxygen singly bonded to carbon (0=C-0*). No changes in these spectra are observed on changing the take off angle and it is concluded both that the contamination of the surface is negligible and that the distribution of these components is reasonably homogeneous within the XPS sampling depth. Furthermore, the relative areas of the components show good agreement with the theoretical values expected for the P B M A structure, and the details are summarized in Tables 4.3 and 4.4. Figure 4.6((c),(d)) represents C ls and O ls spectra from the plasma-treated P B M A surface; new components are now needed in the curve fitting for the spectra in addition to the original components. The new components are at 288.0 eV and 532.5 eV in the C ls and the O ls spectra respectively. These new binding energy values are consistent with the formation of O-C-0 groups, which is as carbon singly bonded to two oxygen atoms.  The introduction of this new species possibly arises by  modification of the ester group to form -O-C(OH)-, and indeed the ester carbon component (0=C-0) shows a significant reduction (by 30%) after the plasma treatment (Table 4.3).  Table 4.2 Relative compositions (%) from curve fitted C Is spectra from PS film prior to and after plasma treatment. Results are given for two take off angles for plasma-treated PS.  Sample Types  C-H, C-C  C-O  C=0  Untreated PS  93.6  -  -  6.4  Treated PS (90°)  92.2  7.8  0.0  0.0  Treated PS (30°)  87.9  9.6  2.5  0.0  JC-Mt  :  115  Table 4.3 Relative compositions (%) from curve fitted C 1 s spectra from P B M A film prior to and after plasma treatment. Results are given for two take off angles for plasma-treated P B M A .  Sample Types Untreated  C-H, C-C  C-O  o-c-o  0=C-0  71.5  14.1  -  14.4  14.3  -  14.3  71.4  T  1  1  Treated (90°)  73.1  13.7  3.0  10.2  Treated (30°)  72.6  13.8  4.0  9.6  Theoretical values based on P B M A formula.  Table 4.4 Relative compositions (%) from curve fitted O ls spectra from P B M A film prior to and after the plasma treatment. Results are given for two take off angles for plasma-treated P B M A .  O/C ratios measured from XPS spectra are  also included.  o*=c-o  o*-c-o  0=C-0*  O/C  49.5 50.0  -  50.5 50.0  0.29 0.28^  Treated (90°)  43.7  11.5  44.8  0.34  Treated (30°)  39.1  13.9  47.0  0.36  Sample Types Untreated  1  Theoretical values based on P B M A formula.  1  The angle dependent measurements confirm that components associated with the new O - C - 0 functional group are emphasized at the smaller take off angle in both C ls and O ls spectra; this supports the view that these groups are particularly formed at the uppermost part of the surface region. Correspondingly, the  0=C-0  component that occurs in the bulk structure  is reduced at the smaller take off angle in the C ls spectrum. Vargo et al. [ 1 6 0 ] also concluded that O - C - O species are incorporated into a related surface, that of poly(methylmethacrylate), on treating with an 0  2  plasma.  Those authors  analyzed their surface with ion scattering spectroscopy (ISS) and attenuated total reflectanceFourier transform infrared (ATR-FTIR) spectroscopy in addition to XPS. The main difference between the results presented here, and from those of Vargo et al., is in the extent of the surface modification reported. The experimental configuration and treatment time are similar in each case, but Vargo et al. found approximately twice as much of the newly incoiporated species as in the present case. This comparison supports the view that the hydrogen plasma has a milder modification effect on the polymer surface compared with the O T plasma.  4.3.3  Polyethyleneterephthalate (PET)  Figure 4.7((a),(b)) shows curve fitted C ls and O ls spectra from the unmodified PET surface, but the assignment of the components has already been discussed in Section  3.2.3.1.  The relative compositions of each component in the C ls and O ls spectra are tabulated in Tables 4 . 5 and 4 . 6 respectively. The curve fitted C ls and O ls spectra from the plasma-treated PET film are shown in Fig. 4.7((c),(d)) respectively. disappearance of the  n—>7t*  After the plasma treatment, the C ls spectrum shows the  shake up feature, and hence the same modification of the phenyl  rings appears to be taking place in the PET surface as has been reported in Section 4 . 3 . 1 for the  1 18  Table 4.5 Relative compositions (%) from curve fitted C ls spectra from PET film prior to and after plasma treatment. Results are given for two take off angles for plasma-treated PET.  C-H, C-C  C-O  O-C-0  0=C-0  TT—>7T*  59.0 60.0  19.8 20.0  -  2.8  -  18.4 20.0  Treated (90°)  63.7  14.4  6.5  15.4  0.0  Treated (30°)  64.6  12.9  8.8  13.7  0.0  Sample Types Untreated  1  1  -  1  Theoretical values based on PET formula  Table 4.6 Relative compositions (%) from curve fitted O ls spectra from PET film prior to and after the plasma treatment. Results are given for two take off angles for plasma-treated PET. O/C ratios measured from XPS spectra are also included. Sample Types Untreated  o*=c-o  o*-c-o  45.6 50.0 1  o=c-o*  O/C  54.4 50.0  0.42 0.40  1  1  Treated (90°)  36.9  21.6  41.5  0.34  Treated (30°)  35.0  24.9  40.1  0.30  Theoretical values based on PET formula.  C M  i  PS sample. Additional new components are fitted in the C ls and O ls spectra at 288.0 eV and 532.6 eV binding energies respectively. These values are very close to those observed in the plasma-treated P B M A surface, and hence it is suggested that O-C-O species are also formed in the treated PET surface. Again angle dependent XPS supports the model that the newly formed O-C-0 species are located closer to the surface compared with the ester (0=C-0) groups that represent a key part of the bulk structure. Details of relative compositions deduced from the C ls and O ls spectra are summarized in Tables 4.5 and 4.6. Although an apparent O-C-0 species was found in both the P B M A and PET surfaces, after the remote hydrogen plasma treatment, larger amounts are indicated for the PET surface. The amounts for P B M A are 3.0% and 11.5% from C ls and O ls respectively; the corresponding amounts for PET are 6.5% and 21.6%. The structural differences between the two polymers, and more specifically the presence of the phenyl groups in PET, may influence the extent of the modification, although the actual mechanisms are still unknown. With more reaction, it is possible that more L M W O M such as C O and C O 2 form in the plasma-treated PET sample than for P B M A . Some of these low molecular weight species are then pumped away from the surface, or alternatively diffuse into the bulk [161]. This interpretation appears consistent with an observed O/C decrease in the plasma-treated PET sample (Table 4.6), whereas a slight increase in the O/C ratio has been indicated for the plasma-treated P B M A sample (Table 4.4).  4.4  Concluding remarks An important point for the research described in this chapter is that it was possible to  perform in situ surface analysis without the samples being exposed to atmospheric  121  contamination after the plasma modification treatment. Also the attempt was made to carry out this work so that all modifications are restricted to the topmost few layers of the polymer. This study clearly shows that the remote hydrogen plasma, which inevitably is associated with minority amounts of water and/or oxygen, is able to introduce new chemical functional groups involving oxygen into the surfaces of the polymers PS, P B M A and PET. The evidence is provided by the new components that are required to be fitted in the C Is and O Is spectra of the plasma-treated samples. The formation of new oxygen functional groups has been attributed by others to improvements in the adhesion properties of the plasma-treated polymers [144,148,149], but the work here may offer a new direction for ensuring a gentler surface modification. The evidence in this chapter suggests that the extent of surface modification depends on the chemical structures of the polymers concerned. Compared with the non-aromatic polymer P B M A , the polymers like PS and PET with the phenyl groups appear to be more reactive to the remote hydrogen plasma. This is shown by the loss of the  Tt—»TT*  shake up feature and by the  higher percentages of newly formed species accordingly to the C ls and O ls spectra. Others have reported comparable observations for PS and PET, compared with polyethylene (PE) and P B M A , using such modification procedures as a remote nitrogen plasma [150], and direct oxygen, argon and nitrogen plasmas [162,163]. Polymers with the aromatic functionality (e.g. polyimide and PET) have been shown to be particularly sensitive to U V irradiation, and can be modified by exposure to much longer wavelengths than are required for the modification of the saturated polymers like PE and polytetrafluroethylene (PTFE) [164].  