Effect of thickness on the physical, thermal and chemical properties of ultra thin polymer films

Acknowledgements Gratitude goes to my two supervisors, Dr Low Hong Yee from Institute of Materials Research and Engineering (IMRE) and Asst Prof Chin Wee Shong from Department of Chemistry, National University of Singapore (NUS). They have patiently guided and imparted invaluable knowledge and advice to me during the past 2 years. Many thanks to Dr Low Hong Yee’s team members especially to Ms Loh Wei Wei, Mr Xu Yongan and Mr Huang Hongliang for the technical assistance and making my work in the laboratory enjoyable. They never fail to render help whenever required. Also to Dr Chin Wee Shong’s team members, Ms Lim Wen Pei, Mr Kerk Wai Tat, Ms Yin Fenfang, Ms Xu Hairuo, Mr Neo Min Shern and Ms Liu Chenmin for offering helpful advice during our regular group meetings and making group meetings and gatherings fun. Appreciation also to many IMRE research staff who had assisted me in the use of the various equipment: Mr Lim Poh Chong for use of XRD, Mr Zheng Yuebing for AFM and ellipsometry, Dr Pan Ji Sheng for XPS, Ms Doreen Lai for TOF-SIMS and Ms Tan Li Wei for contact angle measurements; also to Dr Yang Ping from Singapore Synchrotron Light Source (SSLS) for her help in the density measurements using the XSF. My MSc stint in IMRE and NUS will not have been possible without the constant support and encouragement from family, friends and fellow IMRE colleagues. Lastly, I thank IMRE and NUS for providing me a Graduate Research Scholarship and giving me an opportunity to learn and obtain a higher degree. 1 Abstract Polymer ultra-thin films, defined as films with thickness 0, it corresponds to strong interactions.19-20, 24 Several authors proposed models of Tg behaviour in which the films are considered to consist of three layers.18, 25-27 Thus the dynamics of the material and the Tg of each layer are postulated to be different. In the first layer (near the free surface), the chain mobility is greater than in the bulk polymer; in the second layer (in the centre of the film), the chain mobility is equivalent to the bulk polymer, while in the third layer (at the solid interface), the mobility is restricted relative to the bulk polymer. With an increase in Tg when thickness decrease, there is reduced mobility of 6 the polymer chains as a result of attractive forces at the interface.11 In a free standing film with the same dimensions, Tg was depressed by ~70°C as compared to that of supported film of the same thickness where Tg was depressed by ~20°C. This clearly depicts the role of interfacial interactions.28 Study on Tg of thin poly(2-vinyl pyridine) (P2VP) film coated on silicon wafer also revealed Tg increased with decreasing film thickness. A stronger interaction between the polymer and native oxide surface of silicon wafer was expected as compared to PMMA with the same substrate due to the chemistry of P2VP.29 Fryer et al 11 showed the dependence of Tg of polymer films on interfacial energy and thickness. They reported that the glass transition temperatures of ultra-thin films of polystyrene and poly(methylmethacrylate) depend strongly on the thickness and interfacial energy between the polymer film and the substrate. The substrates were first treated with stable self-assembled films of octadecyltrichlorosilane (OTS) on silicon wafers. The interfacial energy between the substrates and polystyrene or PMMA was tuned by exposing it to different X-ray doses of exposure in the presence of air. Exposure to X-ray radiation modified OTS by producing oxygen containing groups on the surface and this interfacial energy for both polymers increases with increased dosage. It was observed that at low values of interfacial energy, Tg of polymer films was less than bulk value and a reverse phenomenon was observed for that of high values of interfacial energy. They interpreted their results in terms of the layer model. High interfacial energy resulted in decreased segmental mobility and thus an increase in Tg.11 Results of molecular dynamic simulations can give further insight to the decrease in segmental mobility of polymer chains with increasing interfacial energy. Torres et al. represented polymer molecules in their simulations using square-well 7 interaction sites interconnected by fully flexible strings.30 For supported films, an attractive wall represented the substrate. If the attractive potential, ε, between the interaction sites of the polymer chains and wall was the same as the attractive potential between polymer-polymer sites, the wall was considered as weakly attractive and Tg will decrease compared to the bulk. If ε between the interaction sites of the polymer chains and wall was doubled, then the wall was considered as strongly attractive and Tg increased as compared to bulk polymer. Thus an analysis of total mean displacement of segments as a function of temperature and position in the film shows the mobility of the polymer near the substrate decreased with increasing ε. With experimental results, layer models and molecular simulations, it is concluded that the mobility of polymer segments near the substrate decreases as the interfacial energy increases and the nature of substrate-polymer interface dominates the Tg behaviour of ultra-thin films.11 1.2 Effect of Thickness on Moisture Uptake With chain mobility being affected by chain confinement effects3 and hence having an effect on the Tg of the polymer film, we expect similar effect when small molecule diffuses inside the polymer film. Moisture uptake and swelling of the polymers can lead to significant reliability problems. Presence of small amounts of water in polymer thin films can affect a variety of thin film physical properties. Examples are changes in mechanical properties such as tensile stress or hardness, electronic properties, as well as chemical and processing characteristics of polymer thin films in applications. Presence or lack of water in photoresists during the fabrication of intergrated circuits can change the physical and chemical properties as well as affect the imaging characteristics31-32. Small amount of water in the 8 photoresist film can affect the reaction pathways responsible for solubility changes that permit lithography imaging of materials. Too little water in the photoresist during exposure can also lead to considerable side reactions.31 There are many methods to measure the water sorption in polymer films. These methods include gas permeation techniques, electro-microbalances, quartz spring microbalance, FTIR and stress analyzer.31 In the present work, we will study the water uptake as a function of film thickness on gold coated and silicon oxide coated QCM crystals. QCM was chosen as the technique used as it can accurately detect mass uptake in the order of nano grams. The linear relationship between small mass uptake (∆m) on the coated crystal and quartz crystal frequency shift (∆f) is described by the Sauerbrey’s equation in Equation 1.2: ∆m = K∆f Equation 1.2 where K is the proportionality constant which incorporates known properties of the quartz crystal.3, 31, 33-34 1.3 Effect of Thickness on Immobilization Capabilities Surface modification of polymers can be achieved by introducing functional groups that allow the buildup of polyelectrolyte multilayers via layer-by-layer assembly.35-39 Margarita et al 39 reported a method for surface modification of hydrophobic substrates through the absorption of poly(vinyl alcohol). This increased the wettability of the substrates and hence it serves as a platform for other chemistries. They also reported the CVD polymerization of ethyl cyanoacrylate to form ultra-thin poly(ethyl cyanoacrylate) coatings. There is a need to understand the chemistry behind the polymer by determining the orientation and the morphology of the functional groups. From there, we will be able to functionalise our polymer to serve as 9 a platform for other applications such as sensors. 1.4 Scope of work Due to the impending importance of ultra-thin polymer films being utilized in various applications, it becomes necessary to determine how thermal, physical and chemical properties will change with thickness. In this work, we will focus on 2 classes of polymer thin films, polyetherimide and parylenes, prepared using two different deposition processes, spin coating and chemical vapour deposition (CVD) polymerization respectively. We will investigate the thickness effects on the thermal (glass transition temperature, Tg, changes and effects of annealing), physical (moisture sorption studies) and chemical (immobilization capabilities) properties of these ultra-thin films. In general, polyimides are important materials for the electronics industry due to their thermal stability, high chemical resistance characteristics and excellent mechanical toughness. However, polyimides are known to absorb water in small amounts. Water absorbed in polyimide films cause metal corrosion, package cracking, delamination, failures of adhesion to metals and degradation of dielectric properties. The dimensions of the films will also be affected by water due to swelling. Thus the moisture sorption behaviour of polyimide films is to be investigated as it has effects with regard to the reliability and performance of electronic devices fabricated with it. It also becomes important to investigate the moisture sorption behaviour with respect to changes in thickness and substrate influence.10-12 In Chapter 3 of this work, a specific type of polyimides, polyetherimide (PEI), (or also known as ULTEM, with its chemical structure as shown in Figure 1.1) is selected for studied. PEI has high heat-distortion temperature, tensile strength and 10 modulus. It is often used in high-performance electrical and electronic parts, microwave appliances and automotive parts.10 O O N N CH3 O O O O CH3 n Figure 1.1 Chemical structure of PEI The polymer thin films studied in Chapter 4 of this work belong to the parylenes family. Parylene C, with its structure shown in Figure 1.2, is a hydrophobic polymer and is optically transparent. It has excellent mechanical, electrical, thermal and biomedical properties and can be synthesized by CVD polymerization. It has been used as an insulator for electrical passivation in preparation of devices for protein detection. 