Nanofabrication Part I (PHYS 3831, PHYS 3004) Lecture Notes PDF

Nanofabrication Part I (PHYS 3831, PHYS 3004) Dr Catherine Grogan [email protected] SEM Cross section silicon device Intel unveils 22nm 3D Ivy Bridge processor http://www.bbc.co.uk/news/technology-13283...

Nanofabrication Part I (PHYS 3831, PHYS 3004) Dr Catherine Grogan [email protected] SEM Cross section silicon device Intel unveils 22nm 3D Ivy Bridge processor http://www.bbc.co.uk/news/technology-13283882 2011 Traditional planar chip design (left) and Intel's new Tri-Gate technology (right). The company believes that 3D transistors perform more efficiently Intel has unveiled its next generation of microprocessor technology, code named Ivy Bridge.  The chips use a 22 nm manufacturing process, which packs transistors more densely than the 32 nm system. Intel said it would also be using new Tri-Gate "3D" transistors, which are less power hungry. Earlytransistor https://en.wikipedia.org/wiki/Field-effect_transistor Current transistor with scaling and 3D modifications Traditional microprocessor transistors are "planar" The Tri-Gate system features 3D "fins". Intel claims or flat as they pass through the switching gate the greater surface area improves efficiency Now, let’s start our Nanofabrication lectures… Moore’s Law In 1965, Intel co-founder Gordon Moore saw the future. His prediction, now popularly known as Moore's Law, states that the number of transistors on a chip doubles about every two years. This observation about silicon integration, made a reality by Intel, the world's largest silicon supplier, has fuelled the worldwide technology revolution.  Moore’s Law Today, Intel leads the industry with research into new technologies such as high-k/metal gate and tri-gate transistors that will enable Intel to continue the 2-year cycle of Moore's Law for the foreseeable future. Moore’s Law https://en.wikipedia.org/wiki/Moore%27s_law#/media/File:Moore's_Law_Transistor_Count_1970-2020.png Moore’s Law Moore's Law Means More Performance. Processing power, measured in millions of instructions per second (MIPS), has steadily risen because of increased transistor counts. But Moore's Law also means decreasing costs. As silicon-based components and platform ingredients gain in performance, they become exponentially cheaper to produce, and therefore more plentiful, more powerful, and more seamlessly integrated into our daily lives. Moore’s Law As Moore's Law continues, imagine the possibilities: Real-time natural language translation. Imagine being able to speak to someone in a foreign country and having your conversation translated real-time. Facial recognition that works accurately and instantaneously. Imagine being able to capture faces as people enter an airport and match them in real-time against a database of known terrorists, and having a turnstile lock if there is a match. Auto chauffeur. Imagine a car that takes a verbal command for a destination, and can drive you there via the least congested route in the safest possible manner. Evolution of Transistor According to Intel co-founder Gordon Moore, "The implementation of high-k and metal materials marks the biggest change in transistor technology since the introduction of polysilicon gate MOS transistors in the late 1960s." The Intel announcement was quickly followed by an announcement of a similar breakthrough by IBM. Lithography Photolithography Photo-litho-graphy: light-stone-writing Purpose: Transfer same pattern from a template onto a large number of substrates. Common photoresist approach: - Patterns are first transferred to a photoresist (PR) layer after exposure/develop steps - Patterns are then “printed” on the substrate by etching processes using PR as a mask layer - Remove PR (stripping) Pattern Transfer Integrated circuits (IC) and micro-fabricated MEMS devices are formed by defining patterns in the various layers created by wafer-level process steps. Processes -Photo process. The desired pattern is photographically transferred from an optical plate to a photosensitive film coating the wafer/substrate material. -Chemical and physical process. Remove or add materials to create the pattern. - Lift off  Main steps in Lithography Photolithography/lithography is a widely used technique in microelectronics and almost all MEMS processes. - Use a light sensitive material called photoresist. - Spin cast onto wafer/substrate material, partially bake. - Expose parts of wafer/substrate material to light. - Etch away either exposed or unexposed photoresist. -Either etch layer under photoresist, or add material in holes in photoresist. Photoresist: Photoresist is a liquid film that can be spread out onto a substrate, exposed with a desired pattern, and developed into a selectively placed layer for subsequent processing. Typically consists of 3 components -resin A binder that provides mechanical properties (adhesion, chemical resistance, etc.) - sensitizer Photoactive compound - solvent Keep the resist liquid  UV UV mask resist substrate patterned resist substrate Negative resist Positive resist Lithographic Process -Surface preparation (cleaning) -Coating (Spin Casting) -Pre-Bake (Soft Bake) -Alignment -Exposure -Development -Post-Bake (Hard Bake) -Etching using the PR as a masking layer -Stripping/Ashing -Cleaning  Top Down Processing Techniques: Two Primary patterning techniques 1. Subtractive etching/Etch-back: - photoresist is applied on top of the layer to be patterned - unwanted material is etched away 2. Lift-off: - patterned layer is deposited on top of the photoresist - unwanted material is lifted off when resist is removed Subtractive and Additive Creation of Nanostructures The dominant lithographic techniques in Microtechnology are subtractive etching. Subtractive processing techniques can be transferred to nanotechnology. A similar layer deposition technique is used but with reduced thickness. However when progressing onto nanostructures, increased lateral resolution and high precision are required with respect to positioning of masks and the substrate. The processing steps required for nanofabrication compared to microfabrication are similar except a drastic reduction in the amount of material required and in the reaction rate. Because process times are in the second to minute range, the removal of ultrathin layers required a very low rate of removal. Nanostructure Generation by Lift-off Processes Similar to the subtractive process a mask is created. In this case the material layer to be patterned is deposited on top of the mask as opposed to underneath in subtractive. The great advantage of the lift off process is the independence of the process on the etch properties of the patterned layer. The applied etching material only relies on removing the photomask material. It has the potential therefore to fabricate very small structures with high precision. The edge quality of the final patterned material depends on the edge quality of the photomask and the ability to deposit the required material into the photomask openings. Lift-off process Etch-back Novel Lithographic Processes: E-beam lithography Ion-beam lithography X-ray lithography Energy of Light Source Figure 9: Energy of light source Electron Beam Lithography Direct writing the patterns on the wafers. Avoid diffraction limitation (i.e is very short) Low throughput is the fatal drawback One way to improve the run rate is using larger-beam spot size to define central part of elements and minimum beam size (~0.1 mm) to work on element boundary. Comparison of e-beam lithography and optical lithography Optical lithography resolution stops at 20 nm features. Diffraction limits the resolution of features achieved. The main disadvantage to e beam lithography is low throughput and high cost.This ultimately has limited its use to photomask production and low dimensions. The light that exposes the photoresist comes through an aperture defined by the metal pattern of the photomask and the sharpness of the image of the mask produced on the photoresist is limited by the wavelength of the light. This limit determines the minimum linewidth capable of being patterned and what has driven the advances in non-optical based lithography techniques like e-beam, ion beam etc. Implantation Ion implanters are essential to modern integrated-circuit (IC) manufacturing. Doping or otherwise modifying silicon and other semiconductor wafers relies on the technology, which involves generating an ion beam and steering it into the substrate so that the ions come to rest beneath the surface. Ions may be