This results from  differences in the electronic configurations, and appears to depend on the higher photon energies which are required to induce electronic transitions in the more saturated polymers (e.g.  122  there are only  o-»G*  transitions for PE, while the PET structure possesses more lower-energy  transitions involving the n and 7t* levels). In the context of the hydrogen plasma treatment, it can indeed be expected that the greater numbers of the low-lying electronic transitions in polymers such as PS and PET, compared with P B M A , can lead to greater chemical reactivities and hence a higher degree of surface modification. Other studies using different types of plasma systems appear to give greater amounts of oxygen bonded species than was obtained in the present work [160,165,166]. For example, the use of remote 0 and FLO plasmas yielded higher amounts of C-O, C=0 and 0=C-0 species 2  incorporated into surfaces of PS [162,167].  The greater extent of the surface modification  implies more bond breaking and chain scission, and larger amounts of L M W O M can be produced resulting in a weaker boundary layer, and hence poorer adhesion promotion. The lower degree of surface modification achieved in the present work by using the remote hydrogen plasma treatment appears encouraging, although it is still not clear whether the reduction for the aromatic groups seen in PS and PET is good or not when extensions are taken to consider the adhesion of metallic layers, for example in the fabrication of new PLED devices. Certainly it will be necessary to keep a good balance between the extent of chemical modification and the formation of L M W O M in producing effective modifications of polymer surfaces for improved metallic adhesion.  123  Chapter 5 Characterization of aluminum surfaces after different pretreatments and exposure to organosilane coupling agents  5.1  Introduction As mentioned in Section 1.3.2, the question of how to optimize the direct Metal-O-Si  bonding at metal-silane interfaces has proved challenging, but in part this arises because of the often poorly characterized nature of the surfaces to which the silane coatings are applied. As well as the thin oxide layer, surfaces of aluminum and its alloys are generally also covered by a less-strongly bonded carbon layer [168].  In order to limit the role of such  adventitious layers in subsequent processings, a surface pre-treatment is normally applied, although its effect on the surface is often not well established.  For example, Digby and  Packham [169] recently reported on the durability of adhesive bonds, as assessed by wedge tests, when an epoxy resin was applied to an aluminum alloy after different pre-treatments. These observations showed that the bonding effectiveness, and hence the physicochemical properties of the aluminum surfaces, vary significantly according to the type of pre-treatment used. The work in this chapter is involved with similar issues, but the emphasis is on the effects of surface pre-treatment on the adsorption of different silanes to the aluminum surface. High-purity metal samples are used in order to minimize the effect of the secondary elements that are present in alloys. This study has two main objectives. First, surface science methods including X-ray photoelectron spectroscopy (XPS), atomic force microscopy (AFM) and scanning election microscopy (SEM) are used to characterize and compare the natures of these sample surfaces  124  after the different pre-treatments.  Second, these characterizations are extended to assess the  effects of the pre-treatments on the interaction with different organosilanes. Three specific silanes are considered, and their monomer structures are included in Table 1.2. In the text, these silanes are referred to by the abbreviated notations y-GPS, BTSE and y-APS.  5.2  Experimental Samples ( l x l cm") of aluminum panel (99.99% purity) were prepared by machine  polishing with alumina sandpaper (1200 grit). The polished samples were then degreased in acetone, methanol and distilled water.  This pre-treatment defines the sample M (for  mechanical polished). Six distinct pre-treatments were considered, but all started with this procedure. Sample Ac was acid etched by immersing the mechanically polished sample in H S 0 solution (50 vol%) at 40°C for 2 min, followed by a 2 min water rinse. Sample Al was 2  4  alkaline etched by immersing the mechanically polished sample in NaOH solution (10 vol%) at 45°C for 1 min, then treated in HNO3 (10 vol%) for 1 min, followed by a 2 min water rinse [5]. Three further samples were considered (included in Table 5.1) namely those designated AcA (sample Ac but additionally stored in air for 2 weeks), AcW (sample Ac but additionally stored in distilled water for 1 week), and AlW (sample Al but additionally stored in distilled water for 1 week). After each pre-treatment, the sample was dried in air for 30 min before application of a silane coating. This involved dipping a pre-treated aluminum sample in freshly prepared 1 vol% silane solution (y-GPS, BTSE and y-APS, all from Aldrich, no pH adjustment made) in 95/5 vol% mixture of ethanol and distilled water.  Each pre-treated aluminum sample was  immersed in the silane solution at room temperature for 5 min with constant stirring.  Table 5.1 Summary of pre-treatments and their codes used in this work.  Code  Treatment  M  Mechanically polished and degreased  Ac  As for M, then 2 min in H S 0 (50 vol%) at 40°C, 2 min water rinse  Al  As for M, then 1 min in NaOH (10 vol%) at 45°C, 1 min in H N 0 (10 vol%),  2  4  3  2 min water rinse AcA  As for Ac, then stored in air for 2 weeks  AcW  As for Ac, then stored in distilled water for 1 week  AlW  As for Al, then stored in distilled water for 1 week  126  XPS analyses were performed in the MAX200 facility with the non-monochromatized A l K source (1486.6 eV) operated at 10 kV, 20 mA. Survey scan spectra were measured with a  a pass energy of 192 eV, while higher resolution spectra were measured at 96 eV pass energy; binding energies were referenced to the adventitious hydrocarbon component at 285.0 eV in the C ls spectrum. Unless otherwise stated, measurements were made at normal exit direction from the surface (a few measurements were made for the 30° exit direction). A F M observations were performed in tapping mode using a Nanoscope III (Digital Instruments) with a silicon cantilever tip, as described in Section 3.3.2.1. The mean roughness (R ) of the surface a  relative to a reference center plane was calculated using Ly Lx  (5.1)  where f(x,y) gives the perpendicular coordinate relative to the center plane, while Lx and Ly are dimensions of the surface studied in the lateral directions. 5.2.1  Scanning electron microscopy (SEM) The scanning electron microscope (SEM) has proven to be an extremely useful  technique, particularly to characterize surface morphology by imaging the secondary electron emission. The use of an electron beam can provide a thousand-fold increase in resolving power compared with light.  