40 Cl H2C CH2 n Figure 1.2 Chemical structure of parylene C In our study, parylene C is deposited on a substrate via CVD polymerization while PEI is deposited via spin-coating. There are pros and cons with each of these deposition processes. Spin coating is a well-established method for preparing smooth polymeric coatings on flat substrates. This technique is widely used in the microelectronics 11 industry. The polymer is first dissolved in a suitable solvent and applied onto a substrate. By the rotation of the substrate at high speed, excess solution is ejected, leaving a thin film which continues to flow radially outwards by the action of the centrifugal force. As the film thins down, the solvent evaporates. The parameters affecting film thickness are spin speed, volatility of solvent and initial polymer concentration.41 Spin coating is very easy to use, and the cost of equipment is low. It is relatively easy to control the film thickness by varying parameters mentioned above. One disadvantage of spin coating is it requires the use of solvents and the choice of solvent often must fulfill the following considerations: environmental issues, miscibility, and its effects on the film properties. Chemical vapour deposition (CVD) polymerization has gained substantial interest in recent years as it forms polymer films in the absence of solvent and produces conformal pinhole-free coatings. 2,2-Para-cyclophanes are examples of CVD precursors for thin film polymers, commonly known as parylenes. Parylenes are used in a wide range of applications such as automotive, medical, electronics and semiconductor industries. Parylene coatings are inert and transparent and have excellent barrier properties. CVD polymerization to yield parylene is an inherently clean process as the monomer gas is directly converted into polymer without the need for initiator or catalysts and produces linear high molecular weight polymer. Parylene is a semicrystalline polymer with degrees of crystallinity and crystalline modifications that are dependent on the deposition conditions.39, 42 The steps of CVD deposition of parylene consist of 1) the sublimation of dimer in a sublimation furnace, 2) cracking the dimer into monomer in the pyrolysis furnace, 3) transportation of the monomer into the deposition chamber, 4) diffusion of monomer from the region above the substrate through a boundary layer, 5) adsorption 12 of monomer into the substrate, 6) surface migration and bulk diffusion of monomer, and 7) chemical reaction that comprises propagation or initiation. There is no termination reaction. During steady state growth, the density of radical chain ends on the surface remains constant as the new radicals generated by initiation replace those that are buried in the growing film.43 The advantages of CVD polymerization include the following: it is a solventless process; can form structurally continuous, pinhole free and uniform films; it can provide homogenous coating deposited simultaneously on flat surfaces, inside, outside and in corners of deep crevices. However, the initial set up cost of CVD equipment is much higher than for other types of deposition process.40 In Chapter 5, we will investigate the surface modification of parylene with an attached amino group. The chemical structure as shown in Figure 1.3 will be addressed as amino-parylene in this thesis. Usually amino-terminated molecular films are used in the immobilization of enzymes, DNA, and in initiating graft polymerization etc.44 In these applications, it is important that there is a sufficiently high content of the reactive primary moieties exposed on the substrate surface so that there is interaction with other molecules. Zhang et al 44 reported that there were different immobilization capacities of the different aminosilane substrates for pyromellitic dianhydride (PMDA). It was observed that the higher primary amine content favoured a higher uptake of PMDA. They also reported that primary amine content could be a measure of the film morphology and accessibility of the substrate amine groups.44 In this work, we examine the influence of the thickness of the amino terminated parylene films on their capacity for immobilization of pyromellitic dianhydride (PMDA). 13 H2C CH2 CH2 n CH2 CH2 m NH2 Figure 1.3 Structure of amino-parylene 14 Chapter 2 2.1 Characterization Techniques Ellipsometry Ellipsometry is a very sensitive measurement technique that uses polarized light to characterize thin films, surfaces, and material microstructures. It derives its sensitivity from the determination of the relative phase change in a beam of reflected polarized light. Figure 2.1 below illustrates the basic principle of ellipsometry. First, there is an incoming polarized light. The incident beam and the direction normal to the surface define a plane that is perpendicular to the surface which is known as the plane of incidence. The interaction of the light with the sample causes a polarization change in the light, from linear to elliptical polarization. The change in the shape of the polarization is then measured by analyzing the light reflected from the sample. In Figure 2.1, it shows that the amplitude of the electric wave which is in the plane of the incidence as Ep and the amplitude of the electric wave which is perpendicular to the plane of incidence as Es. These are also referred to as the p-waves and s-waves respectively. Figure 2.1 Schematic diagram of the principle of ellipsometry45 Ellipsometry measures two values, Ψ and ∆, which describes the polarization 15 change. Ψ is the relative phase difference of the polarizing light and ∆ is the relative amplitude change. These values are related to the ratio of Fresnel reflection coefficients, Rp and Rs for p- and s- polarized light, respectively. p = tan(Ψ )e i∆ = Rp Eq. 2.1 Rs where ψ is the angle whose tangent is the ratio of the magnitudes of the total reflectance coefficient ( ratio of the outgoing wave amplitude to the incoming amplitude) and p is the complex ratio of the total reflection coefficient. As ellipsometry measures the ratio of two values, it can be highly accurate and very reproducible. From measured quantities of Ψ and ∆, the thickness of the film can be derived by a model fitting. The most commonly used approach to obtain film thickness for transparent material is the Cauchy model. The Cauchy dispersion relation is an inverse power series containing only even terms: n(λ ) = A + B λ 2 + C λ4 + ... Eq. 2.2 where the wavelength λ is given in microns, n is the refractive index and A, B and C are the fit parameters.45 Besides determining the thickness of the film, ellipsometry was used in this project to determine the Tg of the polymer films. 2.1.1 Measurement of Film Thickness Film thicknesses analysis was performed using a Variable Angle Spectroscopic Ellipsometer VASE (J.A.Woollam Lincoln, NE). The Ψ and ∆ data at angles 70° and 75° over wavelength range 500 to 1000nm were fitted using the Cauchy model. Prior to measurements, the film thickness was first estimated by a KLA Tencor-P10 surface profilometer. 16 The fitting of data was carried out by first assuming that the first layer or the substrate layer is 0.6mm Si substrate while the next layer was included as Cauchy layer. From the film thickness obtained from surface profilometer, an estimated value of the thickness was entered. The values of A=1.7, B=0.001 and C=0.0001 which were the values usually used for polymer films were used. Both n and k (extinction coefficient) were first fitted. Once a good fit was obtained, n and k were then fixed and the values of thickness, A, B and C were fitted to obtain the more accurate values. 2.1.2 Measurement of Glass Transition Temperature (Tg) Tg and temperature/thickness dependence measurements were performed by placing the supported film on a Linkam TMS 94 heating/cooling stage. The ellipsometric angles (Ψ and ∆) were continuously recorded at 120s intervals. The samples were heated and cooled at a constant rate of 2°C/min. Tg was determined from the intersection of the best fit of 2 straight lines in the thickness versus temperature curve. 