allowed to travel through a beam line at the energy at which they were extracted from a source material, or they can be accelerated or decelerated by dc or radio-frequency (RF) electric fields. Semiconductor processors today use ion implantation for almost all doping in silicon ICs. The most commonly implanted species are arsenic, phosphorus, boron, boron difluoride, indium, antimony, germanium, silicon, nitrogen, hydrogen, and helium. Molecular Beam Epitaxy Molecular beam epitaxy (MBE) is the deposition of atomically perfect layers by evaporation at very low deposition rates. Molecular beam epitaxy grows films with good crystal structure, it is a well controlled process typically done in ultra-high vacuum. However it is an expensive and slow process. Epitaxy is ordered crystal growth, that is, it is the growth of a film where there is a crystallographic relationship between the deposited film and the substrate. Homoepitaxy is where the film and substrate are the same material. Heteroepitaxy is where the film and substrate are different materials. Solid or gas source deposition of film  Si2H6 (g) molecular beam Si2H6 (g) 2Si (s) + 3H2 (g) Si substrate E.g. Si2H6 (g) 2Si + 3H2 Surface interactions during MBE growth: We have discussed earlier interactions at the surface. During MBE we can expect the same surface processes. Namely, adsorption, desorption, migration and incorporation into film. MBE is also another method of doping substrates: When doping a substrate using MBE, dopant is added during the growth of the film. The advantage of this is that the lattice remains intact. However this is an, expensive doping method along with the fact the there is a limit to the concentration, which can be achieved. Basic process chamber and source layout: Ultra high vacuum chamber Source Si2H6 (g) As (s) wafer Ga (s) Molecular beams Evaporation cells Remember ultra-high vacuum is 10-9 Torr. MBE growth occurs on the surface of a heated crystalline substrate. The sources required are placed in evaporation cells. The evaporation cells are positioned so as to provide an angular distribution of atoms or molecules in a beam. The substrate is heated to the necessary temperature and, when needed continuously rotated, to improve the growth homogeneity. Effusions cells used in MBE systems exploit the evaporation process of condensed materials as molecular flux sources in vacuum. The molecular beam condition that the mean fee path λ of the particles should be larger than the geometrical size of the chamber is easily fulfilled if the total pressure does not exceed 10-5 Torr. Careful variation of the temperatures of the cells permits the control of the intensity of the flux of every component or dopant of better than 1%. The UHV environment of the system is also ideal for the many in-situ characterisation tools, like the RHEED (reflection high energy electron diffraction). The oscillation of the RHEED signal exactly corresponds to the time needed to grow a monolayer and the diffraction pattern on the RHEED window gives direct indication over the state of the surface. Advantages of MBE: Precise control achieved of composition and doping of film. Monolayer resolution. Easy to make novel materials Disadvantages of MBE: Slow growth rate and expensive sophisticated equipment required. Nanofabrication Part II (PHYS 3831, PHYS 3004) Dr Catherine Grogan [email protected]  Assessment topics and date Monday 9th December 9 am Same format as last test, MCQ questions and some ‘short answer’ type questions on Brightspace. (designed to help prepare you for final exam) Plasma Physics Surface Physics and SPM Nanofabrication The preparation of Nanostructures. This section covers the various techniques used in the preparation of Nanostructures. It appears, to date the most advanced machines and devices on the nanometer scale are those produced in nature. The basic reason for this assessment is that nature seems uniquely to have developed an information of base locally available instructions for assembly and replication of complex nanometer scale machines and devices. Source: Nanophysicsand Nanotechnology. An introduction to modern concepts in nanoscience. E. Wolf. Wiley Miniaturization to the nanometer scale has been one of the most important trends in science and technology. (Nanotechnology) The chemistry to fabricate nanolayers, the engineering for nanocomposite design and the physics of nanostructure properties has created many exciting opportunities for research. These new interdisciplinary areas in nanoscience and nanotechnology supersede the more traditional disciplines and demand new paradigms for collaboration.  Nanofabrication: Top down and bottom-up fabrication: Two Broad categories to fabricate: Two different fabrication paths are used: The Top down or the Bottom up approach. Top down approach works towards miniaturising using current technologies. Bottom up works towards building molecular devices atom by atom.  Nanofabrication: Top down and bottom-up fabrication: Top down approaches are good for producing structures with long- range order and for making macroscopic connections, while bottom-up approaches are best suited for assembly and establishing short-range order at nanoscale dimensions. The combination of top-down and bottom-uptechniques is expected to result in the best amalgamation of processes for nanofabrication. Nanofabrication requires new methods for fabrication of devices at this scale. Planar Technology Planar technology is realized by implementing layer-by-layer technology. Nanofabricated devices are produced by the deposition of layer by layer of different materials on the substrate. The thickness of these layers varies from 100 m to 1 nm for ultrathin layers. Of course layers of the order of nm’s is the critical dimensions in nanotechnology. Subtractive and additive process’ Individual structural elements can be prepared by the fabrication of a continuous layer and subsequent fabrication processes (subtractive processes) OR by separate deposition of the individual structures (additive process). The subtractive method starts with: - the deposition of the required material layer, - the subsequent deposition of a masking layer, -followed by the etching of any unprotected portions of the required layer leaving the required portions of the layer behind post the etch. The additive or lift-off process The additive process works by: - initially depositing the mask onto the substrate - patterning the mask - followed by the deposition of the required layer. - The subsequent removal of the masking layer leaves behind the required portions of the material layer. Plasma Applications in Nanofabrication Sputtering Etching Deposition Lithography Surface cleaning/preparation Mask Removal  Surface Processes: Surfaces exposed to plasma experience bombardment by energetic ions, electrons, neutrals and photons. A physical process, being the transfer of kinetic energy, and momentum to the substrate atoms/molecules. The projectiles are massive ions or fast neutrals and if their impact results in substrate material being thrown out and lost to the surface then the process is called “sputtering”. Chemical process; Chemical reactions between the substrateand ions or neutrals with high chemical potential. Physical-Chemical processes – these result from a synergistic combination of a physical component and a chemical reaction. The effect of the physical component is to greatly enhance or to actually induce the chemical reaction. Consequences of such surface processes: Material can be removed from a wafer – etching Material can be added onto a wafer surface – deposition Material can be inserted into the surface region volume –implantation The chemical composition of the wafer surface can be changed, by incorporating the wafer atoms into new compounds. These are key processes required for nanotechnology processes.  