Current S E M instrumentation, with a finely-focussed field emission  electron gun, can resolve detail to approximately 5 nm [170]. For a given accelerating voltage, the geometry of the sample surface relative to the beam has a great influence on the intensity of the emission signal, and this gives rise to the contrast seen in electron micrographs. It is the low-energy secondary electrons (10-50 eV) that are detected, and since they are strongly absorbed by the sample, they must be produced near its surface in order to contribute to image  127  formation. Also these electrons are strongly dependent on the surface topology, including the local angle between the surface and the incident beam. A Hitachi S4100 S E M (Fig. 5.1) was used for this study [171]; it is equipped with a cold-cathode field emission electron gun, and it operates under a vacuum of 10" torr. The main components of the S E M are illustrated in Fig. 6  5.2. The illuminating system consists of an electron gun and condenser lens assembly to focus the beam on to the sample. The electron gun has three main components: a filament, a shield and an anode. The electrons are produced by field emission and accelerated to about 30 keV. The condenser lens assembly then reduces the electron beam diameter from about 25,000 A to as low as 30 A. The narrower the beam the higher the spatial resolution possible, although that also depends on the area from which the secondary electrons are collected at any instant. The electron beam is scanned across the sample by deflection coils, while the detector counts the secondary electrons emitted from each region probed on the surface. The detector has a scintillator and photomultiplier, and the final amplified electron signal is displayed on a cathode-ray tube (CRT), from which photographs can be made for permanent record.  5.3  Results and discussion  5.3.1  Aluminum surfaces  5.3.1.1 XPS observations XPS survey spectra from the pre-treated aluminum samples show signals from aluminum and oxygen, as well as carbon. The percentage of carbon varies with the pretreatment (Table 5.2), but in general the forms of the C ls spectra are consistent in each case with the major part of the surface carbon (>96%) being hydrocarbon, with less than 4% involving C=0 or C-O groups. Accordingly it is concluded that the surface carbon is mainly  Figure 5.1 Illustration of a Hitachi S4100 S E M . Redrawn from ref. [171].  129  Figure 5.2 Diagram of S E M main components.  from air-borne contamination, and further that the majority of the oxygen detected is bonded to aluminum.  Measured A l 2p spectra from these samples (Fig. 5.3) manifest two resolved  components: each shows a metallic component (i.e. A1(0)) at 72.3 eV binding energy, and an oxidized component (Al(III)) at or near 75.5 eV.  Angular dependent XPS (ADXPS)  measurements are fully consistent with the oxide layer being uppermost; thus the oxide component is emphasized relative to the metallic component when these spectra are measured at small exit angles from the sample, compared with when the photoelectrons are collected at the normal exit direction. The relative areas of the oxide and metallic components (I  ox  and I respectively) vary m  between the samples, and this reflects (in part at least) the varying thickness (t ) of the oxide ox  layers. Values of the latter parameter have been estimated assuming that the oxide layer forms a homogeneous film over a flat metallic substrate, and using the approach outlined in Section ,2.3.3. Equations 2.5 and 2.6 can be rearranged to give  tox  —  A.ox In  nm A.m Iox flox TloK Im  (5.2)  +1  when, using the notation of Section 2.3.3, the mean free paths X  moK  and X  o x  are taken as equal,  the take-off angle 9 is fixed at 90°, and the photoelectron cross sections G , G m  o x  are also taken  as equal. Strohmeier [172] evaluated Equation 5.2 with the assumptions that the ratio of the volume densities of aluminum atoms in metal to oxide (n /n ) is 1.5, and that the mean free m  ox  path values for X and A, are 26 A and 28 A respectively to give m  ox  (  t « = 28 In  T ^ 1.4—1  (5.3)  In,  131  O ls  A l 2p  •—i— —i—•—t— —i— —i—•—i 1  1  79  77  75  1  73  71  69  r  - 1  —i— —i— —i— —i— —i—•—i— —i 1  1  1  1  1  540 538 536 534 532 530 528 526  Binding Energy (eV) Figure 5.3 Resolved Al 2p and curve-fitted O ls spectra measured for the pretreated samples designated Al, Ac; M and AW in Table 5.1.  Values oftox for different samples are calculated using Equation 5.3 and included in Table 5.2. Figure 5.3 shows that small variations in the positions of Al(III) peak maxima measured in A l 2p spectra in this work (e.g. from 75.0 to 75.8 eV) are apparently matched by systematic changes in the corresponding O ls spectra; similar relations, although for differently treated aluminum samples, were noted by Amstutz and Textor [173]. From considering the structures of crystalline oxides, hydroxides and oxyhydroxides of aluminum, Thomas and Sherwood [174] emphasized the difficulties in allocating specific O H and oxide groups from core-level spectra of the minerals, and additionally it has been argued that varying defect distributions in different samples can also contribute to the measured binding energy shifts [175,176]. In the context of non-crystalline oxidized films on aluminum, there are inevitably distributions over the different types of bonds and defects, and some have allocated structure in O ls spectra according to a recognition that O H groups from a given sample tend on average to be found at a higher binding energy than oxide groups from the same sample [177-179]. In that spirit, O ls spectra have been fitted in this work with O H groups at 534.0 eV binding energy and the oxide contribution at 532.5 eV. The subsequent estimates of the fraction of oxygen in the OH form (foH ), rather than oxide form, in the samples considered here, are included in Table 5.2. It is clear from the information in Table 5.2 that the details of surface composition vary considerably for the different pre-treatments.  The mechanically polished sample has the  highest carbon contamination (33.8%), but this is reduced for the chemically etched samples, namely the acid etched sample (Ac, 27.7%) and the alkaline etched sample (Al, 21.6%). Exposure to the laboratory atmosphere inevitably increases the contamination, but the presence of carbon is indicated to be reduced by storing samples in distilled water, as also observed by Underhill et al. [180]. However for technological applications the ubiquitous carbon  133  Table 5.2 Percentage compositions, fraction of O H in aluminum oxide film ( T H ) and 0  thickness of oxide film (t ) indicated by XPS after different pre-treatments. ox  o  Sample  Percentage compositions  t (A)  t  0X  0H  O  C  Al  M  49.7  33.8  3.5  13.0  51  0.77  Ac  43.7  27.7  8.1  20.5  42  0.33  Al  42.1  21.6  5.6  20.7  29  0.20  AcA  39.8  37.2  4.3  18.7  55  0.46  AcW  56.0  25.8  0.0  18.2  *  0.94  AlW  54.6  19.7  1.5  24.2  88  0.90  m  Al x 0  Oxide layer is too thick for underlying metal to be detected by XPS.  contamination is virtually unavoidable, and for this reason we need to build up knowledge for adsorption on surfaces with at least some of this contamination. Different pre-treatments lead to different oxide thicknesses, as judged by the application of Equation 5.3, and information is included in Table 5.2 and Fig. 5.4. The chemical etching processes act to reduce the oxide film, as well as reduce the carbon contamination, and on both counts the alkaline etch has a stronger effect than the acid etch as done in this work. A subsequent storage in water does lead to a thickening of the oxide layer, and this is accompanied by an increase in the fraction of hydroxide ( f o H ) in the oxide layer. Figure 5.4 indicates more generally for a series of differently prepared surfaces that there is a tendency f o r foH to increase with oxide thickness. A similar observation was made by Alexander et al. [181] for a series of air-exposed aluminum samples that had been prepared differently from the present work. The reaction with water apparently leads to a greater incorporation of OH as the oxidation proceeds. Studies with A D X P S are fully consistent with the O H being concentrated in the region of the uppermost surface; for example, component peak areas in O ls spectra from the AcA sample correspond to 46% OH, when the photoelectrons are collected at the normal exit angle, but 61% O H when the spectrum is measured at the 30° exit angle. 