2.2 Quartz Crystal Microbalance Quartz crystal microbalance (QCM) is widely used in many applications in the measurement of small masses due to their stability, simplicity of measurement, high precision, high sensitivity and ease of analysis. The frequency change relationship between rigid layers firmly attached to QCM is proportional to the added mass as long as the added mass behaves elastically similar to the quartz crystal itself. The relationship is given in Sauerbrey equation as shown in Equation 2.3: 17 madded = ( f uncoated − f measured ) Cf Eq. 2.3 where madded is film mass per unit area, funcoated and fmeasured are the resonance frequencies of the bare crystal and crystal coated with film (dry), respectively and Cf is a constant determined by the crystal used where it is calculated by Equation 2.4: Cf = 2 f q2 ( ρ qν q ) Eq. 2.4 Here, fq is the resonant frequency of the bare crystal, ρ q is the density of the quartz crystal (2.649gcm-3) and ν q is the shear velocity of the AT cut quartz crystals (332200cms-1). It is important that the Sauerbrey equation is only valid for thin films that can be considered rigid masses. There are other assumptions that are necessary for this expression to hold; e.g. the added mass must be evenly distributed over the electrode, the added mass must be much less than the mass of the quartz crystal itself and the mass is rigidly attached with no deformation from oscillatory motion of the crystal.31, 46 The films prepared for this study meet the above requirements. 2.2.1 Measurement of Mass Change A Maxtek research grade quartz crystal microbalance (RQCM) (PLO-10 phase lock oscillator, 5MHz AT cut, Cr/Au polished quartz crystal, and 0.4cm2 active area) was used to determine mass change. The various thicknesses of polymer were coated on the quartz crystal and their mass changes were measured at room temperature. For the study of mass changes as a function of temperature, the Parylene-C coated crystal was allowed to reach equilibrium initially at room temperature. The crystal resonance frequency was recorded at a rate of 2/min. After which the coated crystal was heated on a Linkam TMS heating/cooling stage at 115°C at various time intervals before measuring the frequency. The frequency shift was converted into 18 mass using Equation 2.3. 2.2.2 Measurement of Moisture Adsorption Polyetherimide was spin-coated on the QCM crystal to obtain the desired thickness. Initially, the polyetherimide coated crystal was allowed to reach equilibrium in a low humidity chamber (Relative Humidity, RH: 20% ± 1%). The sample was then transferred immediately to a high humidity chamber (RH: 95% ± 1%). The whole moisture sorption experiment was carried out at a constant temperature of 25 ± 0.1°C. The crystal resonance frequency was recorded at a rate of 2/min. The frequency shift was converted into mass by Equation 2.3. The percentage of moisture absorbed at steady state can be calculated by Equation 2.5. Moisture( wt %) = 2.3 f dryfilm − f wetfilm f uncoated − f dryfilm × 100% Eq. 2.5 Fourier Transform Infrared (FTIR) Spectroscopy Infrared spectroscopy measures the vibrations of molecules. Each functional group, or structural characteristic, of a molecule has a unique vibrational frequency. The result is a unique molecular "fingerprint" that can be used to confirm the identity of a sample.47 FTIR spectrum was obtained on a Perkin Elmer FTIR spectrometer 2000 using a KBr disc with the respective thicknesses of polymer coated on it. A transmission mode configuration was employed using 32 scans at a resolution of 4cm-1. 2.4 X-ray Photoelectron Spectroscopy (XPS) Electron spectroscopy techniques measure the kinetic energy of electrons that 19 are emitted from matter as a consequence of bombarding it with ionizing radiation or high energy particles. The simplest is the direct ionization of an electron from a valence or inner shell. The kinetic energy, KE, of the ionized electron is equal to the difference between the energy of the incident radiation, hν, and the binding energy or ionization potential, BE, of the electron. This is illustrated in the equation: KE = hν BE. For a given atom, a range of BE values is possible, corresponding to the ionization of electrons from the different inner and outer valence shells and these BE values are characteristic for each element. Measurement of KE, and hence BE values, provides a means of identifying the atoms. XPS has been employed to be a powerful technique for determining the energy levels in atoms and molecules. It has been used to probe the chemical shift of the atom relative to the original molecule and hence obtain information of the structure. This is due to the variation of the binding energies of electrons in a particular atom due to the immediate environment of the atom and its charge or oxidation state. The principal use is for studying surfaces as it is surface sensitive as it probes at the top 2-5nm of the surface. It can be used as an analytical method for detecting the elements (and functional groups) on the surface. 47 XPS measurements were made on a VG Scientific ESCA-LAB-220i XL. The core level signals were obtained at a takeoff angle of 90° with respect to the sample surface. All binding energies (BE) were referenced to the C1s hydrocarbon peak at 285eV in order to compensate for the surface charge effects. The spectra were fitted using the Advantage software and the surface elemental stoichiometries were determined from the fitted peak area ratios. 20 2.5 Time-of-Flight Secondary Ion Mass Spectrometry (TOF-SIMS) Sputtering and ionization in SIMS are due to events caused by the impact of a high velocity ion on a surface. This process is shown schematically in Figure 5. A primary ion (Ar+) strikes the surface, r, with high energy. And this impact of the primary ion causes sputtering of atoms and molecules from a film surface. These sputtered particles include electrons, positive and negative ions. The secondary ions sputtered from the surface are collected by a mass spectrometer and mass analysed. Figure 2.2 Schematic diagram of SIMS sputtering process48 All of the ion formation processes in SIMS are known for other forms of mass spectrometry. The difference is that several ion formation mechanisms may occur simultaneously. The dominant ionization process may vary with the type of polymer involved. One advantage of SIMS is that both positive and negative ions are formed, often in comparable yields. Molecules having low ionization potentials will tend to form positive ions and those with high electron affinities will form negative ions.48 21 In our experiment, TOF-SIMS was performed as a confirmation characterization tool with XPS as it is also a surface sensitive technique. The measurements were performed on ION TOF SIMS IV. The polymer films were coated on Si wafers. The samples were then analysed wth 25keV Ga+, rastered at 500 µm by 500 µm. Both positive and negative polarity spectrums were acquired at 600s. 2.6 X-ray Diffraction (XRD) XRD has been in use for the fingerprint characterization of crystalline materials and for the determination of their crystal structures. X-rays are electromagnetic radiation of wavelength about 1Å. X-rays are produced when high energy charged particles collide with matter. The electrons are then slowed down or stopped by the collision and some of the lost energy is converted into electromagnetic radiation. X-ray wavelength used is emitted by copper Kα, λ=1.5418 Å. Each crystalline phase has a characteristic pattern that can be used as a fingerprint. The two variables are peak position (d-spacing) and intensity. In this experiment, we will determine the peak position and the intensity of the semicrystalline parylene film. 47 XRD of the amino parylene films was measured using a Bruker GADDS diffractometer with CuKα radiation and a graphite monochromator (the accelerating voltage: 40kV; applied current: 40mA). The measurements were carried out at room temperature with the following parameters: scan range: 6~33°; scan time: 30min; incident angle: 1°. Distance from the X-ray source to sample was 15cm using a 0.5mm double pinhole collimator. 22 2.7 Atomic Force Microscopy (AFM) In AFM, the force sensing spring consists of a miniturised cantilever beam clamped at one end and the probing tip at the other end. The atomic force microscope (AFM) probes the surface of a sample with a sharp tip, a couple of microns long and often less than 100Å in diameter. The tip is located at the free end of a cantilever that is 100 to 200µm long. Forces between the tip and the sample surface cause the cantilever to bend, or deflect. A detector measures the cantilever deflection as the tip is scanned over the sample, or the sample is scanned under the tip. The measured cantilever deflections allow a computer to generate a map of surface topography.49 AFM measurements were performed using an AFM Multimode instrument, Digital Instruments, USA. Height images were acquired under ambient conditions in tapping mode using a 0.01-0.025 Ohm-cm Antimony (n) doped Si; cantilever tip, FIB2-100S (source of the tip). 2.8 Contact Angle Analysis Contact angle, θ, is a quantitative measure of the wetting of a solid by a liquid. It is probably the most common method of solid liquid tension. It is defined geometrically as the angle formed by a liquid at the three phase boundary where a liquid, gas and solid intersect as shown in Figure 2.3. The drop of liquid that is put on the solid surface will modify the shape under the pressure of interfacial tensions. It can be seen from this figure that low values of θ indicate that the liquid spreads, or wets well, while high values indicate poor wetting. Typically, if the angle θ is less than 90 the liquid is said to wet the solid. If it is greater than 90 it is said to be nonwetting. A zero contact angle represents complete wetting. 