Deposition: This is the process of coating build up on the substrate surface. Depending on the actual process, some reactions between target materials and the reactive gases may also take place at the substrate surface simultaneously with the deposition process. Physical Vapour Deposition (PVD) and Sputtering: PVD is the physical deposition of thin films of a material by sputtering the required material from a target source. The sputtering process is relatively independent of target materials. PVD processes are carried out under vacuum conditions. The process involved four steps: · Evaporation · Transportation · Reaction · Deposition Evaporation During this stage, a target, consisting of the material to be deposited is bombarded by a high energy source, such as a beam of electrons or ions. This dislodges atoms from the surface of the target, ‘vaporising’ them. Transport This process simply consists of the movement of ‘vaporised’ atoms from the target to the substrate to be coated and will generally be a straight line affair. Reaction In some cases coatings will consist of metal oxides, nitrides, carbides and other such materials. In these cases, the target will consist of the metal. The atoms of metal will then react with the appropriate gas during the transport stage. For the above examples, the reactive gases may be oxygen, nitrogen and methane. In instances where the coating consists of the target material alone, this step would not be part of the process. Advantages of the Physical Vapour Deposition Process ·Materials can be deposited with improved properties compared to the substratematerial ·Almost any type of inorganic material can be used as well as some kinds of organic materials ·The process is more environmentally friendly than processes such as electroplating Disadvantages of the Physical Vapour Deposition Process ·It is a line of sight technique meaning that it is extremely difficult to coat undercuts and similar surface features · High capital cost ·Some processes operate at high vacuums and temperatures requiring skilled operators ·Processes requiring large amounts of heat require appropriate cooling systems · The rate of coating deposition is usually quite slow (could argue a pro or con depending on application) Applications As mentioned previously, PVD coatings are generally used to improve hardness, wear resistance and oxidation resistance. Thus, such coatings use in a wide range of applications such as: · Aerospace · Automotive · Surgical/Medical · Dies and moulds for all manner of material processing · Cutting tools · Fire arms Various Physical Vapour Deposition Techniques: During physical vapour deposition (PVD), a thin film is grown with the atomic species taken part in the growth – atoms, molecules, radicals and/or ions – being deposited from the vapour phase. Among the numerous PVD techniques available, thermal evaporation, magnetron sputtering and pulsed laser deposition (PLD) are some of the most frequently used techniques. The film microstructure – amorphous, polycrystalline or single crystal – and the growth mode – “layer-by-layer” or “nucleation and growth” – are determined by the substrate temperature and the flux and energy of the atomic species hitting the growing film. These growth parameters can be freely chosen using the various PVD techniques. For a film, where a specific microstructure is wanted, only one of the available PVD techniques may be applicable. When growing a film by thermal evaporation, the elements that constitute the film are placed in a crucible and heated by an electron beam or by resistive heating. The deposition flux is controlled by the temperature, while the energy of the atomic species corresponds to thermal energies and cannot be changed. Magnetron Sputtering: In magnetron sputtering, the magnetron is placed in a chamber with an inert gas with a pressure of about 0.3 Pa. A negative bias is applied to the target mounted on the magnetron, whereby a plasma is formed. This plasma is amplified by a magnetic field from permanent magnets in the magnetron. As electrons in the plasma near the target are pushed away from it, an electrical field is formed, which accelerate positive ions towards it. This ion bombardment of the cathode results in sputtering of target atoms, which eventually hit the substrate with energies typically of a few eV The flux of atomic species hitting the cathode is proportional to the electrical power by which the magnetron is driven. Inert-gas ion bombardment of the substrate during growth, which frequently is beneficial, can be obtained by applying a negative bias to the substrate.  Pulsed Laser Deposition: PLD utilises the plasma resulting from the interaction of focused laser radiation pulses with the target surface to deposit thin film coatings composed of one or several target materials. The nature of the solid surface interaction with laser pulses of ns duration determines the energy of the ejected particles in the plasma to the range 10-100 eV with instantaneous deposition rates of up to 1021 cm-2 s- 1. Besides the energy control of the arriving atomic species, with PLD the flux of atomic species hitting the substrate can be accurately controlled. Hereby, stoichiometric line compounds can be grown. With all the three deposition techniques discussed above, nanocrystalline films, including nanocomposites, can be grown. Nanofabrication Part III (PHYS 3831, PHYS 3004) Dr Catherine Grogan [email protected] Continue with Nanofabrication Processes Chemical Vapour Deposition ( CVD) Chemical Vapour deposition the controlled environment of the vacuum chamber to introduce reactants in a vapour form to react and deposit on sample surfaces. Substrates usually heated to enhance reactions. Typically for micrometer thickness ranges, connecting to nm structures. Chemical Vapour Deposition (CVD) Wide variety of Types of CVD reactions: For example: Pyrolysis - thermal decomposition: AB(g) ---> A(s) + B(g) example: Si deposition from Silane at 650 C SiH4(g) ---> Si(s) + 2H2(g) used to deposit: Al, Ti, Pb, Mo, Fe, Ni, B, Zr, C, Si, Ge, SiO2, Al2O3, MnO2, BN, Si3N4, GaN, Si1-xGex,... Atomic Layer Deposition (ALD) Atomic Layer Deposition (ALD) is a technique that deposits ultra-thin films one atomic layer at a time. Reactants are introduced one by one with pump/purge cycles in between. Resulting in a self-saturating surface reaction limited to a single layer on the exposed surface. The result is the deposition of a 100% conformal film, with sequential cycles of these reactions enabling precise control of film thickness. Introduced in 1974 by Dr. Tuomo Suntola and co-workers in Finland to improve the quality of films used in electroluminescent displays. In recent years, it turned out that ALD method also produces outstanding dielectric layers and attracted semiconductor industries for making High-K dielectric materials in silicon transistors used in IC manufacturing. ALD sequence: Releases sequential precursor gas pulses to deposit a film one layer at a time on the substrate. The precursor gas is introduced into the process chamber and produces a monolayer of gas on the wafer surface. A second precursor of gas is then introduced into the chamber reacting with the first precursor to produce a monolayer of film on the wafer surface. Two fundamental mechanisms: -Chemisorption saturation process -Sequential surface chemical reaction Process This unique growth technique can provide atomic layer control and allow conformal films to be deposited on very high aspect ratio structures. ALD methods and applications have developed rapidly over the last few years. ALD film growth is based on chemical reaction of vapors of two precursors on the surface. The substrate is exposed to these precursors one at a time, and therefore ALD is surface controlled growth. This range is determined mainly by the selected precursors. The ALD growth takes place in cycles each producing about 0.1 nm thick layer of the film. The duration of the cycle is typically 1 second. ALD Sequence ALD Applications Including; High-k gate oxides Storage capacitor dielectrics Pinhole-free passivation layers Passivation of crystal silicon solar cells High aspect ratio diffusion barriers for Cu interconnects Adhesion layers Highly conformal coatings for microfluidic and MEMS applications Other nanotechnology and nano-electronic applications Coating of nanoporous structures Fuel cells, e.g. single metal coating for catalyst layers Bio MEMS Etching Etching is the removal of unwanted material. Both Microtechnology and Nanotechnology predominately use a subtractive process to produce structures. Etches remove material from required locations. There are two types of etch, wet and dry. Most plasma etches have a chemical component, also used are physical and chemical combined or purely physical. Wet etching vs dry etching Etch reactants which come from a liquid source are used in a wet etch process. Etch reactants which come from a gas or vapor phase and are typically ionized are used in a dry etch process. Atoms or ions from the gas are the reactive species that etch the exposed film. Selectivity: In general, dry etching has less selectivity than wet etching. Anisotropy: In general, dry etching has higher degree of