The binding of multi functional organic molecules on silicon surfaces 2

The Binding of Multi-functional Organic Molecules on Silicon Surfaces Chapter Experimental 2.1 Principles of surface analysis techniques The commercial availability of reliable UHV technology in the late sixties was a critical step for the development of surface analysis techniques Currently, a truly microscopic understanding of surface phenomena at a molecular / atomic level is possible with the employment of various surface sensitive spectroscopic techniques However, no single spectroscopy is capable of adequately answering all related questions, and modern surface science studies increasingly rely on a multi-technique approach in which a range of surface-sensitive probes are used in tandem to provide complementary information In the present work, a combined experimental and theoretical investigation has been carried out for fundamental understanding of the surface chemistry of a range of multifunctional organic molecules on the Si(111)-7×7 and Si(100)-2×1 surfaces 2.1.1 High-resolution electron energy loss spectroscopy (HREELS) HREELS is a powerful tool to identify vibrational features of surface adsorbates In the vibrational loss regime, the vibration of a specific chemical bond is investigated in terms of the electron energy loss in the inelastic scattering process As an incoming electron approaches the sample surface, it may interact with the adsorbate and be reflected from the surface, leaving the adsorbate vibrationally excited The energy loss of the electron is exactly equal to the vibrational energy of the chemical bond being excited As shown in Figure 2.1, the monoenergetic incoming electron beam can be obtained by passing the thermally emitted electrons through two 127° cylindrical monochromators 31 The Binding of Multi-functional Organic Molecules on Silicon Surfaces arranged in series The reflected electrons are collected by an energy analyzer which has the same structure as the 127° cylindrical monochromator The difference between the incident and reflected electron energies (the energy loss) gives information about the chemical bonds existing on the sample surface The vibrational excitation of HREELS for adsorbed molecules is mainly operated by two independent mechanisms: dipole scattering and impact scattering [1] In the case of dipole scattering, the electric field of the incoming electron interacts with the changing electric field due to molecular vibration, where the incoming electron only senses a composite of the molecular and image dipoles Therefore, only the vibrations with a dipole change perpendicular to the surface are active Dipole scattering is strongly peaked at the specular direction (angle of incidence equals to the angle of reflection, see Figure 2.2) In contrast, the impact scattering is mainly associated with the direct impact between adsorbate and incident electrons This short-range mechanism results in a ‘relaxation’ of the surface selection rule so that all vibrations (both parallel and perpendicular to the surface) may be excited Vibrations excited by impact scattering are best observed off-specular (see also Figure 2.2) In this geometry, dipole forbidden modes may be accessed without their intensities being overwhelmed by the dipole active modes Therefore, compared to other vibrational spectroscopy such as ReflectionAbsorption Infrared Spectroscopy (RAIRS) [1], HREELS has a greater scope for investigating vibrational modes not normally accessible by RAIRS and, deducing information regarding the symmetry of adsorbates in favourable circumstances The particular advantages of electronic electron energy loss spectroscopy with respect to optical techniques are high surface sensitivity, large spectral range, and 32 The Binding of Multi-functional Organic Molecules on Silicon Surfaces additional (non-dipole) excitation mechanisms The major tradeoff is the inherently poorer resolution of electron spectroscopy and the limitation that the experiments have to be performed in vacuum Although the lower resolution might preclude the separation of the vibrational modes in some cases, unique information on the surface interaction can be gathered using EELS in conjunction with symmetry selection rules 2.1.2 X-ray photoelectron spectroscopy (XPS) X-ray photoelectron spectroscopy is one of the most versatile techniques used for surface analysis [3] It provides both the chemical composition and bonding on sample surfaces In XPS studies, X-ray with an energy of 1253.6 eV for Mg Kα1,2 or 1486.6 eV for Al Kα1,2 irradiates a sample surface under UHV conditions Photoionization then takes place in the layers close to the surface The kinetic energy (Ekin) of resultant photoelectrons can be derived from the Einstein relation: Ekin = hν - EB (2a) where EB represents the binding energy of the electron