50 23 Figure 2.3 Contact angle measurement 50 Contact angle measurements were taken on Rame-Hart contact angle goniometer to measure the surface hydrophobicity of the samples before and after modification. The system was first calibrated e to ensure the planarity of the sample stage. Then the focal length was adjusted to ensure that the sample was focused. A drop of 3 µl of deionised water was dropped onto the sample and the contact angle was measured immediately. This step was repeated for at least 5 times to obtain the average contact angle value of the sample. 2.9 X-ray Specular Reflectometry (XSF) High resolution X-ray Specular Reflectometry was used to measure film densities. This measurement was carried by Dr.Yang Ping at the X-ray demonstration and development (XDD) beamline at Singapore Synchrotron Light Source (SSLS). The diffractometer is the Huber 4-circle system 90000-0216/0, with high-precision 0.0001° step size for omega and two-theta circles. The storage ring, Helios 2, was running at 700 MeV, typically stored electron beam current of 300 mA. X-ray beam was conditioned to select CuKα1 radiation (8.048 keV in photon energy) by a Si (111) channel-cut monochromator (CCM) and blocked to be 0.80mm high in vertical 24 direction and 3.50mm wide in horizontal direction by a slit system. Such set-up yielded x-ray beam with about 0.01° in divergence. The detector slit was adjusted to be 1.00mm high to ensure recording all reflected photons. The typical counting time was 5 second for every step and step size of theta varies from 0.02 to 0.0025° for different samples to ensure that the oscillation in reflectivity was well recorded. Diffuse scattering at rocking scan was also measured at some chosen 2-theta in the range of above measurement. As it is much weaker, there is no need to correct raw specular reflectivity by subtracting the diffuse scattering. The simulations were done using simulating software M805 and LEPTOS 1.07 release 2004 (Bruker). The critical angles for parylene layer and Si-substrate (0.223°) are well fitted, the former indicating the density of parylene layer directly. LevenbergMarquardt algorithm for least-square refinement on logarithm of data can be done for all samples. Layer parameters are listed in above table. The trends and oscillating finer structure in the reflectivity were also fitted well. Final χ2-values are below 2.72G10-2. The trends and oscillating finer structure in the reflectivity were fitted well. The layer densities were also obtained from full-profile fitting and can also be seen obviously from the critical angles. Native SiO2 layers between Si-substrates and parylene layers do not play roles or can be described as the interface roughness. 25 Chapter 3 Glass Transition Temperature and Moisture Sorption Studies of Polyetherimide Ultra-thin Films With miniaturization of devices being inevitable in today’s context with the evolution in nanotechnology, thickness of protective films and coatings used in these devices are shrinking in dimensions as well. Polyimides are widely used in various industries such as in aerospace applications as wire and cable insulation, and in electronic industry for flexible printed boards due to their excellent thermal and dielectric properties. As the film thickness reduces to sub-nanometer scale, the competing effects of free surface and surface confinement play important roles in the overall properties of the film. One such property is the glass transition temperature (Tg). Tg is often used as an indicator of the thermal stability of a polymer and has been a useful material parameter for a number of applications. One of the constant concerns in PEI film used for microelectronic application is the moisture absorption since PEI is a relatively polar polymer. While the moisture uptake of bulk PEI films is relatively well studied, few reports are available on the moisture uptake of ultra thin films. The relationship between thickness dependence on Tg and the diffusion of small molecules such as water has been linked to polymer chain dynamics. Diffusion coefficient of dye molecules in PS film has been reported to decrease as the film thickness decreases in conjunction with a decrease in the Tg.51 Tan et al52 evaluated the moisture absorption of thin polyimide films (< 100nm) on silicon substrates by measuring the dimensional changes in humid environment. It was reported that the hydrophilic substrate was responsible for the higher moisture induced swelling detected by x-ray reflectivity measurement. In this work we investigate the moisture uptake of a series of polyimide on both a hydrophobic and a hydrophilic substrate. We will also investigate 26 the changes in Tg and moisture sorption behaviour of polyetherimide films with respect to different thickness and on different substrates. 3.1 Experimental Section 3.1.1 Materials Ultem was purchased from Sigma Aldrich. Chloroform was purchased from Tedia and used as received. 3.1.2 Substrate Preparation Silicon wafers were cut into 1cm x 1cm pieces. They were first cleaned with a solution of deionised water, H2O2 and NH4OH (6:1:1 volume ratio) at 70°C for 10min. They were rinsed with deionised water and then cleaned with a solution mixture of deionised water, H2O2 and concentrated HCl (6:1:1 volume ratio) at 70°C for another 10min. The wafers were then rinsed copiously with deionised water and blown dry with nitrogen. 3.1.3 Film Preparation Various concentrations of Ultem in chloroform were prepared to obtain the various desired film thicknesses. The film thickness were determined with the surface profilometer and confirmed with ellipsometry, fitted with the Cauchy model. Films of different thicknesses were obtained from spin coating at 2000rpm using a Laurell spin coater on clean silicon wafers or quartz crystals. Figure 3.1 showed the graph of the concentrations of solutions used to derive the desired thicknesses on silicon wafer. The thickness obtained for quartz crystal was taken to be the same as that for the films on silicon wafer. 27 250 Thickness (nm) 200 150 100 50 0 0 0.0025 0.005 0.0075 0.01 0.0125 0.015 0.0175 0.02 0.0225 Conc (g/ml) Figure 3.1 Concentrations used to derive the different film thicknesses 3.2 Results and Discussion 3.2.1 Determination of Tg From Ellipsometry Data Tg was determined using the ellipsometry from the kink in the temperature depedendent thickness plot. The discontinuity in either the film thickness or refractive index versus temperature can also be used as measure of Tg. In this study, we used the film thickness which was obtained from fitting the Ψ and ∆ values using the Cauchy model.53 The fitting to Cauchy model was explained in Chapter 2. A typical graph of the raw data obtained from the ellipsometer is shown in Figure 3.2. 28 Figure 3.2 Typical raw data from ellipsometry From the ellipsometry measurement, the Tg is obtained from the second heating curve. Figure 3.3 shows an example of a typical thickness versus temperature data in the second heating curve for a 60nm PEI film. Tg is obtained as the intersection of two linear fitting on the temperature dependent curve. 73 Thickness (nm) 72 71 70 69 Tg 68 140 150 160 170 180 190 200 210 220 230 240 250 260 270 Temp (°C) Figure 3.3 Plot of thickness vs temperature to determine the Tg 29 Figure 3.4 shows a plot of the Tg obtained as above versus the initial film thickness of the prepared films spin-coated onto Si substrates. It was observed that the Tg begins to show deviation from the bulk value for film less than 35nm, where the Tg begins to increase with decreasing thickness in the polymer film. This phenomenon of a greater Tg in the ultra-thin film points to the fact that PEI has rather strong attractive interactions with the SiOx surface (from Si wafers). This can be explained in terms of the structure of PEI as shown in Figure 1.1, i.e. the polar C=O groups of PEI could be interacting strongly with the SiOx surface. A restricted mobility of PEI molecules at the solid interface is expected due to this interaction. Thus there is a substantially positive deviation of about 10°C for films thinner than 35nm. In general, when the film-substrate interaction is strong, an increase in Tg is observed. When a hydrophobic substrate is used, on the other hand, a weaker interaction and a reduction in Tg are expected. 220 218 216 214 Tg (°C) 212 210 208 206 204 202 200 10 20 30 40 50 60 70 80 90 Thickness (nm) Figure 3.4 Plot of Tg vs film thickness spin-coated onto wafers 30 The determination of Tg becomes more difficult, for films thinner than 15nm. This is a result of several reasons: 1) there is a reduction in signal because less material is being probed in the experiments; 2) there is a reduction in the contrast between the slopes characterizing the glassy and rubbery regions and 3) a broadening of the transitions. 