anisotropy than wet etching. Etch Rate: In general, dry etch has lower etch than wet etching Etch Control: Dry etching is much easier to start and stop than wet etching. The three important parameters measured in an etch process are: (1) Etch Rate, (2) Selectivity), and (3) Uniformity. (1) Etch Rate: The etch rate depends on the chemical reaction at the solid surface, or the removal of reaction products, depending on the rate of individual stages. An important measured parameter of any etch process step is the etch rate of the required material. Mask layer Before Silicon dioxide etch Etch rate = thickness removed/ time of etch. Substrate: Silicon Silicon dioxide after etch Substrate: Silicon 2) Selectivity: The aim of any etch is to have as high as possible selectivity of the etch to the required material to be etched and as low as possible selectivity of the etch to the underlying material. The aim is that while etching your target material, you don’t also etch away any underlying layers. At the beginning of the etch, only the target material will be exposed to the plasma and etch. However as the etch nears the bottom of the target layer, the underlying layer then begins to be exposed for a brief amount of time. Consequently the etc rate of the underlying layer must be very low in plasma etch chemistry. Example: Material to be etched: SiO2. Underlying material: Si In this case the etch is required to have a high selectivity (etch rate) to silicon dioxide and a low selectivity (etch rate) to the Silicon. A masking layer is used to protect the area of material which is required to remain post the etch. Masking principle: A patterned masking material defines the pattern of the required layer that will be transferred to the underlying structure. Mask layer Before Silicon dioxide etch Substrate: Silicon Silicon dioxide after etch Substrate: Silicon Silicon dioxide layer pre and post plasma etch. The selectivity of an etch is quoted as, the ratio of the etch rate of the target material over the etch rate of the underlying material. Selectivity depends on both chemistry and ion bombardment. (3) Uniformity: Uniformity across a wafer is needed so that underlying films are not subjected to extended plasma exposure at varying points on a wafer. The level of uniformity depends partly on selectivity. Advantages of plasma etching over wet etching: It is easier to automate for a commercial process. Plasma processes allow for high purity and less contamination. There is a reduction in the chemical hazards and waste treatment. Plasma etching is anisotropic and so can achieve good geometries and line-width resolution. https://www.google.com/imgres?imgurl=https%3A%2F%2Fmedia.springernature.com%2Foriginal%2Fspringer-static%2Fimage%2Fprt%253A978-0-387-48998-8%252F9%2FMediaObjects%2F978-0-387- 48998-8_9_Part_Fig57_HTML.gif&imgrefurl=https%3A%2F%2Flink.springer.com%2F10.1007%2F978-0-387-48998- 8_751&tbnid=GacF2eaQP5LEfM&vet=12ahUKEwib5tb7x7v0AhXfZBUIHSGuATgQMygBegUIARCvAQ..i&docid=_RqFT2OOrsi0EM&w=365&h=224&itg=1&q=anisotropic%20and%20isotropic%20etch%20profile s&ved=2ahUKEwib5tb7x7v0AhXfZBUIHSGuATgQMygBegUIARCvAQ Physical-Chemical Plasma Etching Process: Plasma is created. This in turn creates: Chemically reactive species Ions Both of these are then transported to the wafer surface Chemical reaction on the substrate material is physically induced by ions striking the substrate Order of the etching process in Physical-Chemical Plasma etching: Generation of etching species Without generating the etching species etching will not proceed. Diffusion to surfaces -Etching species must get to the surface to react with the thin or substrate molecules -The mechanics of getting to the surface can limit aspect ratio, undercutting, uniformity Adsorption -Can also effect aspect ratio Reaction: -Strong function of temperature -Obviously effect the etch rate Desorption -Can stop etch if the reacted species is not volatile Diffusion to bulk gas -Can lead to non-uniform etching due un-reacted etching species.  Sidewall Passivation In order to increase the Degree of Anisotropy, sidewall passivation is used: Formation of Sidewall Passivating Films Formation of nonvolatile fluorocarbons that deposit on the surfaces (Polymerization) The deposit can only be removed by physical collisions with incident ions Ions and inert species Photoresist masking layer Polysilicon Oxide Silicon substrate Fluorocarbon films deposits on all surfaces, but the ion velocity is nearly vertical. As a result, as the etching proceeds there is little ion bombardment of the sidewalls and the fluorocarbon film accumulates Adding hydrogen encourages the formation of the fluorocarbon films because hydrogen scavenge fluorine, creating a carbon-rich plasma (same thing happened when C2F6 is used instead of CF4) Ions and inert species Trade-off between selectivity and Anisotropy Photoresist masking layer Polysilicon Oxide Silicon substrate Scanning Probe Microscopy (PHYS 3831, PHYS 3004, PHYS 3706 Dr Catherine Grogan [email protected] Scanning Probe Miscroscopy SPM Scanning Tunnelling Microscopy Atomic Force Microscopy  Scanning Tunneling Microscopy (STM) Scanning tunneling microscopy (STM) was invented by Binning and Rohrer in 1981 and was awarded the Nobel Prize for Physics in 1986. It can truly be said to have revolutionised surface science and is a key instrument in nanotechnology. The engineering approach used in STM has lent itself to a whole family of different techniques collectively known as scanning probe microscopy. The STM instrument can operate in vacuum or in air, but best resolution is obtained in vacuum due to minimising interference from ambient atmosphere and water adsorption. Scanning Tunneling Microscopy (STM) In the vacuum based technique: an atomically sharp metallic tip (usually tungsten, but gold or platinum-iridium are also used) mounted on a set of three piezoelectric transducers is positioned extremely close to the specimen surface at a distance d (within a few angstroms) and a bias voltage Vb applied between tip and sample in the range from a few tens of mV to about 5V. The tip and sample are not in electrical contact as a vacuum gap Vo exists between them. However a current can still flow if the gap is physically small enough due to electrons quantum mechanically tunnelling across the vacuum gap.  incident vacuum vacuum e wave barrier barrier amplitude transmitted e wave amplitude eVo exp decay eVb d The vacuum gap of width d represents an energy barrier to the electron of eVo which is higher than the bias-driven electron energy of eVb (left). Figuratively, in wave terms, electron wave amplitude rapidly decays exponentially in a barrier, the wave within the barrier being called evanescent. If the barrier width is narrow enough (right), the exponential decay may not fully occur leading to a portion of the evanescent wave surviving and exciting the barrier. Tunneling has occurred.  In wave terms, electron waves will exponentially decay on hitting a potential barrier. The rapidly decaying wave in the barrier is known as an evanescent wave and is characterised by it’s exponential amplitude fall-off during oscillation. The decay is rapid usually leading to an extinction of the electron wave in the barrier – zero transmission. However, if the barrier is narrow enough, the exponential function does not cause the amplitude to fully decay and the remaining wave amplitude can exit the barrier on the other side. This is tunneling and in effect means that a fraction of electron amplitude can penetrate the barrier. In particle terms, the analogy is that there is a low but non- zero probability that an electron can penetrate the barrier and be transmitted. The percentage transmitted is given by the transmission function as dealt with in other courses on Quantum Mechanics. Scanning Tunneling Microscopy (STM) The tunneling current is an exponential function of the sample- tip distance. The tunneling current across the gap is given by I  exp(−2 d ) where  is the decay constant of the electron wavefunction. The decay constant is directly related to the workfunction of the material It is this exponential sensitivity of the current to d that gives the STM its excellent sub-angstrom resolution, coupled with the low voltages involved which ensure that only the states close to the Fermi level are being probed. Atomic resolution can be routinely achieved with this instrument – though as will become apparent, we do not actually see atoms.  