5.3.1.2 Morphological features The discussion of results in Section 5.3.1.1 was made using the working hypothesis that the samples had essentially flat surfaces, although it is generally recognized that samples of this type are likely to have considerable roughness and even porous characteristics [5]. XPS is less sensitive to such features, although the defect characteristics of these samples are likely t o affect their chemistries and adsorption properties. Accordingly, a fuller understanding of the  33  o  tox  (A)  Figure 5.4 Values of oxide thickness (t ) and fraction of hydroxide (f ) measured from XPS for different samples considered in this work. The circles identify samples defined in Table 5.1; the triangles represent a series of samples given different air exposures starting from Ac. The dashed line is a visual guide only. ox  0H  effects of the various pre-treatments should be helped by having independent characterizations of their respective topographies. The A F M and S E M micrographs (Figs. 5.5 and 5.6) emphasize the different physical natures of these variously-treated samples. Line-like features from the mechanical polishing for sample M are strongly apparent in the micrographs from A F M (Fig. 5.5(a)) and S E M (Fig. 5.6(a)).  Such features are notably reduced by the chemical etching treatments (Ac and Al),  although they are still visible in the S E M images taken from the acid-etched sample (Fig. 5.6(b)).  Features from the mechanical polishing are completely removed by the alkaline-  etching treatment to form sample Al (Fig. 5.6(c)), which shows larger pitting holes compared with Ac. A F M measurements of mean roughness values (Ra), made over 8 x 8 um areas, were 2  deduced as 37, 61 and 77 nm for the samples M, Ac and Al respectively. Evidence from both A F M and S E M fits the conclusion that the etching attack is more severe in the alkaline treatment used. S E M images suggest that the effect of water storage depends on the initial state of the sample. No significant change is observed for the sizes of the pitting holes after the acidetched sample is stored in water (compare S E M images from samples Ac and AcW in Fig. 5.6((b),(d)), but the pitting holes are larger after storing an alkaline-etched sample in water (Fig. 5.6(c),(e)), and also more aluminum particulates are formed [seen as white structures in Fig. 5.6((d),(e))]. The latter observations are consistent with enhanced oxide dissolution and re-deposition [182] during the water storage for the alkaline-etched sample compared with the acid-etched sample. The greater reactivity detected by S E M for the alkaline-etched sample appears broadly consistent with the thinner oxide layer estimated for sample Al compared with that for Ac (29 A vs. 42 A), and hence the greater passivation in the latter case. Nevertheless,  137  Figure 5.5 A F M micrographs measured for differently pre-treated samples (a) M, (b) Ac, and (c) Al.  30.0 um  Figure 5.6 S E M micrographs measured for aluminum samples given five different treatments (Af, Ac, Al, AcW and AlW) and for the alkaline-etched samples treated with y-GPS [GAl).  XPS has already indicated that the surface region of sample AlW is considerably changed compared with Al; thus Table 5.2 reports that the fraction of OH in the surface region increases markedly from Ac to AcW, as well as from Al to AlW.  5.3.2  Silane interactions  The three types of organosilanes, y-GPS, BTSE and y-APS, were separately applied to five pre-treated aluminum surfaces, as indicated in Table 5.3, which lists elemental composition information from XPS. S E M micrographs show changes in surface morphology (e.g. Fig. 5.6(f)) as a result of the adsorbed silane layers, and these changes are most marked for the alkaline etched surfaces. For example, the open (hole-like) areas seen by S E M in the pretreated sample Al (Fig. 5.6(c)) appear relatively filled after the addition of y-GPS (Fig. 5.6(f)). The presence of silane layers on these samples is confirmed in XPS by the detection of Si 2p peaks (for all organosilanes considered) and N ls peaks for y-APS. Angle-dependent XPS observations for the GM, BM and AM samples show that the C ls and Si 2p signals, as well as N ls for sample AM, are emphasized at low take-off angle (30°), whereas the reverse is seen for the A l 2p signal. Although the samples can have a relatively complex topography, these trends are fully consistent with the silane films being on-top of the surface (and the oxidized aluminum below). After the silane exposures, higher Si 2p (and N ls for y-APS) signals are detected in each case on the aluminum substrates that have been stored in distilled water, or in air, compared with the amounts of adsorption detected on the corresponding samples without the air or water exposure. In particular, XPS indicates greater silane adsorption on the sample designated AcA than on Ac, and similarly more silane adsorption occurs on AlW than on Al; those trends apply for each of the three silanes considered in this work. This fits the situation  140  Table 5.3 Percentage compositions from XPS after different silane treatments.  Sample*  O  C  Al  Si  N  GM  38.4  47.3  9.1  5.2  GAc  36.4  45.6  14.6  3.4  GAl  32.4  55.6  5.4  6.6  GAcA  36.2  45.8  10.6  7.4  GAIW  34 A  43.1  11.2  11.3  BM  40.0  30.8  20.0  9.2  BAc  45.4  24.0  20.1  10.5  BAl  37.7  28.6  26.4  7.3  BAcW  48.4  24.4  15.0  12.2  BAIW  39.0  43.0  0.0  18.0  AM  12.6  66.4  12.5  4.9  3.6  AAc  10.4  65.8  15.0  5.1  3.7  AAl  15.7  59.5  9.7  8.3  6.8  AAcW  17.4  61.4  2.5  9.7  9.0  AAIW  21.4  55.7  2.3  11.6  9.0  The first letter G, B or A abbreviates y-GPS, BTSE or y-APS respectively; the rest of the notation is for the pre-treatment, as in Table 5.1.  141  expected with increasing OH on the surfaces, since then there is greater scope for condensation reactions with hydrolyzed O H groups on the silanes. However, that does not appear to be the only factor since relatively high silane coverages are observed on the alkaline etched surfaces (i.e. samples that received pre-treatment Al) even though its O H content appears to be less than on the mechanically polished and acid etched surfaces (M and Ac respectively). One additional factor is that the alkaline etch appears to be more effective at removing the organic contamination from the top of the surface, thereby possibly aiding the direct bonding process from silane to substrate. A second possibility is that the alkaline etched surface has the rougher topography, and this may in turn give more sites for adsorption and polymerization within the pore and crevice structures formed. N ls spectra from the y-APS films show two types of nitrogen (Fig. 5.7), namely the primary amine group N H , which is hydrogen-bonded to an O H group, and the protonated form 2  NH  + 3  .  The former component is at 400.9 eV binding energy, while the latter is at 402.5 eV  [183,184].  Two additional trends are apparent from observations on the samples with the  different pre-treatments: (i) increasing y-APS adsorption, as shown by larger N ls and Si 2p signals, is accompanied by an increasing N H 2 / N H 3 ratio (Fig. 5.8); and (ii) angle-dependent +  measurements indicate that the  N H  and N H - ? groups are emphasized more at 30° and 90° exit +  2  angles respectively. Both observations are consistent with the protonated form  NH  preferentially deeper down in the sample compared with the situation for the N H  groups, and  2  + 3  being  therefore that the former groups are especially involved in interfacial bonding to the oxide, presumably via H-bonding. Similar conclusions have been reached for the interaction of y-APS with other metals, and this has been termed as an upside-down interaction [46,50,185] insofar as the interfacial bonding is through the N end rather than  142  408  406  404  402  400  398  396  Binding Energy (eV) Figure 5.7 Curve-fitted N ls spectra measured from the sample designated AM (i.e. y-APS added after mechanical polishing) for different takeoff angles: (a) 90°, and (b) 30°.  