3.2.2 Moisture Sorption Studies Measuring the moisture sorption of PEI films is the other focus of this work. The aim was to probe the thickness dependence on the moisture sorption behaviour and since there was thickness dependence on Tg, to determine if there was a relationship between Tg and the diffusion of water molecules in ultra-thin films. The effect of moisture sorption of PEI ultra-thin films was investigated by measuring the frequency shift of PEI coated on QCM crystals exposed to varying relative humidity environment. Using the Sauerbrey equation31, the QCM frequency change was converted into mass change. The frequency of bare QCM crystal used in each experiment was measured before being coated with the polymer film, while the initial or dry mass of the film is also obtained. The dynamic frequency data was acquired as a function of time. Figure 3.5 shows a typical set of sorption frequency data as a function of time. The initial portion of the frequency-time graph is the stabilization of the QCM crystal in the “dry” environment (relative humidity of 25%) before switching to the “wet” environment (relative humidity of 95%). We observed an immediate drop in the frequency when the film is switched from dry to wet chamber. This phenomenon will be addressed in a later section. After the initial frequency drop, there is more gradual frequency reduction over time. This is attributed to the moisture absorbed by the polymer film over time. The frequency 31 change eventually reached a plateau after ~ 100 minutes. 32 4994980 “Wet” “Dry” Frequency (Hz) 4994970 4994960 4994950 4994940 100 150 200 250 Time (min) Figure 3.5 Frequency-Time graph obtained from QCM The data from QCM was then converted into percentage mass change in order to quantify the amount of moisture that was absorbed by the film. This was done using the Sauerbrey’s equation which converts the frequency change of the QCM crystal into mass change, and later the mass is converted into percentage mass change: PercentageMass = Masst − Masst 0 × 100% Masst 0 Equation 3.1 where Masst is mass at time t and Masst0 is mass at initial. Figure 3.6 shows the plot of the percentage mass change of the different film thicknesses spin-coated on gold-coated and SiOx-coated QCM crystals respectively as a function of time. The exposure time in the high humidity environment was kept at 100min because this is sufficiently long enough to allow the PEI films to reach a moisture content that will remain essentially constant. According to Van Alsten and Coburn10, polyimide film reaches constant moisture content after being exposed to moisture for 1 hour. It was observed that in our experiment, moisture absorption 33 occurred continuously even after 1h but at an extremely slow rate. Generally, similar trends were observed for both the SiOx and Au substrates. There is an increase in the percentage mass change or percentage moisture uptake with a decrease in film thickness with 20nm film having the highest percentage of moisture uptake as shown in Figure 3.6. Our earlier study on Tg of PEI shows that a decrease in thickness of film led to an increase in Tg due to interfacial effects of PEI polymer on the SiOx substrate. Tan et al5 suggested that higher moisture absorption in thinner films was attributed to the moisture rich layer at the film/substrate interface. However, in the current study, both the hydrophilic (SiOx) and hydrophobic (Au) substrates were showing an increase in the moisture absorbed by PEI films as the film thickness reduces. 34 a) Au-coated QCM crystal 7 6 20nm %mass change 5 4 45nm 65nm 3 2 80nm 1 0 0 20 40 60 80 100 Time (min) b) SiOx-coated QCM crystal 5 20nm %mass change 4 45nm 3 2 65nm 1 80nm 0 0 20 40 60 80 100 Time (min) Figure 3.6 Moisture uptake of film of various thicknesses and on different substrates 35 It was, however, observed that the moisture uptake for the films on the SiOx coated QCM crystal reaches a constant value at a faster rate (in ~5min) as compared to that of the gold-coated QCM crystal which took about 20min on average to reach a plateau. At first glance, the initial fast moisture uptake as mentioned earlier by the initial drop in frequency could be attributed to surface adsorption. However, since the top surface of PEI film coated on SiOx substrate is expected to be similar to that coated on Au substrate, the difference in uptake rate observed here is likely caused by moisture absorbed by region nearer to the substrate. Since SiOx is a more hydrophilic substrate compare to Au, it has a higher potential for moisture uptake. Table 3.1 shows the results of the contact angle measurements of a ~60nm PEI film on different substrates using 3µl of deionised water. The results showed that, after spin coating PEI films on the substrates, the resulting contact angle was approximately the same (between 87° ±2) regardless of the substrate used. This result confirms that the initial moisture uptake in the films was not due to the structure or surface of the films but the substrate interfacial layer. Table 3.1. Contact angle measurements Material Contact angle Si wafer 69 ± 4 PEI on Si wafer 87± 2 Au-coated wafer 94 ± 1 PEI on Au-coated wafer 87± 1 SiO-coated QCM crystal 51± 4 PEI on SiO-coated QCM crystal 86± 1 Au-coated QCM crystal 92 ± 1 PEI on Au-coated QCM crystal 86± 2 36 We would thus propose a simple bilayer model where the polymer film is composed of an interfacial layer having a higher propensity for moisture absorption and positioned directly adjacent to the silicon substrate. The second layer extends from the interfacial layer boundary to the free surface having a moisture affinity that is typical of bulk PEI. A schematic diagram is shown in Figure 3.7. Si wafer Free surface Interfacial layer Figure 3.7 Schematic diagram of the simple bilayer model This phenomenon of increased acceleration in moisture uptake and film thickness dependency are in agreement with the findings by Vogt et al54 where they reported that the absorption could be described by interfacial effects on the absorbed water concentration. They suggested a segregation of water near the silicon oxide due to the hydrophilic nature of the interface. It was learnt that a water layer will form on silicon oxide when exposed to a given vapour pressure due to surface ionization.55 The chemical potential of the absorbing system could be written as a summation of the external potentials on the surface and the internal intrinsic potentials as shown in the following equation54 37 µ = Σµ external + Σµ int ernal Equation 3.2 In this experiment, the external contribution is negligible. At the interface, the additional contributions to the chemical potential from the substrate are important. The attractive nature of silicon oxide leads to positive external contribution. This leads to the increase in the chemical potential near the substrate and hence an increase in the water concentration. This is the reason why the initial moisture absorption is much higher for the silicon oxide surface than the gold surface. In the case of the SiOx coated QCM crystal, the rate of moisture uptake is much faster than that of the gold-coated QCM crystal. Despite the initial water uptake difference, PEI films on both the SiOx and the Au substrate shows an increasing moisture uptake as the film thickness reduces. Figure 3.8 shows the plot of percentage moisture absorption of the films as a function of initial film thickness. For film less than about 50nm, the amount of moisture absorbed by PEI and SiOx are similar; but for film more than 50nm thick, the amount of moisture absorbed by PEI film on Au is marginally higher than that of SiOx. This shows that the effect of the more hydrophilic substrate, SiOx, is not substantial in thicker films; and is evident only when the film thickness is below 50nm. 38 Au SiO 5.5 5.0 4.5 % mass change 4.0 3.5 3.0 2.5 2.0 1.5 1.0 0.5 20 30 40 50 60 70 80 Film thickness Figure 3.8 Plot of percentage mass change of the films after exposure to moisture for 100 min 3.3 Conclusion We have observed that the nature of the substrate has an important influence in the determination of the deviation in Tg of the ultra-thin films as compared to the bulk. PEI, being a polar and hydrophilic polymer, has good interfacial adhesion with Si wafer and hence an increase in Tg was observed with decreasing thickness due to the restricted mobility. Substrate interface also had an effect on the moisture sorption of the polymer film. A hydrophilic polymer spin coated on a hydrophilic substrate (i.e. SiOx coated quartz crystal) had a greater initial moisture uptake as compared to that of a hydrophobic substrate (i.e. Au coated quartz crystal). 39 Chapter 4 Thermal Annealing of Parylene-C Ultra-thin Films Parylene C is a widely used protective barrier coating in various applications. It is a hydrophobic, linear and semicrystalline polymer. Recent reports have shown that ultra thin parylene films exhibited reduced crystallinity due to surface confinement56 ; and the moisture uptake behavior of the ultra thin parylene C film was attributed to the lower crystallinity.57 Bulk parylene typically has two crystalline phase, the α and β phase. It has been reported annealing at temperature above 200 oC, the α phase undergoes recrystallization to the β phase and an increase in crystallinity or crystallite size was observed.58-59 In this work, we investigate the effect of film thickness on the annealing of parylene-c films. 4.1 Experimental Section 4.1.1 Materials Silicon (100) wafers, single side polished and 0.6mm thick were purchased from Prime Research Pte Ltd. Parylene dimer (di-chloro-di-p-xylylene) was purchased from Daisan Kasei Co. Ltd. 4.1.2 Substrate Preparation Silicon wafers were cut into 1cm x 1cm pieces. They were first cleaned with a solution of deionised water, H2O2 and NH4OH (6:1:1 volume ratio) at 70°C for 10min. They were rinsed with deionised water and then cleaned with a solution of deionised water, H2O2 and concentrated HCl (6:1:1 volume ratio) at 70°C for another 10min. The waters were then rinsed copiously with deionised water and blown dry with nitrogen. 