Iref Vz Vy Itip Vx Vb Schematic of the x,y,z piezo transducer stage of an STM and the associated voltages. The tip is biased relative to the sample and a feedback mechanism allows the various modes of operation to be controlled. Scanning Tunneling Microscopy (STM) Experimentally, the piezo transducers are extremely sensitive to voltage difference and expand or contract their length depending on the potential across them. They can be used to finely position the tip at any position across the sample surface to within fractions of an angstrom. The STM is understandably extremely sensitive to vibrations and electrical noise, as its measurements are extremely delicate. The STM is therefore suspended on dampening springs and magnetic dampeners. A typical vacuum STM today is precision engineered to minimise outside interference and as a result a system, with UHV system, costs typically between €250k and €500k.  Scanning Tunneling Microscopy (STM) Suspension springs STM head Magnetic/eddy current dampeners An Omicron Nanotechnology (omicron.de) UHV flange mounted STM system  STM Modes of Operation In band energy terms, the first major assumption often used in quickly interpreting STM images is that the band structure of the tip is essentially “flat”. Therefore any effects we see in the image are considered to be due to the electronic structure of the surface only. The bias between sample and tip determines what the STM is “seeing”. Biasing essentially leads to a misalignment in energy terms between the Fermi levels of the tip and the sample. Where the tip is biased more positive than the sample, tunneling will occur from sample to tip meaning that the filled electron states in the sample are supplying the electrons and therefore these states are being probed. Where the bias is reversed, electrons will tunnel from tip to sample but can only do so where an empty electronic state exists on the surface Therefore a current map in this case will map the available empty states on the sample surface. Of note is that insulators cannot be studied since they cannot carry a current. The electrons available for tunneling are only those which lie in the misaligned region, i.e. those electrons in states represented by the band eVb. Thus by careful choice of bias voltage, we can choose to look only at states lying within a certain energy range of the Fermi level – in other words we can be selective about the states we choose to probe. zero or “vacuum” level tip Fermi level eVb tip states sample states There are two major modes of STM scanning operation: In constant height mode, the tip is rastered over the surface line by line at a fixed height and the variation in current flowing between tip and sample is measured (usually current is of the order of nA). A plot of the current levels in the x,y plane gives an electronic density map of the surface. Alternatively, a feedback mechanism is set up so that the tunneling current is kept constant as the tip moves across the surface by changing the z position of the tip. This is known as the constant current mode of operation. As a result, the feedback mechanism moves the tip away from the surface if a large current flows in order to reduce it and moves the tip closer to the surface if a small current flows in order to increase it. This gives a topological map of the surface. Scanning Tunneling Microscopy (STM) Tips which are not sharp lead to multiple tunneling current routes and effectively “blur” any image we are studying, sometimes severely. The lateral resolution of the instrument is thus closely associated with the lateral width of the tip. scan direction scan direction multiple potential current paths in a non- the ideal single tunneling path atomically sharp tip with a sharp tip  Scanning Tunneling Microscopy (STM) The map of the variations in current registered can be analysed by computer to form an electron density map of a surface that, though not strictly an image of atoms, is a good representation of the symmetry of the surface structure. In reality the image is a “mix” of the electronic structures of the tip and sample since tunneling into the tip can only occur where there are empty states in the tip. For a metallic tip, this is a good approximation but as mentioned earlier we assume the tip electronic structure is completely featureless so that the image we see is just surface related. This is a big (and incorrect) assumption and for proper structural analysis, a computer model of the entire system needs to be constructed which properly convolutes the electron states of the sample and of the surface. This makes true STM analysis extremely difficult.  Scanning Tunneling Spectroscopy Scanning tunneling spectroscopy (STS) can also be carried out at any specific point on the surface. In this mode, the tip is positioned at a particular point and the distance is fixed. The bias is then varied from negative to positive (for example from – 3V to +3V) and the. This gives a plot of the current tunneling current recorded voltage (IV) characteristic at that point. This is an important measurement in surface physics because if the IV characteristic is linear then the surface follows Ohm’s law and that point of the surface is metallic. If the IV characteristic looks more like a diode for example, then that point of the surface is semiconducting.  With modern software, STS is usually carried out at every point along the surface simultaneously with the topographic scan so from the topographical image, using a cursor on screen, any point can be chosen and the IV characteristic for that point displayed This gives unprecedented information about the physical and electronic structure of surfaces. However, insulating surfaces cannot be studied as no current flows in or out. STS image of a surface point showing diode like IV characteristics  Atomic Force Microscopy (AFM) A related scanning probe technique, atomic force microscopy (AFM), can also be used to “image” surfaces. In this case, a tiny cantilever with a tip (metallic or diamond) is moved over the surface, usually using piezo transducers as before, and is held just a few angstroms from the atoms. As it moves, tiny repulsions and attractions between tip and sample due, for example, to van der Waals forces, deflect the cantilever up and down. If a laser beam is bounced from the cantilever, then these tiny deflections can be recorded. Again, the variations can be converted into a topographical map which is a representation of surface structure. Atomic Force Microscopy The system is in many ways simpler, since a tunneling current does not have to be set up between tip and sample, a difficult procedure in STM. However, the tip must be brought extremely close to the surface without “crashing”, which is difficult on rough surfaces. Because there is no tunneling current, the interaction between tip and sample is no longer an exponential function of the distance and so the resolution is not as good as STM. But it is still excellent, in the tens to hundreds of angstroms region depending on the sample being probed. AFM has a number of advantages over STM, if resolution can be sacrificed. It runs in vacuum, but does not require vacuum.  Atomic Force Microscopy As there is no tunneling, insulators can be readily studied. Also liquids or solid structures within liquids can be studied and other forces can be made use of, beyond the electrostatic, for example capillary forces in liquids or surface tension effects. Magnetic tips can probe surface magnetic structure in the variation known as magnetic force microscopy - MFM. Chemically active tips structures can even be used to probe specific chemical sites on a surface giving a molecular sensitivity map in a technique known as chemical probe microscopy - CPM. The potential for the future in the use of the wide variety of known and as yet undiscovered probe techniques is huge. A typical non-vacuum AFM can be as cheap as €20k and still provide for excellent surface analysis.  