referenced to vacuum level In general, Fermi level (EF) is taken as the reference energy level for conducting solid samples, illustrated by the energy level diagram of Figure 2.3 The sample is electrically connected with the spectrometer so that their Fermi energies are at the same level Accordingly, the kinetic energy of a photoelectron is Ekin = hν - EB - Φ (2b) where Φ is the work function of the spectrometer and EB the binding energy referenced to Fermi level 33 The Binding of Multi-functional Organic Molecules on Silicon Surfaces Specifically, in our XPS experiments, the binding energy (BE) values are given by referencing them to the peak maximum of the Si 2p line (BE=99.3 eV calibrated to Au 4f7/2) [4] of a clean Si substrate with an full width at half maximum of less than 1.2 eV The resolution of a XPS instrument mainly depends on the X-ray source employed The simplest type of X-ray source for XPS is the characteristic emission lines from an anode bombarded by high-energy (14.8 keV) electrons Al Kα1,2 (hν = 1486.6 eV) and Mg Kα1,2 (hν = 1253.6 eV) radiations are most popular due to (1) their energies sufficiently high for ionizing core-level electrons of all elements (i.e keV or more) and (2) relatively “clean” with very few satellites or other peaks, resulting in relatively narrow line-widths (full width at half maximum, FWHM = 0.70 eV for Mg Kα1,2 and 0.85 eV for Al Kα1,2) XPS can provide information about core-level chemical shifts and valence band structure of materials Core-level binding energy of a certain element may be subject to variations (typically up to a few eVs) depending on its chemical environment (i.e oxidization state, lattice sites, molecular environment etc.) The information on the changes in electronic structure, valence states, formation and breakage of chemical bonds is very useful for us to understand the physical and chemical processes occurring on sample surfaces This is extensively reviewed by Egelhoff [5] 2.1.3 Ultraviolet photoelectron spectroscopy (UPS) Instead of using photons of kiloelectronvolt energy, much lower energy photons may be used to excite electrons in the solid In ultraviolet photoelectron spectroscopy (UPS) the source of photons is a differentially pumped inert gas discharge “lamp” This produces discrete low energy resonance lines (e.g He I, hν = 21.2 eV, and He II, hν = 34 The Binding of Multi-functional Organic Molecules on Silicon Surfaces 40.8 eV) with an inherent width of a few millielectronvolts, which can only have sufficient energy to emit electrons from the valence band (electrons with a binding energy of 30 eV or less are generally classified as valence electrons) The obtained spectra reflect the joint density of states of the ground and final states and also depend upon the photon energy The work function (Φ) can also be extracted from the secondary electron cut-offs of the spectra This technique has been widely used in the study of adsorption phenomena and of the valence band structure of metals, alloys and semiconductors Although valence band structures can also be obtained in XPS studies, the photoionization cross section of valence electrons using ultraviolet light (e.g He I and He II) is much higher Angle-resolved UPS (ARUPS) was developed using polarized incident light coupled with variable angle of detection It was demonstrated to be useful in providing detailed information about the orientation of molecular orbitals and structural modifications of adsorbed species on surfaces 2.1.4 Scanning tunneling microscopy (STM) The STM was invented by Gert Binnig and Heinreich Rohrer at IBM, Switzerland, for which they received the Nobel Prize for physics in 1986 It is a powerful tool that reveals the direct and local information such as the reaction sites, the spatial distribution of adsorbates and the local structural and electronic changes upon surface reconstruction Understanding the interaction of species with semiconductor surfaces has been considerably advanced by the now widespread use of STM STM can probe the spatial distribution of electron density on a surface at a level of atomic resolution Biasdependent STM studies, in which images are taken at different voltages between the STM 35 The Binding of Multi-functional Organic Molecules on Silicon Surfaces tip and the sample, allow the determination of surface electronic states, and in special cases, enable the discrimination between different chemical species [6] Figure 2.4 schematically presents a scanning tunnelling microscopy system The principle of STM is not complicated An atomically sharp tip approaches a conducting surface within a separation of a few nanometers and a small potential difference is applied between the tip and sample If the tip is biased positively relative to the sample, an energetic incentive is provided for electrons to flow from the sample to the