40 4.1.3 Parylene-C film Preparation The deposition of Parylene-C film was prepared on clean silicon wafers and quartz crystals as the Gorham method60-61 using a commercial parylene coating unit. The reactor consists mainly of three sections: vaporization, pyrolysis and deposition. The dimer was first sublimed at 100°C followed by pyrolysis at 700°C and deposition at below room temperature. The base pressure was kept at ~2x10-2 Torr. The thickness of the film was varied by adding various amounts of feed dimer and reaction time was kept at 1hour 30min. Figure 4.1 shows the polymerization mechanism of parylene C. Cl Cl CH2 CH2 2 H2C CH2 H2C CH2 CH2 CH2 n Cl Dimer Monomer Polymer Figure 4.1 Polymerisation of parylene C 4.2 Results and Discussion 4.2.1 Ellipsometry Results Figure 4.2 shows the ellipsometry results fitted with the Cauchy model of a freshly deposited ~40nm parylene film on silicon wafer. The film was first heated (Heating 1) from 0°C to 180°C, cooled down to 0°C (Cooling 1) and then reheated to 150°C (Heating 2). It was observed that in the initial heating cycle (Heating1), there was an increase in the thickness of film from from 0°C-60°C, then the film thickness decreased 60°C to 120°C and increases again when heated to 180°C. This phenomenon was not observed when the second heating cycle was carried out 41 (Heating 2) where there was an increase in thickness of the film and the 2 best fit lines could be determined to obtain the glass transition temperature (Tg) of the film. This transition is a result of secondary crystallization, as will be discussed in the following. Heating1 Cooling1 Heating2 44.0 43.5 Thickness (nm) 43.0 42.5 42.0 41.5 41.0 40.5 40.0 0 10 20 30 40 50 60 70 80 90 100110120130140150160170180 Temperature (°C) Figure 4.2 Ellipsometry results of a freshly deposited 40nm film as it undergoes the heating and cooling cycles Since the transition observed in the ellipsometry result is a reduction in thickness, it is important to ensure that the transition is not due to thermal degradation of the film. To further explore the temperature transitions from ~60°C to ~120°C, the polymer film of ~40nm was annealed at 3 different temperatures: 40°C which was before the transition, 70°C which was the start of the transition and 115°C which 42 marked the end of the transition. For the films held isothermal at 70°C and 115°C, ellipsometry experiments showed a decrease in the thickness in the first hour and after which the thickness remained constant. Plots of percentage thickness change of the films that were held isothermal at the different temperatures : 40°C, 70°C and 115°C, are shown in Figure 4.3 For film annealed at 40°C, which was before the transition, there was negligible change in the film thickness with respect to time. However for 70°C and 115°C, it was observed that a decrease in thickness was observed in the first 60min and then slowdown in the decrease was observed thereafter. The phenomenon of decreasing thickness was not observed for the 40°C temperature. iso70 iso40 iso115 0.0 -0.2 % thickness change -0.4 -0.6 -0.8 -1.0 -1.2 -1.4 -1.6 -1.8 0 10 20 30 40 50 60 70 80 90 100 110 120 Time (min) Figure 4.3 Comparison of thickness change with respect to time at the different annealing temperatures. 43 The ellipsometry measures thickness change, a reduction in thickness could be due to thermal degradation or evaporation of absorbed moisture. We have ruled out the possibility of thermal degradation because the glass transition temperature of film remains the same after the first heating, where the transition was detected. Although parylene-c is a hydrophobic polymer, residual moisture that is absorbed during storage could be evaporating during the temperature scan. In order to verify that the thickness reduction is not dominated by a loss of mass, either caused by thermal degradation or evaporation of moisture, QCM experiments were conducted 4.2.2 QCM Measurements QCM measurements were carried out to determine if the decrease in thickness was due to mass loss during the annealing since degradation of the film was ruled out. Frequency change in the QCM measurements were converted to mass change using the Sauerbrey equation. This was later converted to percentage mass change for easier comparison. QCM experiments where 40nm parylene-c films coated on the QCM crystals were carried out at 70°C, 115°C and 140°C. This was a similar experiment that was carried out in ellipsometry earlier where the films were held isothermal at the temperatures stated. Figures 4.5-4.7 show the results obtained from the QCM results. Although, there may be an initial sharp mass loss which we have attributed it to the memory effect of thermal exposure, there is insignificant change in the mass thereafter. Thus we can safely conclude that there is no mass loss during the temperature over the period of time. Hence the transition is not due to mass loss. 44 4.2.3 XRD Measurements In order to further characterize the transition observed in the ellipsometry measurement, XRD was carried out. Based on the transition observed in ellipsometry experiments, we carried out the XRD measurement of the parylene films at 2 different temperatures, 120°C and 180°C. XRD patterns for all films show that there is an increase in the peak intensity but not peak shift. This indicates that the thermal annealing in this series of samples resulted in an increase in the crystallinity but not the crystallite size. It is also worth noting that unlike the bulk phase, the annealing did not result in the recrystallization to β phase, but merely an increased in the intensity of the α phase. The XRD results are summarized in Table 4.1. The integrated peak intensities obtained from the XRD peak was normalized to the film thickness. As shown in Table 4.1, there is an increase in the normalized peak intensity for all films annealed at both temperatures. It is also observed that the extent of increase in the normalized peak intensity is lower for thinner films, indicating that the secondary crystallization is restricted by surface confinement. It has also been reported that the crystallinity of parylene-C decreases with decreasing film thickness. 56-57 . 45 2 0 -2 % mass change -4 -6 -8 -10 -12 -14 -16 0 10 20 30 40 50 60 70 80 90 Time (min) Figure 4.4 Mass change obtained from QCM measurements for the 40nm films annealed at 70°C 46 0 %mass change -1 -2 -3 -4 -5 0 20 40 60 80 100 120 140 Time (min) Figure 4.5 Mass change obtained from QCM measurements for the 40nm films annealed at 115°C 2 0 -2 % mass change -4 -6 -8 -10 -12 -14 -16 -18 0 10 20 30 40 50 60 70 80 90 100 Time (min) Figure 4.6 Mass change obtained from QCM measurements for the 40nm films annealed at 140°C 47 0.60 0.56 a) Unannealed film 0.52 Intensity(A.U.) 0.48 0.44 0.40 0.36 0.32 0.28 0.24 0.20 0.16 0.12 0.08 0.04 0.00 10 12 14 16 2 Theta 0.60 0.55 b) Film annealed at 120°C 0.50 Intensity (A.U.) 0.45 0.40 0.35 0.30 0.25 0.20 0.15 0.10 0.05 0.00 10 12 14 16 2 Theta 48 0.6 c) Film annealed at 180°C Intensity (A.U.) 0.5 0.4 0.3 0.2 0.1 0.0 10 12 14 16 2 Theta Figure 4.7 XRD spectrums of a) unannealed films, b) films annealed at 120°C for 10min and c) films annealed at 180°C Table 4.1 Comparison of integrated peak intensities from XRD Thickness (nm) 10 27 38 60 80 150 4.3 Unannealed (Intensity) 0.00485 0.02516 0.05106 0.04623 0.08596 0.16309 120°C (Intensity) 0.02062 0.02403 0.05348 0.0719 0.11718 0.24877 180°C (Intensity) 0.02253 0.029 0.04912 0.10037 0.13691 0.25997 Conclusion Thermal annealing of a series of ultra thin parylene films was carried out. The temperatures for the thermal annealing were selected based on a detectable transition during the heating scan in ellipsometry experiment. QCM was further used to rule out 49 the possibility of mass loss during the heating scan, while XRD data support the hypothesis of secondary crystallization in these thin films. The secondary crystallization upon thermal annealing resulted in an increase in the degree of crystallinity but no observable crystallite size change or recrystallization of the alpha phase. Finally, it was also observed that the extent of secondary crystallization was dependent on the film thickness. Specifically, thinner films show a restricted secondary crystallization due to surface confinement. 50 Chapter 5 Effects of Film Thickness on the Surface Chemistry of Aminoparylene Amino-terminated molecular films are frequently used in the immobilization of enzymes, DNA sensing and adhesion promotion.44 In these examples, the reactive primary amine moieties are often exposed on the surface for interaction with other molecules. The amino group acts as an anchor on the surface and a linkage point for the attachment of other molecules. Thus it becomes important to have a sufficiently high content of primary amines available on the surface. This is the unique properties imparted by the amine group with its capability for reaction with other functionalities.62 Amino-parylene may be use for the above mentioned applications. As a polymer coating with a reactive functional group, it can be useful for biomedical applications. Polymer coatings on surfaces such as poly(ethylene oxide), PEO, have been widely used for producing biocompatible as well as antibiotic materials as the coating exhibit low degrees of protein adsorption, cell adhesion and bacterial adhesion. Gong et al63 developed an approach to grafting PEO-amine chains onto an aldehyde surface. This allows most biomacromolecules such as proteins and peptides, which contain amino groups to be immobilized onto biomaterial surfaces after the aldehyde plasma activation. As mentioned in Chapter 4, parylene is used as a coating in medical device because it is inert and has an excellent barrier to moisture. Aminoparylene could thus be useful in applications that might require the use of the amine functional group for interaction with protein molecules. In this study, we examined the effects of thickness on the concentration of the amino groups in amino-parylene prepared by CVD method. 