Surface Analysis Techniques There are many techniques in use today for the study of surface and interface properties. Such techniques are capable of studying chemical composition, surface structure and electronic structure and properties. No one technique can give all this information adequately. Technique Acronym Physical Chemical Electronic Structure Composition Structur e Scanning Electron Microscopy SEM ( ) Scanning Tunneling Microscopy STM Atomic Force Microscopy AFM Low Energy Electron Diffraction LEED Reflection High Energy Electron RHEED Diffraction X-ray Photoelectron Spectroscopy XPS Auger Electron Spectroscopy AES Ultraviolet Photoelectron Spectroscopy UPS ( ) Ellipsometry / Reflectivity / RAS ( ) Reflection Anisotropy Spectroscopy  Electron Microscopy The scanning electron microscope (SEM) is well-established as an instrument which images a variety of physical properties at or near the surface of a solid. It works by scanning a focused electron beam across a specimen surface in a TV-like raster. As it moves from position to position, a number of measurements can be made. Either the current flowing from specimen to ground is measured (called the absorption current), or emitted characteristic x-rays are registered, or, most commonly, emitted secondary electron fluxes are recorded. The information is then displayed in a picture format where high intensity signals are represented by bright areas and low intensity signals by dark areas. Modern image processing can convert this to a colour image also showing variations in x-ray emission which can be related to chemical composition.  SEM As a technique relying on a current flowing into the sample, the sample has to be conducting. This is usually accomplished by sputtering a thin layer of gold onto the sample prior to scanning. In many cases, this may destroy the very features that are being studied. Since an electron gun is used, the technique requires a low vacuum to operate. Other variants of the technique include transmission electron microscopy (TEM), where a thinned sample is used and the electron beam passes through it, and reflection electron microscopy (REM) where the beam is reflected off the sample. Optical Methods Materials analysis using optics is quite a broad and varied field. One great advantage is that while the particular samples may or may not require vacuum environments, light sources and optical techniques do not and so are applicable in all environments. Optical techniques also tend to be much safer in that the relatively low energy of photons, in comparison to the energy of electron beams or ion beams, means that they are less likely to cause damage to a material. One of the most important optical techniques in use is spectroscopic ellipsometry. Vacuum Science Lecture 7 and 8 (PHYS 3831, PHYS 3004, PHYS 3706) Dr Catherine Grogan [email protected] Types of Pumps used to reach required pressure in a vacuum chamber In this lecture, understand the types of pumps available and how they operate. Vacuum Pumps Pumps can be considered to generally operate on one of two principles in order to remove gas molecules from a vacuum chamber. Either they mechanically impart a change in momentum to the gas molecules to direct them out of the chamber, or they chemically or non-mechanically trap the gas molecules. There are also pumps that use both approaches. We will discuss the principles of the following, but be aware that there are countless variations and manufacturer specific versions. The standard vacuum symbol for a general pump is shown here to the right.  □ Mechanical □ Non-mechanical □ Momentum Removal □ Momentum Removal □ (gas transfer - compression) □ (entrapment) □ Diaphragm pump □ Sorption pump □ Rotary oil pump □ Getter-ion pump □ Roots pump □ Sublimation pump □ Diffusion pump □ Cryopump □ Turbomolecular pump Gas transfer vacuum pumps use one or several compression stages – the gas particles are removed from the volume which is to be pumped and ejected into the atmosphere (compression pumps). Entrapment vacuum pumps are those where the gas particles which are to be removed condense on or are bonded by other means (e.g. adsorption) to a solid surface, which often is part of the boundary of the volume itself – the particles are not actually removed from the system..  Rotary Vane Pump First stage (backing) pump outlet inlet Indicative operating range 103 – 10-3 mbar (ATM – RV) Used from atmosphere down six orders of magnitude or so, this is a very common first stage pump used to “back” the second stage pump in a system, such as a Roots or turbo. It consists of a fixed steel cylinder, the stator, inside of which a solid cylinder, the rotor, rotates eccentrically. In the two-vane model, the rotor is slotted across one diameter and contains a pair of spring loaded vanes which tightly push against the lubricated stator wall. All surfaces are precision ground and immersed in hydrocarbon oil within the pump casing.  As the rotor rotates, gas from the chamber that reaches the larger trap A is pushed around and compressed to a higher pressure at B and out into the oil reservoir via a pressure activated valve. This reduces the pressure in the trapped section, allowing more chamber gas to expand into A again as the rotor rotates. OUTLET GAS BALLAST Typical small rotary assembly (Edwards) with motor on left, rotor/stator chamber on right. Gas ballast inlet and exhaust outlet are above the rotor/stator chamber Roots Pump or Roots Blower Second stage pump, usually requires backing by first stage Indicative operating range 10-2 – 10-5 mbar (RV – MV) Roots pumps, or Roots blowers, are almost always used in combination with a backing pump such as a rotary. They cannot be used from atmosphere. Two-stage versions are capable of working down to the HV region with highest pumping speeds at 5 to 7 orders below atmosphere. The pump consists of two precision figure-of-eight or similar shaped symmetrical rotors which rotate in opposite directions. The rotors are synchronised and do not touch each other or the chamber wall. Roots Pump or Roots Blower The width of the gap between rotors and between rotor and inner wall is no more than a few tenths of a millimetre which allows the pump to run at high speeds without friction, without lubrication and without mechanical wear. The addition of backing allows the Roots to operate at a lower inlet pressure than direct discharge to atmosphere, which also lowers the power required to drive the pump. Two configurations of Roots blower  Roots Pump or Roots Blower The pump is a rotary positive-displacement type pump and operates by trapping gas in the inlet section and then, through the action of the rotors compressing and expelling the gas to the outlet section on each half rotation, similar to the action of the rotary pump. The pump can quickly shift large quantities of gas in this way and is particularly suited to processes which produce or use large gas loads, such as plasma processes. The Roots has the great advantage of being almost insensitive to dirt. Compression ratio for a pump is the ratio of the pressure at the forevacuum or pump output side to the pressure at the intake side p o k= pi  Turbomolecular (turbo) pump Second stage pump, requires backing by first stage Indicative operating range 10-3 – 10-9 mbar (MV – HV/UHV) Turbomolecular pumps are typically used from MV right down to UHV. They are required to be backed, usually with a rotary. The turbo consists of a stack of rotor discs mounted on a central axis which can spin rapidly, interleaved with stators. The discs have radial blades protruding at a carefully chosen angle which impart momentum to gas molecules, forcing them in a direction along the axis of the stack out of the chamber. The stators are similarly formed, with the blade angle in the opposite direction to the rotors to aid maximum momentum transfer to the gas molecules. On collision with the blades of each subsequent disc, molecules receive further momentum to speed their exit from the chamber. Turbo Pump rotor stack   Schematic of the turbomolecular pump action in altering gas momentum stator stack momentum transfer direction With discs just a few mm in thickness, only a small change in pressure is attained between discs, so that a number of discs are needed to attain useful pressure gradients. If the angle between the slots and the surface of the disc is decreased, a higher compression ratio for the gas is obtained (which is desirable for efficient expulsion of the gas), but at the expense of a slower pumping speed. In practice, the slot angle is varied as you travel down the rotor shaft, with the discs near the high vacuum port having large slot-to- surface angles to ensure high pumping speed while those at the other end, operating at higher pressure, have smaller angles. Photo of a small turbomolecular pump rotor blades outlet to backing pump Difffusion (diff) pump Second stage pump, requires backing by first stage Indicative operating range 10-3 – 10-9 mbar The vapour diffusion pump was developed by two people at around the same time, Gaede in Germany and Langmuir in the USA. The basic pumping action is that a fast stream of vapour molecules, generated by heating a pumping fluid, is forced up a central chimney and emerges from a limited orifice annular nozzle into an initally evacuated (backed) pump casing, the stream eventually condensing on the cooled walls where it is returned to the fluid reservoir for re-evaporation. Like an umbrella of heavy molecules, the streaming direction of the vapour imparts momentum to the gas molecules so creating a pressure gradient between the high vacuum inlet and the discharge outlet. The dense vapour umbrella simultaneously creates a seal across the pump inlet preventing back diffusion.  