tip, where their potential energy will be lowered Such a current flowing across “free space” is surprising since, classically, the electrons in the sample are bound within the solid and a minimum amount of energy, equal to the sample work function (typically several eV) must be applied At room temperature, the average thermal energy available is only of the order of tens of meV However, at small tip-surface separations, instead of surmounting the activation energy barrier for electron transfer (the work function), electrons are able to pass through the small vacuum gap by a “special” channel called “tunnelling” This small tunnelling current is readily measurable with modern electronics technology The magnitude of this current is exponentially dependent on the tip-surface separation The larger distance between the tip and surface, the smaller the current Therefore, by measuring the current as a function of separations between tip and sample, a topographic image of the surface can be obtained In order to control the relative position of the tip and surface to within a fraction of an angstrom, piezoelectric actuators are used The physical dimension of this electromechanical device changes as a function of applied electric voltage In practice, it is also necessary to shield the STM against vibrations from outside of the instrument as 36 The Binding of Multi-functional Organic Molecules on Silicon Surfaces well as mechanical resonance induced by motions within the instrument One of the earliest and most striking successes of STM is resolving the structure of the Si(100)-2×1 and Si(111)-7×7 surfaces [7-10] 2.1.5 DFT theoretical calculations More than ever before, there are increasing usages of calculations among surface scientists for investigating the chemical reactions on silicon surfaces Calculations are serving not only as a means of interpreting experimental data, but as a mechanism of supplementing limited data and predicting possible configurations One recent central focus in semiconductor area is the attempt to fundamentally understand the adsorption and binding mechanisms of unsaturated hydrocarbons on silicon surfaces The Si(100) surface has been extensively studied due to its technological importance in semiconductor device manufacturing Several calculation methods, ranging from molecular mechanics to quantum chemistry, have been employed for obtaining information about adsorption kinetics and thermodynamics, and for optimizing adsorbate geometries The relatively simple (2×1) structure of the Si(100) surface can be easily modeled using a small cluster of Si9H12 for chemical binding, only involving one Si=Si dimer For multi-binding modes involving two adjacent dimers in same dimer row or two dimers in neighboring dimer rows, larger clusters, Si15H16 and Si23H24, should be employed, respectively In contrast, it is much more difficult to perform DFT theoretical calculations for the unsaturated organic molecules/Si(111)-7×7 system A cluster including all the reconstruction Si-atoms is intrinsically complex based on the DAS model proposed by 37 The Binding of Multi-functional Organic Molecules on Silicon Surfaces Takayanagi et al.[11] The reconstructed (7×7) unit cell has over 100 atoms in the first three layers Adding an impinging foreign molecule to this system makes it even more difficult to deal with Facing this complicated structure, we first optimize a large cluster that can well represent the spatial distribution of all related atoms Smaller clusters obtained from further reduction were employed to model the possible binding configurations of molecules with density functional theory calculations As shown in the left-bottom panel of Figure 2.5, Cluster II (Si30H28) was cut from the central part of MMFF94 [12] optimized Cluster I containing 973 atoms including the capping H-atoms (the top panel of Figure 2.5), where the precision of atomic positions suffers the least from boundary effects It contains an adatom and an adjacent rest atom from an unfaulted subunit, serving as a “di-σ” binding site for the attachment of one molecule Capping H atoms at the cluster boundaries are kept frozen Silicon atoms in the bottom double layer are placed at bulk lattice positions prior to the geometry optimization process, with each Si-Si bond length set to 2.3517 Å and all bond angles adjusted to 109.4712o Cluster III (Si9H12) was obtained from further reduction of Cluster II Similarly, all capping H-atoms were frozen during geometry optimization Calculation clusters corresponding to possible binding modes are constructed by addition of molecule onto the mother cluster (Cluster III of Figure 2.5) Clusters I, II and III were constructed by us in successful prediction of the adsorption energy of benzene on Si(111)-7×7 [1314] Calculations were performed using SPARTAN package [15].The adsorption energy of