51 5.1 Experimental Section 5.1.1 Materials Silicon (100) wafers, single side polished and 0.6mm thick, were purchased from Prime Research Pte Ltd. Pyromellitic dianhydride (PMDA) and N,Ndimethylacetamide (DMAc) were purchased from Aldrich and used as received . Amino parylene precursor, 4-aminomethyl-di-p-xylylene was purchased from Daisan Kasei Co. Ltd and used as received. 5.1.2 Substrate Preparation Si wafers were cut into 1cm x 1cm pieces. They were first cleaned with a solution of deionised water, H2O2 and NH4OH (6:1:1 volume ratio) at 70°C for 10min. They were rinsed with deionised water and then cleaned with a solution of deionised water, H2O2 and concentrated HCl (6:1:1 volume ratio) at 70°C for another 10min. The waters were then rinsed copiously with deionised water and blown dry with nitrogen. 5.1.3 CVD Polymerization Amino parylene was deposited by chemical vapour deposition of the precursor using a commercial parylene coating unit designed for the Gorham process60-61. The deposition process is similar to that described for parylene C film in Chapter 4. The temperature in each section was maintained constant throughout the reaction. The dimer was first sublimed at 130°C followed by pyrolysis at 650°C and deposition at below room temperature. The base pressure was kept at ~2x10-2 Torr. The thickness of the film was varied by adding various amounts of feed dimer and reaction time was kept at 1 hour 15 min. 52 5.1.4 Immobilisation of PMDA The polymer films of different thicknesses were immersed in a 0.5%(w/v) DMAc solution of PMDA for 30min and then rinsed with DMAc, a mixture of DMAc/methanol (1:1) and pure methanol in succession and blown dry with nitrogen. The samples were then kept under vacuum. 5.2 Results and Discussion 5.2.1 Deposition of Amino-Parylene Amino-parylene samples were prepared as described in the experimental section. The different thicknesses were prepared by varying the mass of dimer. Figure 5.1 shows the mass of dimer vs film thickness. The thickness of the film was measured using ellipsometry and surface profilometer. The plot shows that there is a linear relationship between the mass of dimer and the thickness of the polymer film, indicating a fairly controllable deposition process. 140 Thickness of film (nm) 120 100 80 60 40 20 0.12 0.14 0.16 0.18 0.20 0.22 Mass of dimer (g) Figure 5.1 Plot of thickness vs. the mass of dimer 53 5.2.2 Immobilization Capacities of Amino-Parylene The primary amine group of the amino-parylene is reactive towards the dianhydride group of PMDA. The reaction resulted in an amide bond. This reaction is schematically shown in Figure 5.2. In this experiment, we expect the reaction to occur primarily on the surface of the film, and the following surface characterization was carried out to quantify the concentration of the available surface amino group as a function of film thickness. H2C CH2 CH2 n H2C + CH2 m O O O O O O NH2 PMDA H2C CH2 O CH2 HN CH2 n CH2 O m O OH O O Figure 5.2 Reaction between amino-parylene and PMDA 5.2.3 FTIR Analysis FTIR was used to confirm that the reaction between PMDA and the amino parylene has occurred. Figure 5.3 shows the FTIR spectra of both amino parylene before and after immobilization. 54 2210 518 611 928 1120 1273 3460 1638 2347 2929 After immobilization 1055 1439 1385 755 2929 3435 1631 2541 3928 3783 Before immobilization 4000 3500 3000 2500 2000 1500 1000 500 -1 Wavenumber (cm ) Figure 5.3 FTIR spectrum of amino parylene before and after immobilization After immobilization, the appearance of a shoulder at around 1630-1850cm-1 is assigned to the amide bond. Figure 5.4 show the enlarged spectra of the parylene film around 1600 cm-1 clearly showing that amide bond was formed after the immobilization reaction. Furthermore, the FTIR spectra for the freshly prepared film and the film after the reaction with PMDA also show the presence of hydrogen bonding. Hydrogen bonding in the freshly prepared film was assigned to the intermolecular hydrogen bonding by the amide group, as discussed in the XPS results. Hydrogen bonding in the film after the immobilization reaction is assigned to both inter and intramolecular hydrogen bonding by the anchored PMDA. 55 %T Fresh film reacted with PMDA 2000 1900 1800 1700 1600 1500 1400 1300 1200 1100 1000 -1 wavenumber (cm ) Figure 5.4 FTIR spectra showing the presence of amide bond in the parylene film after the immobilization reaction, the arrow indicate the amide bond center at 1650 cm-1 fresh %T reacted with PMDA 4000 3800 3600 3400 3200 3000 -1 wavenumber (cm ) Figure 5.5 FTIR spectra showing the hydrogen bonding the parylene film before and after the immobilization 56 5.2.4 XPS Analysis Figures 5.6 and 5.7 show a series of XPS spectra of the amino-parylene before and after the immobilization reaction. The N 1s peak was used to quantify the concentration of the surface amine group. Before immobilization, the initial amino parylene has 2 distinct N 1s peaks, one at 399.8 eV which corresponds to the 1° amine group and the other smaller peak at 401 eV which corresponds to the H-bonded amine. 46, 56 After the reaction with PMDA, a new peak is detected at 402 eV. This peak was assigned to the amide group after PMDA immobilization.44, 64 By taking the ratio of the normalized intensity of the 3 fitted N 1s components after deconvolution, the percentage of amide formation can be estimated. This fitting data is shown in Table 5.1. By taking the ratio of the peak at 402eV (the peak attributed to amide formation) and the peak at ~399.5eV, the percentage of amide formation is determined. The intensity of the amide N peak is used as an indication of the surface concentration of the amine group. Table 5.2 shows the computed percentage amide formation, where the area ratio of peak N1s C (peak assigned to the amide group) was taken with respect to the N1s A peak (peak assigned to 1° amine group from the unreacted film). Figure 5.7 is a plot showing the relationship of percentage amide formation versus the initial film thickness. As the thickness of the films decreases, there is an increasing amount of amine groups present on the surface. 57 37nm before (a) 58nm before Binding energy (eV) 37nm after (b) Binding energy (eV) 58nm before Binding energy (eV) Binding energy (eV) Figure 5.6 (a) and (b). XPS spectra of 37 and 58nm films 58 (c) 95nm before (d) 122nm before Binding energy (eV) Binding energy (eV) 122nm after 95nm after Binding energy (eV) Binding energy (eV) Figure 5.7 (a) and (b). XPS spectra of 95 and 122nm films 59 Table 5.1 Fitted data where A shows the 1° amine peak position, B shows the H- bonded amine position and C shows the position of the amide peak Thickness (nm) 122 N 1s A B Before Immobilization 399.76 401.35 Area Ratio 1 0.33 N 1s A B C After immobilization 399.75 401.07 402.29 Area Ratio 1 0.35 0.19 95 A B 399.83 401.44 1 0.38 A B C 399.87 401.18 402.4 1 0.37 0.18 58 A B 399.63 401.07 1 0.37 A B C 399.83 401.08 402.25 1 0.4 0.22 37 A B 399.47 401.07 1 0.34 A B C 399.88 400.8 402.09 1 0.54 0.37 Table 5.2 Percentage of amide formation with respect to thickness Thickness of Film (nm) 122 95 58 37 % amide formation 12.3 11.6 13.6 19.4 60 22 20 % amide formation 18 16 14 12 10 8 40 60 80 100 120 Thickness of film (nm) Figure 5.7 Plot of percentage of amide formation vs thickness of amino parylene film The higher surface concentration of the amino group as a function of film thickness is further confirm with TOF-SIMS experiments. 5.2.5 TOF-SIMS Analysis TOF-SIMS was carried out to further compliment the XPS data. Comparatively, TOF-SIMS is a much more surface sensitive technique and hence it can reveal more information of the surface functionalities of the prepared aminoparylene films. In contrast to the XPS measurement, unreacted amino parylene films are measured using the TOF-SIMS. The positive ions data was used in this study. Various peaks were analysed to show the essential functional groups and substituents that were present within the polymer molecule. The typical fragments are depicted in Table 5.3. The intensities of the fragments were obtained after the peaks were normalized to the total ion intensities. In Table 5.4, we tabulated the ratios of CH2NH2 61 peak with respect to the various fragmented phenyl groups. The CH2NH2 fragment came from the amino side group, while the fragmented phenyl group came from the backbone of the amino-parylene. Thus, the ratio of CH2NH2 to the backbone phenyl group was used to quantify the surface concentration of the amino groups. A higher concentration of CH2NH2 in relation to the phenyl groups is used as an indication of more amino groups are directed outward on the surface of the amino-parylene films these ratios consistently show that the intensity of the CH2NH2 fragment is higher in thinner films. Figure 5.8 shows the plot of ratio of CH2NH2 with respect to the different phenyl groups and its substituents. The TOF-SIMS results thus confirm that the film surface is more enriched