Diffusion Pump inlet pi pv shockwave front water cooled supersonic casing jet vapour backing pressure pb outlet heater Schematic (left) of diff pup action and a typical pump stack shown right Diffusion pumps are usually operated with a very stable high purity hydrocarbon oil, but mercury may also be used where hydrocarbons cannot be tolerated. The pump must be backed to about 1 mbar or less, usually with a rotary. Diffusion pumps have a high pumping speed over a wide pressure range. Depending on the pumping fluid, speeds will vary but because pumping is based on collision processes, speeds vary inversely with √M for any gas being pumped. Multiple stage pumps are now commonplace, together with pump fluid purifying mechanisms to prolong the operating life of the pump fluids.  Sorption (sorb) pump Second stage pump, used in conjunction with backing by first stage pump Indicative operating range 103 – 10-2 mbar These are physically the simplest of entrapment pumps. They provide a clean and convenient method for backing high vacuum pumps such as the getter ion pump. The sorb consists of a cylindrical container filled with a molecular sieve – a material which is highly porous with a large surface area capable of adsorbing gas by physisorption. Charcoal is an historical example of such a material, though nowadays pumps use zeolites such as alumino-silicates. The sorb container is connected directly to the vacuum chamber via a valve. The sorption capacity of the material is greatly increased by cooling the material to liquid nitrogen temperatures, so the sorb container is wrapped in a blanket container which is filled with liquid nitrogen. Once equilibrium temperature is reached, the valve is opened and the sorb quickly traps large volumes of gas. The valve can then be closed after a few minutes once no further drop in pressure is registered. The pump can later be regenerated by heating to expel the trapped gas. The pumps do not, however, pump inert gases as well as oxygen and nitrogen, and demand a constant laboratory supply of liquid nitrogen. Getter-ion or Sputter-ion (ion) pump Third stage pump, used in conjunction with backing by first and second stages Indicative operating range: low 10-5 – 10-12 mbar The pumping action here is based on the adsorption of ions and the arrangement in the pump is of a massively parallel array of elements. Getter ion or sputter ion pump The sputtering of titanium from the cathodes leads to a fresh film of titanium being deposited on the anodes which can getter gas and it is this process which in fact results in the entrapment of most of the gas molecules pumped. For inert gases which cannot combine with titanium, the ionisation process allows them to be pumped by direct attraction of the positive ions to the cathode. The pumps operate by first ionising gas in a discharge formed by applying a 5kV potential difference between anode and cathode, and then using positive ion collisions to sputter titanium from the titanium cathode. Using a permanent external magnet around the pump, a uniform magnetic field also acts in the pump with its lines of force along the common axis of symmetry of the electrodes. Within this configuration, the electrons are confined to spiral paths between electrodes and consequently can ionise any gas molecules they collide with. Positive gas ions produced in this way are collected by attraction to the cathode and eject titanium atoms by sputtering and yet more electrons leading to further gas ionisation etc.  Getter ion or sputter ion pump Ion pumps can only function at pressure of 10-5 mbar or less, since at higher pressures the load of gas is so large that the electrodes are shorted out and cannot sustain the potential difference required to operate. Ion pumps are connected directly to chambers and are only switched on when the chamber has been roughed out and brought to sufficiently low pressures by other pumps, usually a turbo-rotary combination. Ion pumps do not actually remove the pumped gas to the outside but retain it on the electrodes. But because the actual volume of gas involved is so tiny, there is no need for regular regeneration like for sorption pumps. The current flowing due to movement of gas ions is a direct measure of the number of gas ions and therefore can be used as a measure of pressure; There are no moving parts so the pump is both silent and vibration free. The stray magnetic field from ion pumps may be a disadvantage in certain processes however. Ion pumps are commonly used as final stage pumps in UHV, reaching pressures at the measurable limit of 10-12 mbar. They come in various sizes, from tiny sardine-tin sized pumps used for evacuating tube attachments up to large oven-sized devices with speeds of hundreds of litres per second. Photo of diode style sputter-ion pump arrangement  Sublimation (TSP) pump Fourth stage pump Indicative operating range: low 10-9 – 10-12 mbar The principle of gettering has been used since the early part of the 20th century as a pumping system, particularly in the production of vacuum tubes. Most transition metal surfaces can getter all manner of gas molecules quite effectively but titanium is the main getter used as a pump in vacuum technology. Titanium can getter most gases, with the notable exception of methane. In order to use a transition metal as a getter, it is necessary to produce a clean film or surface of the metal. Any gas molecules in collision with the reactive surface will then be chemisorbed and trapped. One of the simplest ways to do this in UHV is to use a titanium sublimation pump. In this system, titanium elements, or titanium alloy elements, are placed directly into the vacuum chamber and, at sufficiently low pressures (10-8 mbar or less), a high current of 40-50A is passed through the element at a low voltage of 6V or so. The power is sufficient to evaporate the titanium metal from the glowing element which then condenses as a thin film on the wall of the surrounding vacuum chamber. There, it quickly getters surrounding gas molecules, trapping them. The gas molecules are not actually removed from the chamber. In practice, a feedthrough head with 3 filaments is used so that filaments can be cycled through as the titanium is depleted. Additionally, power supplies are timer-controlled since the “firing” of a TSP pump is typically carried out for just 1 to 2 minutes every few hours or so in order to “top up” the pressure of a UHV chamber. The TSP is therefore essentially a third or fourth stage pump usually used in conjunction with an ion pump Photo of three-filament TSP feedthrough arrangement  Cryopump Second/third stage pump Indicative operating range: low 10-3 – 10-7 mbar An increasingly popular pump is the cryopump. There are various types – all consist of a series of manifolds (complex folded surfaces) internally which allow for a large surface area. These surfaces are cooled to liquid helium (4K) or liquid nitrogen (77K) temperatures in the case of the cold trap and as such trap large quantities of gas molecules on supercooled surfaces by physisorption.  Cryopump Cryopumps are clean, quiet, efficient, vibration free. But they demand a continuous supply of liquid helium and as such are expensive. Just like the ion pump, it is an entrapment pump and therefore has limited capacity. It requires backing since it cannot operate from atmospheric pressure. Pressure Measurement Industrial and research processes employ an enormous range of pressure values, some 15 decades or more. This inevitably means that no one gauge will do the job throughout the vacuum spectrum. The general symbol for a vacuum gauge is shown here to the right. There are a number of important considerations involved in choosing a gauge: a gauge in use may actually affect the pressure we are trying to measure; it must be located in as ideal a geometry as possible in order to give meaningful measurements; it must be kept well calibrated; its reading may well depend on the type of gas being measured. In general, gauges are either absolute, in that they directly measure pressure or the number of molecules, or they rely on a physical effect such as thermal conductivity or ionisation and relate pressure to that.  