AM  AAc  AAl  AAcW  AAIW  Sample type  Figure 5.8 Plots of Si and N composition and N H , / N H ratio estimated by XPS (normal take-off angle) for different samples treated with y-APS. +  3  hydrolyzed groups at Si as shown in Fig. 5.9. The lack of strong Si-O-Al interfacial bonding with y-APS may be associated with reports of poorer corrosion resistance accorded aluminum surfaces by coating with y-APS, rather than with other organosilanes like BTSE or y-GPS [43,186].  5.4  Conclusions The objective of this project is to gain more basic information for the adsorption  between hydrolyzed organosilanes and oxidized surfaces of aluminum. Earlier reports have noted how the silane adsorption can be affected by the specific pre-treatment given to the aluminum [169]. This work shows that the physicochemical characteristics of the differently pre-treated surfaces can vary quite markedly, and further that this can in turn affect the subsequent adsorbed coverages. O ls spectra have been used to gauge the amounts of OH present in the surface region of the oxidized films involved.  For surfaces of comparable  roughness, this work is fully consistent with a tendency for the silane adsorption to increase with the number of O H groups on the oxidized aluminum surface.  In turn, that conclusion  helps underpin observations of a more phenomenological nature; for example, Stralin and Hjertberg [187] showed that more robust mechanical adhesion occurs for silane-containing ethylene co-polymers on aluminum surfaces that have been fully hydrated.  145  OH HO  \  NH,  O  H |f 2  NH^  / O H  Si  Si.  /  NH  + 3  NH,+ O  <T  f  ?  Al substrate  Figure 5.9 Al surface.  Illustration of 'up-side interaction' between y-APS and  Chapter 6 Concluding remarks and future directions This thesis attempts to elucidate chemical events occurring at interfaces, especially between polymer and metal, with the view to designing approaches for optimal chemical bonding in relation to technologies associated with polymer light-emitting diodes (PLED) and the passivation of aluminum surfaces. The three following sections identify some new results provided by each part of this research, and suggestions are made for follow-up studies in each case.  6.1  Investigation of some metal-polymer interfaces The work discussed in Chapter 3 focused on thermally deposited metals interacting with  polymer surfaces with relevance to issues concerning the degradation of cathode and polymer interfaces observed in real P L E D devices [29, 37-41]. Three themes were considered. The first concerns the form of the chemical bonding interactions between metal and polymer, and the emphasis was on the systems Zr/PET, Al/PET and Al/PS. Evidence was provided to show that different types of interaction exist in these interfaces, at least for the specific metallic coverages considered in this work. At the relatively low coverages used, the Zr bonds especially to O atoms in the PET surface.  There is no  evidence for the presence of either direct Zr-C bonding, or of the breaking of major bonds in the PET structure, and this situation appears similar to that reported by Gerenser for Ag/PET [92]. The spectral changes on depositing metal are clearly much greater for the Al/PET system. Evidence is presented for the formation of direct bonds of both Al-O and Al-C types,  147  and the latter include interactions with the aromatic ring. Most previous XPS studies for this system have emphasized Al-O-C interactions [81,83,85,86,188], although Novis et al. [82] using HREELS concluded that the interactions were especially of the -COO" A l type, and +  hence more in line with conclusions reached in this work for the Zr/PET system. There has been considerable discussion about whether chemical bonds form between A l and the aromatic ring [80]. Our observations for Al/PET support this, particular through changes in the C ls spectrum, including new structure, and a shift to higher binding energy and enhancement in the shake-up peak. The latter has been predicted by Pireaux to be consistent with a more localized 7t-electron system [80], although it has not been reported in other XPS studies on this system. On the other hand, we do not detect evidence for direct chemical bonding in the Al/PS interface, an observation that appears to emphasize the role of ester groups in strengthening the metallic adhesion to these aromatic structures. The second theme considered the stability of a Mg/PET interface in the presence of a top layer coating of aluminum. Evidence was presented that the use of A l may be able to provide a protective P L E D layer, as an alternative to the more bulky and less flexible glass encapsulation method [98,99]. The third theme considered the effect of heating the Al/PPV interfacial system to 80°C.  This approach was suggested by the report of Friend and co-  workers [115] that the active region of such a P L E D device can experience a temperature rise of as much as 60°C due to current heating during normal operation. However those workers did not clarify specific changes that may be occurring, although there have been many reports on the cathode-polymer interface degradation observed in failed P L E D devices [111-114]. The present study showed substantial changes in A l 2p and O ls spectra for the Al/PPV system as a result of heating to 80°C. Oxidation occurs at this interface, and it was concluded that the  148  oxygen source was in the substrate PPV layer, rather than residual oxygen in the vacuum background. Much follow-up research can be anticipated for the work described in Chapter 3. Our understanding of the metal-polymer interface can at best still only be seen as fragmentary, and it seems clear that a wide arsenal of characterization techniques need to be applied, including more use of vibrational spectroscopy (e.g. HREELS) and microcalorimetry [189-191], for example as a function of metal coverage. These and other methods used in Chapter 3 also need applying to the modified surfaces prepared in Chapter 4. Further work on the Al/Mg/PET system should be directed at fabricating actual devices using the proposed protective coating method, and testing their operating characteristics. The need to reduce the presence of oxygen in Al/PPV systems has been established by this work, and that means new preparation procedures are needed compared with the method used in Section 3.4. The reduction of oxygen content may be achieved by improved annealing schemes or by new preparation procedures [192-194]. It has been established that different conversion conditions give dramatic variations in the efficiencies of devices [195-198].  6.2  Plasma modification of polymer surfaces It has been widely believed that a strengthening in the metal-to-polymer bonding can be  achieved by incorporating adhesion-promoting functional groups at the polymer surface without degrading the structural integrity of its near surface region.  Among the methods  proposed, plasma modification has been of interest, but Burger and Gerenser [143] found that treatment in oxygen plasmas can produced extensive chain scission, leading to a polymer surface that is rich in low molecular weight fragments.  Such fragments are not helpful for  enhanced adhesion to a subsequent added metal layer, and therefore there is a need for gentler  149  surface modification treatments.  