with amino group in the thinner films, as in agreement with the trend seen in XPS analysis. Table 5.3 The substituents fitted in the TOF-SIMS experiment Ion NH3 NH4 C2H5 CH4N C2H6 SiO SiC2H6 C6H4 C6H5 C6H4CH2 (C7H6) C6H5CH2CH (C8H8) C8H9 37 55 95 Int Int Int 2.11 2.24 2.78 3.94 5.19 5.36 1524.51 2126.01 2122.42 60.33 65.91 70.06 43.01 59.2 57.89 23.68 14.65 12.13 322.36 221.86 185.1 83.3 88.96 102.53 380.29 523.35 616.85 42.75 56.33 68.11 119.04 168.03 204.61 347.46 573.15 693.2 122 Int 1.56 3.72 1148.1 63.79 37.21 25.82 374.97 98.64 444.25 60.17 161.51 683.36 Table 5.4 Ratio of CH2NH2 with respect to the phenyl groups and their substituents. CH2NH2/C6H4CH2 CH2NH2/C6H5CH2CH CH2NH2/C8H9 37nm 1.41 0.51 0.17 55nm 1.17 0.39 0.11 95nm 1.03 0.34 0.10 122nm 1.06 0.39 0.09 62 CH2NH2/C6H4CH2 CH2NH2/C6H5CH2CH CH2NH2/C8H9 1.4 1.2 Ratio 1.0 0.8 0.6 0.4 0.2 0.0 40 60 80 100 120 Thickness (nm) Figure 5.8 Plot of ratio of CH2NH2 with respect to different phenyl groups and their substituents 5.2.6 Contact Angle Measurement We expect that, if the surface concentration of amino group is indeed higher in the thinner films, the film would be more hydrophilic. In order to further verify the results from XPS and TOF-SIMS, contact angle measurement was thus carried out. Figure 5.9 shows that, the contact angle increases with decreasing film thickness before immobilization. After the immobilization, the contact angles are consistently lower than those before and there is now no clear dependence on the film thickness. A lower contact angle after the immobilization of PMDA is indicative of the more polar PMDA molecules. We believe that for the same reason, the contact angle 63 doesn’t show a dependence on the film thickness, as compare to the unreacted films; although it must be mentioned that we have not characterized the thickness of the immobilized PMDA. before after 110 108 106 Contact angle (°) 104 102 100 98 96 94 92 90 30 40 50 60 70 80 90 100 110 120 130 Thickness (nm) Figure 5.9 Contact angle measurements 5.2.7 AFM Analysis Contact angle measurement is affected by the surface topology or smoothness. It is known from earlier report57 that the surface roughness of a series of parylene films depends on their film thickness. Therefore, AFM was carried out to ensure that the data obtained from the contact angle experiments was due to the hydrophobicity of the polymer film and not the surface roughness of the film. Figures 5.10 show the AFM images of the films before and after 64 immobilization. There is also no significant change to the average surface roughness of the films before and after the immobilization. This is believed to be a result of the amorphous nature of the amino parylene, in contrast to the semi-crystalline nature of parylene C. XRD measurements shown in Figure 5.11 showed no clear diffraction peaks are detected at low angles. 65 (a) 37nm 37nm film before immobilization RMS = 0.810nm 37nm after immobilization RMS = 0.853nm 66 (b) 58nm 58nm film before immobilization RMS = 0.871nm 58nm after immobilization RMS = 0.717nm 67 (a) 95nm 95nm film before immobilization RMS = 1.075nm 95nm after immobilization RMS = 0.710nm Figure 5.10 (a), (b) and (c) are AFM images of the various films before and after immobilization 68 37nm 58nm 95nm 122nm 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 2theta Figure 5.11 XRD measurements of amino parylene films 69 5.2.8 XSF measurements XSF measurements are carried out to determine the density of the polymer films on silicon substrate. The measurements as shown in Table 5.5 indicated that there were no significant changes in the film density as a function of thickness. This implied that the morphology of the amino films was not due to the packing of the film. This is also confirmed by the XRD measurements of the film which showed that the film was amorphous at all thicknesses. Table 5.5 XSF measurements Thickness (nm) Roughness (nm) Density (g/cm-3) Interface Roughness (nm) 5.3 114.7±0.5 1.4±0.2 1.11±0.06 0.4±0.1 98.1±0.5 1.5±0.2 1.18±0.04 0.1±0.1 73.4±0.7 1.5±0.2 1.11±0.04 0.3±0.2 49.8±0.5 1.0±0.1 1.16±0.06 0.16±0.08 26.7±0.7 1.2±0.2 1.1±0.2 0.5±0.2 Conclusion We have investigated the effects of thickness on the chemical properties of amino-parylene ultra-thin films. It was demonstrated by the different characterization results that, with a decrease in thickness of the polymer films, there is an increase in the amount of primary amine functional groups directed on the surface. 70 Chapter 6 Overall Conclusion This project aims to investigate the various physical, thermal and chemical properties of ultra-thin polymer films as a function of thickness. All these properties have significant impact on applications that require the use of polymer films with thickness [...]... sorption of PEI films is the other focus of this work The aim was to probe the thickness dependence on the moisture sorption behaviour and since there was thickness dependence on Tg, to determine if there was a relationship between Tg and the diffusion of water molecules in ultra- thin films The effect of moisture sorption of PEI ultra- thin films was investigated by measuring the frequency shift of PEI... into the deposition chamber, 4) diffusion of monomer from the region above the substrate through a boundary layer, 5) adsorption 12 of monomer into the substrate, 6) surface migration and bulk diffusion of monomer, and 7) chemical reaction that comprises propagation or initiation There is no termination reaction During steady state growth, the density of radical chain ends on the surface remains constant... indicator of the thermal stability of a polymer and has been a useful material parameter for a number of applications One of the constant concerns in PEI film used for microelectronic application is the moisture absorption since PEI is a relatively polar polymer While the moisture uptake of bulk PEI films is relatively well studied, few reports are available on the moisture uptake of ultra thin films The. .. catalysts and produces linear high molecular weight polymer Parylene is a semicrystalline polymer with degrees of crystallinity and crystalline modifications that are dependent on the deposition conditions.39, 42 The steps of CVD deposition of parylene consist of 1) the sublimation of dimer in a sublimation furnace, 2) cracking the dimer into monomer in the pyrolysis furnace, 3) transportation of the monomer... energy, KE, of the ionized electron is equal to the difference between the energy of the incident radiation, hν, and the binding energy or ionization potential, BE, of the electron This is illustrated in the equation: KE = hν BE For a given atom, a range of BE values is possible, corresponding to the ionization of electrons from the different inner and outer valence shells and these BE values are characteristic... requires the use of solvents and the choice of solvent often must fulfill the following considerations: environmental issues, miscibility, and its effects on the film properties Chemical vapour deposition (CVD) polymerization has gained substantial interest in recent years as it forms polymer films in the absence of solvent and produces conformal pinhole-free coatings 2,2-Para-cyclophanes are examples of. .. applications as wire and cable insulation, and in electronic industry for flexible printed boards due to their excellent thermal and dielectric properties As the film thickness reduces to sub-nanometer scale, the competing effects of free surface and surface confinement play important roles in the overall properties of the film One such property is the glass transition temperature (Tg) Tg is often used... polarized light The incident beam and the direction normal to the surface define a plane that is perpendicular to the surface which is known as the plane of incidence The interaction of the light with the sample causes a polarization change in the light, from linear to elliptical polarization The change in the shape of the polarization is then measured by analyzing the light reflected from the sample In... AFM, the force sensing spring consists of a miniturised cantilever beam clamped at one end and the probing tip at the other end The atomic force microscope (AFM) probes the surface of a sample with a sharp tip, a couple of microns long and often less than 100Å in diameter The tip is located at the free end of a cantilever that is 100 to 200µm long Forces between the tip and the sample surface cause the. .. wafers 30 The determination of Tg becomes more difficult, for films thinner than 15nm This is a result of several reasons: 1) there is a reduction in signal because less material is being probed in the experiments; 2) there is a reduction in the contrast between the slopes characterizing the glassy and rubbery regions and 3) a broadening of the transitions 3.2.2 Moisture Sorption Studies Measuring the moisture ... transparent and hydrophobic polymer We investigated the effect of thermal annealing on the secondary crystallization as a function of film thickness The effect of film thickness on the surface chemical. .. of the films and the chemical properties of the polymer, it becomes important to understand how the polymer properties changes with thickness 1.1 Effect of Thickness on Glass Transition Temperature... indicator of the thermal stability of a polymer and has been a useful material parameter for a number of applications One of the constant concerns in PEI film used for microelectronic application is the