binding configurations was calculated at the DFT theory level using perturbative BeckPerdew functional (pBP86) in conjugation with a basis set of DN** (comparable 6-31 38 The Binding of Multi-functional Organic Molecules on Silicon Surfaces G**) [15] Geometric optimizations were conducted under SPARTAN default criteria Heat of formation, synonymous to adsorption energy, is obtained after subtracting the energy of the adsorbate/substrate complex from the total sum of the substrate and gaseous molecule 2.2 Experimental procedures 2.2.1 Ultra-high vacuum systems The experiments were performed on three separate ultra-high vacuum (UHV) chambers All of them have a base pressure of less than × 10-10 Torr (the routine pressure for the chamber equipped scanning tunnelling microscopy is ~ × 10-11 Torr), achieved with turbo-molecular and sputtered-ion pumps (Perkin Elmer) The first chamber is equipped with a high-resolution electron energy loss spectrometer (HREELS, LK-2000-14R, LK Technologies, USA) and a quadruple mass spectrometer (UTI-100, USA) for gas analyses The chamber designed to investigate electronic properties of solid surfaces mainly consists of an X-ray gun (both Mg and Al anodes), He I and II UV sources and hemispherical energy analyzer (CLAM 2, VG Scientific, UK) for X-ray photoelectron spectroscopy (XPS) and ultraviolet photoelectron spectroscopy (UPS) The STM system includes a sample preparation chamber and Omicron VT STM chamber Thus, several techniques including HREELS, XPS, UPS and STM are available for studying the silicon surface chemistry HREELS is commanded to obtain vibrational features of organic molecules adsorbed on Si surfaces; XPS and UPS are employed to investigate electronic properties including core-level shifts and variations of valence band structures; STM is used to provide the information on surface images at a resolution of 39 The Binding of Multi-functional Organic Molecules on Silicon Surfaces atomic level In addition, DFT theoretical calculation was performed to optimize the possible configurations upon chemisorption and calculate their adsorption energies For HREELS measurements, an electron beam with a primary energy (Ep) of 5.0 eV collides on the surface at an incident angle (θi) of 60o from the surface normal The energy resolution of the spectrometer (full width at half maximum, FWHM) was determined to be ~ 6-7 meV The XPS spectra were acquired using the non-monochromatized Mg Kα or Al Kα X-ray source with a pass energy of 20 eV, and a take-off angle of about 75o X-ray source was operated at a power of 296 W (14.8 kV and 20 mA) For XPS, the binding energy scale was referenced to the peak maximum of Si 2p line (99.3 eV) for a clean Si(100) with FWHM of less than 1.2 eV The UPS spectra were collected using He II 40.8 eV for a wider energy window with a photon incidence angle of ∼ 45o and a pass energy of 10 eV and referenced to the Fermi level of the clean metallic tantalum sample holder STM studies were performed on both clean and adsorbate covered Si(111)-7×7 surfaces Both filled and unoccupied state images were collected 2.2.2 Sample preparation The Si samples cut from p-type boron-doped Si(100) or Si(111) wafers (99.999%, 1-30 Ω⋅cm, Goodfellow) were mounted in the following Two pieces of Si single crystals with same dimension (18×10×0.38 mm3) were first covered by evaporation with a thin Ta layer on their unpolished back-sides for homogeneous heating and cooling Then they were ultrasonically cleaned in HPLC methanol and millipore water (18.2 MΩ⋅cm) successively to remove carbonaceous contaminants A piece of Ta foil (0.025 mm thick, Goodfellow) was sandwiched between the two so-treated silicon samples as heater The 40 The Binding of Multi-functional Organic Molecules on Silicon Surfaces samples were clamped together using two Ta clips for fixation The in-between Ta foil was then spot-welded to Ta rods at the bottom of our manipulator A 0.003" W-5%Re/W26%Re thermocouple was attached to the center of one silicon sample by high temperature ceramic adhesive (Aremco 516) for temperature measurement and control Such mounted silicon samples can be resistively heated to 1400 K and conductively cooled to 110 K using liquid nitrogen The temperature distribution on the sample is within ± 10 K at 1000 K as identified by an IR pyrometer (TR-630, ε=0.74, Minolta) After bake-out, the silicon sample was thoroughly degassed at 900 K under a background pressure lower than 3×10-10 Torr Surface contaminants, such as carbon and oxygen, were removed by repeated Ar+ bombardment (1.0 to 1.5 keV, sample current 510 àAãcm-2) and thermal annealing to 1300 K for Si (100) or 1200 K for Si(111) for 1020 minutes Surface cleanliness was confirmed by STM, XPS, UPS and HREELS 2.2.3 Organic Chemicals Pyrazine (99.0 %), pyrazine-d4 (95.0 % D), Pyrimidine (99.0%), s-triazine (99.0 %), acetylethyne (95.0 %), benzadehyde (99.0 %), benzadeyde-α-d1 (99.0 % D), acetophenone (99.0 %) and acetophenone-methyl-d3 (98 % D) used in the experiments were obtained from Aldrich Chemical Co Cyanoacetylene was synthesized by the method of Colos and Waluk [16] In short, the starting compound, methyl propiolate (Aldrich, 99.0%), was transformed into an amide, HCCCONH2, with the excess of liquid ammonia The amide lost water and yielded the nitrile when mixed with P2O5 and heated in vacuum Diacetylene was prepared by dehydrochlorination of 1, 4-dichloro-2-butyne (Aldrich, 98.0%) [17] Cyanoacetylene and diacetylene were all trapped with an acetone / liquid nitrogen slush The final products were characterized using FTIR spectroscopy 41 The Binding of Multi-functional Organic Molecules on Silicon Surfaces (Nicolet Nexus 870) In order to remove air or other volatile contaminants dissolved in chemicals, the freeze-pump-thaw procedure was repeated for several cycles before the vapour of target molecules was delivered onto sample surfaces All organic molecules were introduced into the chamber from a glass bulb through a stainless steel tube with a 0.5 mm inner diameter or directly backfilling the chamber via a precision leak valve The exposure was obtained from the chamber pressure increase without ion gauge sensitivity calibration and expressed in the unit of Langmuir (L) (1 L = 1×10-6 Torr · S) 42 The Binding of Multi-functional Organic Molecules on Silicon Surfaces To view: rotation in plane of paper (Horizontal) 14‫ ״‬Flange θ Variable 13º - 68º θ 30º Sample (a) (b) Figure 2.1 The schematic of HREELS set-up (a) and principle (b) 43 The Binding of Multi-functional Organic Molecules on Silicon Surfaces Figure 2.2 The schematic presentation of specular and off-specular geometry experimental methods 44 The Binding of Multi-functional Organic Molecules on Silicon Surfaces Ekin Ekin Evac hν Φ EF Φ EB Spectrometer Sample Figure 2.3 Energy level diagram for X-ray photoemission 45 The Binding of Multi-functional Organic Molecules on Silicon Surfaces Figure 2.4 The schematic of scanning tunneling microscopy (STM) system 46 The Binding of Multi-functional Organic Molecules on Silicon Surfaces Faulted halves Dimers Adatoms Rest atoms Unfaulted halves Cluster I (Top view) Further reduction Cluster II (Side view) Cluster III (Side view) Figure 2.5 A large cluster of the top five silicon layers constructed based on the DAS model to present three Si(111)-7×7 surface unit cells surrounding a corner hole It (Cluster I) has 973 atoms including the capping H-atoms (not displayed for clarity) Cluster II (Si30H28) and III (Si9H12) are reduced from Cluster I (Ref 12) 47 The Binding of Multi-functional Organic Molecules on Silicon Surfaces References H Ibach, D L Mills, Electron Energy Loss Spectroscopy and Surface Vibrations, Academic Press INC, London, 1982 M K Weldon, C M Friend, Chem Rev 96, (1391) 1996 J F Moulder, W F Stickle, P E Sobol, K D Bomben, Handbook of X-ray Photoelectron Spectroscopy, Physical Electronics Division, Perkin-Elmer Corporation, Minnesota, 1991 D Briggs, M P Seah, Practical Surface Analysis, 2nd ed., Vol 1, Auger and X-ray Photoeletron Spectroscopy, edited by John Wiley & Sons, New York, 1995 W E Egelhoff, Jr Surf Sci Rep 6, 253 (1987) R J Hamers, Y Wang, Chem Rev 96, 1261 (1996) R M Tromp, R J Hamers, J E Demuth, Phys Rev Lett 55, 1303 (1985) R J Hamers, R M Tromp, J E Demuth, Phys Rev B 34, 5343 (1986) R J Hamers, R M.Tromp, J E Demuth, Surf Sci 181, 346 (1987) 10 G Binnig, H Rohrer, Ch Gerber, E Weibel, Phys Rev Lett 50, 120 (1983) 11 K Takayanagi, Y Tanishiro, M Takahashi, S Takahashi, J Vac Sci Technol A 3, 1502 (1985) 12 T A Halgren, J Comput Chem 17, 490 (1996) 13 Y Cao, X M Wei, W S Chin, Y H Lai, J F Deng, S L Bernasek, G Q Xu, J Phys Chem B 103, 5698 (1999) 14 Z H Wang, Y Cao, G Q Xu, Chem Phys Lett 338, 7(2001) 48 The Binding of Multi-functional Organic Molecules on Silicon Surfaces 15 W J Hehre, J Yu, P E Klunzinger, A Guide to Molecular Mechanics and Molecular Orbital Calculations in SPARTAN, Wavefunction, Inc., Irvine, CA, 1997 16 R Colos, J Waluk, J Mol Struct 409, 473 (1997) 17 M Khlifi, P Paillous, C Delpech, M Nishio, P Bruston, F Raulin, J Mol Spectrosc 174, 116 (1995) 49 ... and 32 The Binding of Multi- functional Organic Molecules on Silicon Surfaces additional (non-dipole) excitation mechanisms The major tradeoff is the inherently poorer resolution of electron spectroscopy... Molecules on Silicon Surfaces Figure 2. 2 The schematic presentation of specular and off-specular geometry experimental methods 44 The Binding of Multi- functional Organic Molecules on Silicon Surfaces. .. between the STM 35 The Binding of Multi- functional Organic Molecules on Silicon Surfaces tip and the sample, allow the determination of surface electronic states, and in special cases, enable the