Accordingly, in the present work, the use of a remote  hydrogen plasma was investigated with the hope that the presence of oxygen-containing molecules from the background was sufficient to introduce functional groups involving oxygen. This was found to be the case for the three polymer surfaces investigated, namely PS, P B M A and PET. The extent of modification depends on the specific polymer structure, but overall it was relatively small compared to results reported by others using different treatments [160,165167]. Although this work does appear to have achieved a gentler modification of the polymer surfaces, extensions are required in at least two directions. For the first, it is necessary to assess the metallic bonding formed on these surfaces after they have been plasma treated.  That  amounts to combining the approaches used for the work described in Chapters 3 and 4. Second, although it appears that the production of oxygen functionalities may have worked well in the study in Chapter 4, the simultaneous addition of H has been quite significant. At this point it is open to discussion whether the complete loss of phenyl groups in the surface region is in fact advantageous for the subsequent metal adhesion or not. Therefore it may well be worth investigating the use of other less-reactive plasmas, for example the use of helium or neon. They should still allow the addition of oxygen groups as a result of trace involvement by H 0 and 0 molecules in the background gas. 2  6.3  2  Silane treatment on Al surfaces Observations made in Chapter 5 with XPS, A F M and S E M analyses show that surfaces  of aluminum which have been given different pre-treatments, and subsequent exposure to air and water, differ considerably in their physicochemical natures.  Oxide thicknesses, and  fractions of oxygen in O H and oxide forms, are quantified for each sample using Al 2p spectra and O ls data. The use of the latter has been somewhat controversial, especially because of  150  discussions by Sherwood and his co-workers [174,199]. Nevertheless a parallel independent study to ours made by Alexander et al. [179,181] for aluminum alloy and magnetron-sputtered aluminum surfaces used a very similar approach by fitting O ls spectra with hydroxide and oxide components, and their results showed a remarkable congruity with the conclusions presented in this thesis.  This similarity in conclusions only became apparent when Drs.  Mitchell and Sun attended the 2nd International Symposium on Aluminum Surface Science and Technology at Manchester in M a y 2000, after both groups had completed their studies. The specific reason for characterizing the various pre-treated aluminum surfaces was to understand the subsequent silane adsorption phenomena.  The observation with X P S that the  amount of silane adsorption (for all three organosilanes considered) increases directly with the fraction of surface hydroxide (as deduced from O ls spectra) appears fully consistent with the mechanism of silane-metal interaction illustrated in F i g . 1.5.  This work also showed for  comparable surface O H involvement that the amount of organosilane adsorption tends to increase with the roughness of the original aluminum surface as assessed by A F M . The study for organosilane-aluminum interactions requires extension in several ways. First, more direct probes are still needed for the extent of S i - O - A l interfacial bonding in these various systems.  Second, although the present work only used high-purity aluminum, which  can be considered to provide model surfaces, extensions to real alloy surfaces must consider influences introduced by second phase particles and segregating elements.  For example, in  other work done recently in this laboratory a substantial enrichment of copper was formed at a 2024 A l alloy surface after an acid etch pre-treatment [200], and that can greatly influence chemical properties for this surface. However the effect of such segregations on organosilane adsorption are not known, although it is important to investigate. This is particularly since an  151  underlying reason for studying the organosilane-aluminum interactions is to investigate the development of alternative passivating procedures to the currently-used chromating method, which is now generally considered as environmentally hazardous.  Therefore corrosion tests  (e.g. exposure to salt solution, electrochemical polarization measurements) are needed for aluminum surfaces that have been treated with organosilane in order to optimize the S i - O - A l bonding, and to assess the roles of the different parameters (e.g. oxide thickness, amounts of surface impurity, degree of O H involvement) in providing a protective influence.  152  References  1.  P. Ziegler, M.F. Vallat, H. Hirada and J. Schulz, J. Mat. Sci. 32 (1997) 1809.  2.  Y . Misawa, N . Kinjo, M . Hirao, S. Numata and N . Momma, IEEE Trans. Electron Devices 34(1987) 621.  3.  W. Patric, W.S. Mackie, S.P. Beaumont and C.D.W. Wilkinson, J. Vac. Sci. Technol. B 4 (1986) 390.  4.  R. Baigent, N.C. Greenham, J. Griiner, R.N. Marks, R.H. Friend, S.C. Moratti and A . B . Holmes, Synthetic Metals 67 (1994) 3.  5.  S. Wernick, R. Pinner and P.G. Sheasby, The Surface Treatment and Finishing of Aluminum and its Alloys, Finishing Publications Ltd., Teddington, 1987.  6.  H . Leidheiser, Jr. and P.D. Deck, Science 241 (1998) 1176.  7.  S. Nowak, M . Collaud, G. Dietler, P. Groning and L. Schlapbach, J. Vac. Sci. Technol. A 11 (1993)481.  8.  P. Stoyanov, S. Akhter and J.M. White, Surf. Interface Anal. 15 (1990) 509.  9.  R. Haight, R.C. White, B.D. Silverman and P.S. Ho, J. Vac. Sci. Technol. A 6 (1988) 2188.  10.  P. Marcus, C. Hinnen, D. Imbert and J.M. Siffre, Surf. Interface Anal. 19 (1992) 127.  11.  A . Atanasoska, S.G. Anderson, H . M . Meyer III, Z. Lin and J.H. Weaver, J. Vac. Sci. Technol. A 5 (1987) 3325.  12.  M.J. Vasile and B.J. Bachman, J. Vac. Sci. Technol. A 10 (1992) 2992.  13.  T.P. Nguyen and J.L. Mansot, Thin Solid Films 283 (1996) 135.  14.  A.J. Kinlock, Adhesion and Adhesives, Chapman & Hall, London, 1990.  15.  Y . Martin, Ed. Selected Papers on Scanning Probe Microscopes, Design and Applications, SPIE, Bellingham, 1995.  16.  D. Briggs and M.P. Seah, Eds. Practical Surface Analysis by Auger and X-ray Photoelectron Spectroscopy, Wiley, New York, 1990.  153  17.  J.J. Pireaux, Ch. Gregoire, P.A. Thiry, R. Caudano and T.C. Clarke, J. Chem. Phys. 88 (1988) 3353.  18.  L.C. Feldman and J.W. Mayer, Eds. Fundamental of Surface and Thin Film Analysis, North-Holland, New York, 1986.  19.  D.S. Dunn and J.L. Grant, J. Vac. Sci. Technol. A 7(1989) 253.  20.  A. Benninghoven, K.T. F. Janssen, J. Ttimpner and H.W. Werner, Eds. Secondary Ion Mass Spectroscopy SIMS VIII, Wiley, New York, 1992.  21.  K.L. Mittal, Ed. Physicochemical Aspect of Polymer Surfaces, Plenum Press, New York, 1989.  22.  F. Garbassi, M. Morra and E. Occhiello, Eds. Polymer Surfaces-From Physics to Technology, John Wiley & Sons, Chichester, 1994.  23.  D.R. Fitchmun, S. Newman and S. Newman and R. Wiggle, J. Appl. Polymer Sci. 14 (1970) 2441.  24.  J.N. Israelachvili and D. Tabor, Proc. Roy. Soc. A 321 (1972) 435.  25.  B.V. Derjaguin, Research 8 (1955) 70.  26.  C. Weaver, J. Vac. Sci. Technol. A 12 (1975) 18.  27.  L. H. Lee, Ed. Fundamentals of Adhesion, Plenum Press, New York, 1991.  28.  J.H. Burroughes, D.D.C Bradley, A.R. Brown, R.N. Marks, K. Mackay, R.H. Friend, P.L. Burn and A.B. Holmes, Nature 347 (1990) 539.  29.  R.F. Service, Science 267 (1995) 1262.  30.  P.L. Burn, A.B. Holmes, A. Kraft, D.D.C. Bradley, A.R. Brown, R. H. Friend and R.W. Gymer, Nature 356 (1992) 47.  31.  M.C. Lonergan, Science 278 (1997) 2103.  32.  J.C.Scott, Science 278 (1997) 2071.  33.  E.J. Rohling, M. Fenton, F.J. Jorissen, P. Bertrand, G. Ganssen and J.P. Caulet, Nature 394 (1998) 162.  34.  R.K. Kasim, M. Pomerantz and R.L. Elsenbaumer, Chem. Mater. 10 (1998) 235.  35.  R.H. Friend, R.W. Gymer, A.B. Holmes, J.H. Burroughes, R.N. Marks, C. Taliani, D.D.C. Bradley, D.A. Dos Santos, J.L. Bredas, M. Logdlund and W.R. Salaneck, Nature 397 (1999) 121.  154  36.  J.R. Sheats, H . Antoniadis, M . Hueschen, W. Leonard, J. Millter, R. Moon, D. Roitman and A . Stocking, Science 273 (1996) 884.  37.  A . Kraft, A.C. Grimsdale and A . B . Holmes, Angew. Chem. Int. Ed. 37 (1998) 402.  38.  B.H. Compston and K . F. Jensen, Appl. Phys. Lett. 69 (1996), 3941.  39.  V . N . Savvate'ev, A . V . Yakimov, D. Davidov, R . M . Pogreb, R. Neumann and Y . Avny, Appl. Phys. Lett. 71 (1997) 3344.  40.  P.E. Burrows, V . Bulovic, S.R. Forrest, L.S. Sapochak, D . M . McCarty and M.E. Thompson, Appl. Phys. Lett. 65 (1994) 2922.  41.  M . Fujihira, L . Do, A. Koike and H . Han, Appl. Phys. Lett. 68 (1996) 1787.  42.  E.P. Plueddemann, Silane Coupling Agents, Plenum, New York, 1982.  43.  W.J. van Ooij and T. Child, Chemtech. 35 (1998) 26.  44.  S.M. Cohen, Corrosion 57 (1995) 71.  45.  S.G. Hong and F.J. Boerio, Surf. Interf. Anal. 21 (1994) 650.  46.  R.P. Digby and S.J. Shaw, Int. J. Adhesion & Adhesives 18 (1998) 261.  47.  E.P. Plueddemann, in Silanes and Other Coupling Agents, Ed. K . L . Mittal, VSP, Utrecht, 1992, p.3.  48.  P. Walker, in Silanes and Other Coupling Agents, Ed. K . L . Mittal, VSP, Utrecht, 1992, p. 21.  49.  K.W. Allen and M . G . Stevens, J. Adhesion 14 (1982) 137.  50.  F.J. Boerio and J.W. Williams, Appl. Surf. Sci. P (1981) 19.  51.  F.J. Boerio, C A . Gosseilin, R.G. Dillingham and H.W. Liu, J. Adhesion 13 (1993) 159.  52.  J. Ondrus and F. J. Boerio, J. Coll. Interf. Sci. 124 (1998) 349.  53.  R. Chen and F. J. Boerio, J. Adhes. Sci. Tech. 4 (199) 453.  54.  J. D. Miller and I. Ishida, J. Phys. Chem. 86 (1987) 1593.  55.  Y . L . Leung, M . Y . Zhou, P.C. Wong, K.A.R. Mitchell and T. Foster, Appl. Surf. Sci. 59(1992) 23.  56.  Y . L . Leung, Y.P.Yang, P.C. Wong, K.A.R. Mitchell and T. Foster, J. Mater. Sci. 12 (1993) 844.  155  57.  A.J. Pertsin and Y . M . Pashunin, Appl. Surf. Sci. 44 (1990) 171.  58.  J. Fang, B.J. Flinn, Y . L . Leung, P.C. Wong, K.A.R. Mitchell and T. Foster, J. Mater. Sci. Lett. 76(1997) 1675.  59.  M . Getting and A.J. Kinlock, J. Mat. Sci. 12 (1997) 2511.  60.  H . Hertz, Anal. Phys. 31 (1887) 982.  61.  K . Siegbahn, Science 217 (1982) 4555.  62.  K . Siegbahn, C.N. Nording, A . Fahlman, R. Nordberg, K . Hamrin, J. Hedman, G. Hohansson, T. Bermark, S.E. Karlsson, I. Lindgren and B. Lindberg, ESCA: AtomicMolecular and Solid Sate Structure Studied by Means of Electron Spectroscopy, Almqvist and Wiksells, Uppsala, 1960.  63.  E. Bauer, Vacuum 22 (1972) 539.  64.  C D . Wagner, W . M . Riggs, L.E. Davis, J.F. Moulder and G.E. Muilerberg, Eds. Handbook of X-ray Photoelectron Spectroscopy, Perkin-Elmer, Minnesota, 1979.  65.  G. Beamson and D. Briggs, High Resolution XPS of Organic Polymers, Wiley, Chichester, 1992.  66.  M.P. Seah and W.A. Dench, Surf. Interf. Anal. 1 (1979) 2.  67.  G.A. Somorjai, Chemistry in Two Dimensions: Surfaces, Cornell University, New York, 1981.  68.  B.D. Ratner, T.A. Horbett, D. Shuttleworth and H.R. Thomas, J. Colloid Interf. Sci. 53(1981) 630.  69.  Y . L . Leung, Surface Studies of Planar Model HDN Catalysts, PhD. Thesis, University of British Columbia, 1998.  70.  G. Ertl and J. Kuppers, Low Energy Electron and Surface Chemistry, V C H , Weinheim, 1985.  71.  MAX200 Manual, Leybold, Koln, 1986.  72.  D.A. Shirley, Phys.Rev. B 5 (1972) 4709.  73.  H.E. Bishop, Surf. Interf. Anal. 3 (1981) 372.  74.  P.S. Ho, R.H. Haight, R.C. White, B.D. Silverman and F. Faupel, Fundamentals of Adhesion, L - H . Lee, Plenum Press, New York, 1991.  75.  J.H. Burronughes, D.D.C. Bradley, A.R. Brown, R.N. Marks, K . Mackay, R.H. Friend, P.L. Burns and A.B. Holmes, Nature 347 (1990) 539.  156  76.  B . R . Hsieh, E . Ettedgui and Y . Gao, Synthetic Metals 78 (1996) 269.  77.  P. Dannetun, M . Fahlman, C . Fauquet, K . Kaerijima, Y . Sonoda, R. Lazzaroni, J.L. Bredas and W . R . Salaneck, Synthetic Metals 67 (1994) 133.  78.  J.C. Carter, I. G r i z z i , S . K . Heeks, D . J . Lacey, S.G. Latham, P . G . M a y , O.R. de los Panos, K . Piehler, C R . Towns and H.F. Wittman, A p p l . Phys. Lett. 71 (1997) 34.  79.  P . E . Burrows, V . Bulovic, S.R. Forrest, L . S . Sapochak, D . M . McCarty and M . E . Thompson, A p p l . Phys. Lett. 65 (1994) 2922.  80.  J.J. Pireaux, Synthetic Metals 67 (1994) 39.  81.  M . B o u , J . M . Martin, Th.Le Mogne and L . Vovelle,, A p p l . Surf. Sci. 47 (1991) 149.  82.  Y . Novis, N . Degosserie, M . Chtai'b, J.J. Pireaux, R. Caudano, P. Lutgen and G . Feyder, J. Adhesion S c i . Technol. 7 (1993) 699.  83.  A . Calderone, R. Lazzaroni, J . L . Bredas, Q.T. Le and J.J. Pireaux, J. Chem. Phys. 102 (1995)4299.  84.  P . C . Wong, Y . S . L i and K . A . R . Mitchell, A p p l . Surf. Sci. 84 (1995) 245.  85.  S.Akhter, X . L . Z h o u and J . M . White, A p p l . Surf., Sci. 37 (1989) 201.  86.  P. Stoyanov, S. Akhter and J . M . White, Surf. Interface Anal. 15 (1990) 509.  87.  A . Proctor and P . M . A . Sherwood, Anal. Chem. 54 (1982) 13.  88.  S. Tanuma, C J . Powell and D.R. Penn, Surf. Interface Anal. 11 (1988) 577.  89.  A . Proctor and P . M . A . Sherwood, Anal. Chem. 52 (1980) 2315.  90.  P . C . Wong, Y . S . L i and K . A . R . Mitchell, Surf. Rev. Lett. 2 (1995) 297.  91.  M . Textor and R. Grauer, Corrosion Sci. 23 (1983) 41.  92.  L . J . Gerenser, J. Vac. Sci. Technol. A 8 (1990) 3682.  93.  R. Hauert, J. Patscheider, M . Tobler and R. Zehringer, Surf. Sci. 292 (1993) 121.  94.  P. Broms, J. Birgerson and W . R . Salaneck, Synthetic Metals 88 (1997) 255.  95.  J . H . Burronughes, D . D . C . Bradley, A . R . Brown, R . N . Marks, K . Mackay, R . H . Friend, P . L . Burns and A . B . Holmes, Nature 347 (1990) 539.  96.  B . R . Hsieh, E . Ettedgui and Y . Gao, Synthetic Metals 78 (1996) 269.  157  97.  P. Dannetun, M . Fahlman, C . Fauquet, K . Kaerijima, Y . Sonoda, R. Lazzaroni, J.L. Bredas and W . R . Salaneck, Synthetic Metals 67 (1994) 133.  98.  J.C. Carter, I. G r i z z i , S.K. Heeks, D . J . Lacey, S.G. Latham, P . G . M a y , O.R. de los Panos, K . Pichler, C R . Towns and H . F . Wittman, A p p l . Phys. Lett. 71 (1997) 34.  99.  P . E . Burrows, V . Bulovic, S.R. Forrest, L . S . Sapochak, D . M . McCarty and M . E . Thompson, A p p l . Phys. Lett. 65 (1994) 2922.  100.  J . M . Burkstrand, J. A p p l . Phys. 52 (1981) 4795.  101.  R. Hoogewijs, L . Fiermans and J. Vennik, J. Electron Spect. 11 (1977) 171.  102.  J.C. Fuggle, L . M . Watson, D . J . Fabian and S. Affrossman, J. Phys. F 5 (1975) 375.  103.  A . J . Pertsin and Y u . M . Pashunin, A p p l . Surf. Sci. 47 (1991) 115.  104.  X . D . Peng and M . A . Barteau, Surf. Sci. 233 (1990) 283.  105.  P. Steiner, S. Hiifner, N . Martensson and B . Johansoon, Solid State Commun. 37 (1981) 73.  106.  Y . M . Wang, Y . S . L i , P . C Wong and K . A . R . Mitchell, A p p l . Surf. Sci. 72 (1993) 237.  107.  H . Piao and N . S . Mclntyre, Surf. Sci. Lett. 421 (1999) L171.  108.  J . A . Taylor, J. Vac. Sci. Technol. 20 (1982) 751.  109.  C . Lea and J. Ball, A p p l . Surf. Sci., 17 (1984) 344.  110.  X . Zhou, G . E . Thompson, P. Skeldon, G . C W o o d , K . Shimizu and H . Habazaki, Corrosion Sci., 47 (1999) 1599.  111.  H . A z i z and G . X u , J. Phys. Chem. B 101 (1997) 4009.  112.  W . R . Saleneck, S. Stafstrom and J.L. Bredas, Conjugated Polymer Surfaces and Interfaces, Cambridge University Press, Cambridge, 1996.  113.  B . H . Cumpston, I.D. Parker and K . F . Jensen, J. A p p l . Phys. 81 (1997) 3716.  114.  N . V a d i m , A . V . Savvate'ev, D . Yakimov, R . M . Davidov, R. Pogreb, T. Neumann and Y . A v n y , A p p l . Phys. Lett. 77 (1997) 3344.  115.  N . Tessler, N.T.Harrison, D.S.Thomas and R . H . Friend, A p p l . Phys. Lett. 73 (1998) 732.  116.  R . W . Lenz, C - C . Han, J. Sternger-Smith, and F. E . Karasz, Polymer Chemistry 26 (1998) 3241.  158  117.  G . Binning, H . Roher, C . Gerber and E . Weibel, Phys. Rev. Lett. 49 (1982) 57.  118.  G . Binning, C F . Quate and C . Gerber, Phys. Rev. Lett. 56 (1986) 69.  119.  D . Sharid and V . Elings, J. Vac. Sci. Technol. B P (1991) 431.  120.  S. Myhra, in Handbook of Surface and Interface Analysis, Eds. J . C Riviere and S. Myhra, Marcel Dekker, Inc., New York, 1998, p395.  121.  MultiMode SPM Instructional Manual, Digital Instruments, California, 1997.  122.  L B . Beech, International Biodeterioration and Biodegradation 141 (1996) 141.  123.  Application Note, Digital Instrument  124.  D . Susac, Interfaces for Polymer Light Emitting Diodes, M . S c . Thesis, University of British Columbia, 1999.  125.  R . A . Wessling, J.Polym.Sci.Polym. Syn. 72 (1985) 55.  126.  F. Papadimitrakopoulos, K . Konstandinidis, T . M . M i l l e r , R. Opila, E . A . Chandross, and M . E . Galvin, Chem. Mater. 6 (1994) 1563.  127.  K . X i n g , M . Fahlman, M . Logdlund, D . A . dos Santos, V . Parente, R. Lazzaroni, J.L. Bredas, R . W . Gymer and W . R . Salaneck, Adv.Mater. 8 (1996) 971.  128.  F . D . Ettedgui, H . Razafitrimo, K . T . Park, and Y . Gao, A p p l . Phys. 75 (1997) 7536.  129.  P . H . Citrin and G . K . Wertheim. Phys. Rev. B 27 (1983) 3176.  130.  M . G . Mason. Phys. Rev. 5 27(1983) 748.  131.  K . Konstandinidis, F. Papadimitrakopoulos, M . Galvin, and R . L . Opila, J. A p p l . Phys 77(1995) 5642.  132.  F . K . LeGoues, B . D . Silvermans and P.S. Ho, J.Vac.Sci. Technol. A 6 (1988) 2200.  133.  M . Atreya, S. L i , E.T. Kang, K . G . Neoh, Z . H . M a and K . L . Tan, J. Vac.Sci. Technol. A 77(1999) 853.  134.  S. L i , E . T . Kang, K . G . Neoh, Z . H . M a and K . L . Tan, Surf. Sci. 454 (2000) 990.  135.  S. Karg, J.C. Scott, J.R. Salem and M . Angelopoulos, Synth. Met. 80 (1996) 111.  136.  V . H . Tran, V . Massardier, A . Guyot and T.P. Nguyen, Polymer 34 (1992) 3179.  137.  J.R. Rasmusson, P. Broms, J. Birgeson, R. Erlandsson, and W . R . Salanek, Synth. Met. 79(1996) 754.  (http://www.di.com).  159  138.  C M . Chan, T . M . K o , H . Hiraoka, Surf. Sci. Reports 24 (1996) 1.  139.  F . D . Egitto and J.L. Matienzo, I B M J. Res. Develop. 38 (1994) 423.  140.  O. Auciello, A . Gras-Marti, J . A . Valles-Abarca and D . L . Flamm, Eds. PlasmaSurface Interactions and Processing of Materials, Kluwer Academic Publishers, Dordrecht, 1988.  141.  M . Ferreira and M . Mosan, Eds. Microwave Discharges: Fundamentals and Applications, Plenum Press, New York, 1993.  142.  M . A . Uman, Introduction to Plasma Physics, M c G r a w - H i l l , Inc., New York, 1964.  143.  R . W . Burger and L . J . Gerenser, in Metallized Plastics 3: Fundamentals and Applied Aspects, E d . K . L . Mittal, Plenum Press, New York, 1992, p l 7 9 .  144.  N . Inagaki, S. Tasaka and K . H i b i , J. Adhesion Sci. Technol. 8 (1994) 395.  145.  L . J . Gerenser, J. Vac. Sci. Technol. A 6 (1988) 2897.  146.  L . J. Gerenser, J. Vac. Sci. Technol. A 8 (1990) 3682.  147.  Q. T. Le, J. J. Pireaux and J. J. Verbist, Surf. Inter. Anal. 22 (1994) 224.  148.  F . D . Eggito, L . J . Matienzo, K . J . Backwell and A . R . K n o l l , J. Adhesion Sci. Technol. 5(1994)411.  149.  M . Collaud, S. Nowark, O. M . Kiittel, and L . Shclapback, J. Adhesion Sci. Technol. 8 (1994)435.  150.  R. Foerch, N . S. Mclntyre and R. N . S. Sodhi, J. A p p l . Poly. Sci. 40 (1990) 1903.  151.  M . Collaud, S. Nowark, O. M . Kiittel, and L . Shclapback A p p l . Surf. Sci. 72 (1993) 19.  152.  M . Morra, E . Occhiello and F. Garbassi, J. Adhesion Sci. Technol. 7 (1993) 1051.  153.  D . Beamson and D . Briggs, Molecular Physics 76 (1992) 919.  154.  J. Lub, F . C . B . M van Vroonhoven, E . Brunix and A . Benninghoven, Polymer 30 (1989) 40.  155.  B . W . Callen, M . L . Ridge, S. Lahooti, A . W . Neuman and R . N . S . Sodhi, J. Vac. Sci. Technol. A 13 (1995) 2023.  156.  E . Occhiello, M . Morra, P. Cinquina and F. Garbassi, Polymer 33 (1992) 3007.  157.  L . Lianos, D . Parrat, T. Q. Hoc and T. M . Due, J. Vac. Sci. Technol. A 12 (1994) 2491.  160  158.  R. K . Wells, J.P.S. Badyal, L W . Drummond, K . S . Robinson and F.J. Street, Polymer 54(1993) 3611.  159.  T.P. Tepermiesister and Ft. H . Sawin, J. Vac. Sci. Technol. A 10 (1992) 3149.  160.  T . G . Vargo and J . A . Gardella Jr., J. Polymer Sci. 27 (1989) 1267.  161.  D . Briggs, D . G . Ranee, C R . Kendall and A . R . Blythe, Polymer 21 (1990) 98.  162.  R . K . Wells, J.P.S Badyal, L W . Drummond, K . S . Robinson and F.J. Street, J. Adhesion Sci. Technol. 7(1993) 1129.  163.  L . J. Gerenser, J. Adhesion Sci. Technol. 7 (1993) 1019.  164.  W . Kesting, J. Bahner and E . Schollmeyer, J. Polym. Sci. Part B : Polym. Phys. 31 (1993)887.  165.  P. Groning, M . Collaud, G . Dietler and L . Schalapbach, J. A p p l . Phys. 76 (1994) 887.  166.  R. Cueff, G . Baud, J.P. Besse, M . Jacquet and M . Menmalek, J. Adhesion 42 (1993) 249.  167.  J.F. Evans, J . H . Gibson, J.F. Moulder, J.S. Hammond and H . Goretzki, Fresenius Z . Anal. Chem. 379(1984) 319.  168.  J . A . Treverton and M . P . Thomas, Int. J. Adhesion & Adhesives 9 (1989) 211.  169.  R . P . Digby and D . E . Packham, Int. J. Adhesion & Adhesives 75 (1995) 61.  170.  N . S . Mclntyre, R . D . Davidson, L Z . Hyder and A . M . Brennenstuhl, in Handbook of Surface and Interface Analysis, Eds. J.C. Riviere and S. Myhra, Marcel Dekker Inc., New York, 1998, p648.  171.  Instruction Manual for S-4I00 Field Emission Scanning Microscope, Hitachi Ltd., Tokyo, 1991.  172.  B . R . Strohmeier, Surf. Interface Anal. 75 (1990) 51.  173.  M . Amstutz and M . Textor, Surf. Interface Anal. 79 (1992) 595.  174.  S. Thomas and P . M . A . Sherwood, Anal. Chem. 64 (1992) 2488.  175.  C . Ocal, B . Basurco and S. Ferrer, Surf. Sci. 757 (1985) 233.  176.  A . Fritsch and P. Legare, Surf. Sci. 756 (1987) 247.  177.  T. Ronnhult, U . Rilby and I. Olefjord, Mat. Sci. Eng. 42 (1980) 329.  178.  C . Chen, S.J. Splinter, T. D o and N . S . Mclntyre, Surf. Sci. 382 (1997) L652.  161  179.  M . R . Alexander, G . E . Thompson and G . Beamson, Surf. Interface Anal. 29 (2000) 468.  180.  P.R. Underhill, G . Goring and D . L . DuQuesnay, Int. J. Adhesion & Adhesives 18 (1998) 307.  181.  M . R . Alexander, G . E . Thompson and G . Beamson, to be published in Proceedings of 2nd International Symposium on Aluminum Surface Science and Technology, Manchester, M a y 2000.  182.  W . Vedder and D . A . Vermilyea, Trans. Faraday Soc. 65 (1969) 561.  183.  R . G . Dillingham and F.J. Boerio, J. Adhesion Sc. Technol. 6 (1992) 207.  184.  M . L . A b e l , J.F. Watts and R . P . Digby, Int. J. Adhesion & Adhesives 18 (1998) 179.  185.  M . R . Horner, F.J. Boerio and H . M . Clearfield, in Silanes and Other Coupling Agents, Ed. K . L . Mittal, V S P , Utrecht, 1992, p. 241.  186.  L . J . Matienzo, D . K . Shaffer, W . C . Moshier and G . D . Davis, J. Mat. Sci. 21 (1986) 1601.  187.  A . Stralin and T. Hjertberg, J. Adhesion 41 (1993) 51.  188.  J . M . Burkstrand, J. A p p l . Phys. 52 (1981) 4795.  189.  H . Ibach and D . L . M i l l s , Electron Energy Loss Spectroscopy and Surface Vibration, Academic Press, 1982.  190.  J.J. Pireaux, C h . Gregoire, P . A . Thiry, R. Caudano and T . C . Clarke, J. Chem. Phys., 55(1988) 3353.  191.  R. J. Murdey and J. T. Stuckless, Proc. of the Soc. Plastics Eng., Ann. Tech. Conf. 57 (1999) 2207.  192.  M . Ferreira and M . F . Rubner, Macromolecules 25 (1995) 7107.  193.  A . C . Fou, O . Onitsuka, M . Ferreira, M . F . Rubner and B . R . Hsieh, J. A p p l . Phys. 79 (1996) 7501.  194.  J.W. Baur, S. K i m and M . F . Rubner, Advanced Material 10 (1998) 1452.  195.  C . Zhang, D . Braum and A . J . Heeger, J. A p p l . Phys. 73 (1993) 5177.  196.  M . Herold, J. Gemeiner and M . Schwoerer, Act Polym. 47 (1996) 436.  197.  J. Morgado, F. Cacialli, J. Gruner, N . C . Greenham and R . H . Friend, J. A p p l . Phys. 55 (1999) 1784.  162  198.  J. Morgado, R . H . Friend and F. Cacialli, Synthetic Metals 114 (2000) 189.  199.  P . M . Sherwood, Surf. S c i . Spectra 5 (1998) 1.  200.  X . Sun, W . H . K o k , K . C . Wong, R. L i , K . A . R . Mitchell and T. Foster, Proceedings of 2nd International Symposium on Aluminum Surface Science and Technology ( A S S T 2000), in press.  201.  L . J . Rothberg, M . Y a n , F. PapadimitrakopOulos, M . E . Galvin, E . W . Kwock and T . M . Miller, Synthetic Metals 80 (1996) 41.  202.  L . J . Rothberg and A . J . Lovinger, J. Mater.Res. 11 (1996) 3174.  163  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0061347/manifest

Comment

Related Items