C1 C2 Capacitance Manometer Indicative effective range: low 10+3 – 10-3 mbar An example of an absolute gauge is the capacitance D manometer, a form of diaphragm gauge. Here, a metal diaphragm D sits symmetrically between two centrally-bored capacitor plates C1 and C2. The volume between the diaphragm and one of the plates is usually evacuated (the right hand side here). The diaphragm is exposed to the chamber and the force of molecular collisions, i.e. the absolute pressure, displaces the diaphragm towards one plate and away from the other. Capacitance Manometer As the spacing is altered, the capacitance is changed and this is measured by means of a sensitive capacitance bridge. This approach gives a system of pressure measurement which is extremely stable and has high accuracy, the symmetrical arrangement aiding this considerably. A range of 103 mbar to 10-3 mbar can be achieved in certain circumstances. C1 C2 D  Pirani Gauge Indicative effective range: low 10+3 – 10-3 mbar At regular “high” pressures, above 1mbar, the thermal conductivity of a gas is largely independent of pressure. At lower pressures however, the conductivity is dependent and decreases linearly reaching a value of zero in the high vacuum region. This dependency is utilised in thermal conductivity gauges such as the Pirani gauge. This is a rather insensitive gauge (in comparison to others used at HV and UHV) and works by sensing the heat loss from a hot filament due to the surrounding gas. A thin wire is placed along the axis of a tube, the open end of which connects to the chamber or roughing line. This axial wire is heated to about 200°C by Joule heating. The gauge forms one arm of a Wheatstone bridge arrangement. The heating voltage applied is regulated so that the resistance and therefore the temperature of the filament remains constant. Thus the Wheatstone bridge is always balanced and the voltage applied is directly related to the pressure. compensator pirani The lowest pressure measurable is typically low 10-3 mbar. As the gauge meter depends on the heat conduction of heated the surrounding gas, the pressure filament reading is clearly dependent on the fixed resistor calibration resistor gas composition and is therefore not absolute. gauge power supply vacuum  Ionisation Gauge Anode grid cage Indicative effective range: +150V MV to UHV is the ionisation gauge. Central ion collector Filament -40V assembly Most popular is the Bayard-Alpert gauge. It consists of a fine filament placed outside a cylindrical anode grid or cage which Conflat feedthrough surrounds an axial ion collector. The ion collector is held at around –40V while the anode grid is at +150V. The heated filament produces electrons by the thermionic effect. These electrons travel to the anode and collide with gas molecules on the way, ionising them. The positive gas ions formed are collected by the collector where they give up their charge. Bayard-Alpert ionisation gauge assembly for UHV The ion current registered at the collector is therefore a measure of the gas ions produced and therefore related to the gas pressure. Anode grid cage +150V Central ion collector -40V Filament assembly Conflat feedthrough  Ionisation gauge for UHV Electrons emitted from the filament which do not collide with a gas molecule but just hit the anode can release photons there. These photons can then release further electrons in collisions with the ion collector by the photoelectric effect. But photoelectrons released from the ion collector have the same effect as positive ions reaching the collector and so are measured as a false pressure. This is known as the x-ray effect. This falsely elevated pressure reading is a negligible effect in HV, but in lower pressures, UHV to XHV, the effect can dominate the reading and constitute the ultimate pressure reading of the gauge. Ionisation gauge for UHV The geometry of the Bayard-Alpert gauge minimises the effect by having a fine wire ion collector, which minimises surface area, but the effect still limits the ability of the ion gauge to measure pressure accurately lower than the mid 11s (mbar). Pressure readings at this level should be treated with caution as the gauge is operating at the x-ray limit.  A Typical UHV Vent/Pump Routine A sample UHV chamber used in the treatment of crystalline silicon surfaces (research grade) is shown. The chamber (1) is backed by a rotary (2) whose pressure is monitored by a pirani (3), and this backs a turbo (4) with the entire roughing line connecting to the chamber through a right-angled all-metal valve (5). Third stage pumping consists of a sputter-ion pump (6) which is periodically topped up by a TSP (7) and UHV pressure is monitored by a Bayard-Alpert ionisation gauge (8).  Sample Vent Process Assuming the chamber is at UHV and requires venting up to atmospheric pressure, the following procedure might be followed. The UHV initial state would be chamber held on ion pump, all-metal valve closed. We will vent to pure dry nitrogen at a slightly higher pressure than atmosphere so that on opening the chamber gas flow is outwards rather than contamination air inwards. Connect pure dry nitrogen gas line to Turn of all gauges Turn off TSP and ion pump Sample Vent Process cont. Turn off rotary and power down turbo Turbo should power down gracefully and auto vent through its vent valve With backing line vented to pure nitrogen, slowly “crack” open the all- metal valve Allow slow bleed into vacuum chamber over many minutes while slowly opening all-metal valve Once all-metal valve is fully open chamber should be vented to pure dry nitrogen at a positive pressure  Sample Pump-down Process We assume the initial chamber state is sealed and relatively uncontaminated. All gauges and pumps are initially off. Utilise a system logbook with previous reports/graphs on pump times and pressure behaviour as a guide. Fully open all-metal valve Engage rotary on the roughing line and turn on Pirani When Pirani reaches low 3’s, wait a further 10 minutes or so and then attempt to engage turbo Allow turbo to fully spin up Attempt to engage ionisation gauge – you may have to let the chamber sit on the turbo for an hour or so before this is possible Obtain a pressure in the low 5’s  Sample Pump-down Process Attempt to engage ion pump and observe its current and operating voltage. If over a few minutes the voltage doesn’t continue to rise to high voltage, then the ion pump efficiency is not improving and the system still has too much gas. You should disengage the ion and hold on the turbo for another period of time before trying the ion pump again Assuming the ion pump engages and rises to a high (but not full) operating voltage then the system is ready for bakeout Bake the chamber and the ion pump at about 200°C (either using internal heating elements or external heating tapes wrapped around the system) so that desorbed gases are pumped by the roughing line – this may take 24-48 hours but a plot of pressure over time should show pressure first rise, then plateau then fall back Switch off bake and allow chamber to cool. While still warm, degas any filaments (electron guns, ionisation gauge, evaporation ovens, TSP) to clear adsorbed gases through the roughing line. As the chamber is warm, these gases will not tend to stick to the chamber wall  Sample Pump-down Process Allow chamber to cool and close all-metal valve to hold the chamber on ion pump only as it falls into the 9’s or 10’s. Check for leaks if the chamber is not heading for its ultimate pressure. Engage and schedule TSP for regular pressure “top-up” Plasma Physics cont. (PHYS 3831, PHYS 3004, PHYS 3706) Dr Catherine Grogan [email protected] Plasma Physics cont. Reminder: Plasma Sheaths in Plasma Physics Sheaths and Ambipolar Diffusion. High field regions called sheaths develop at all solid surfaces. Is this always true? If we had some way of generating a plasma without applying a voltage across it, would it be possible to get rid of the sheaths entirely? It turns out that this is not possible. Sheaths will always form around a plasma composed of electrons and ions. Plasma Sheaths: The formation of sheaths arises out of the difference in average velocities between the electrons and the ions. Plasma Sheaths: Consider a slab of plasma between two plates. It is initially uniform and there is no potential difference between the plasma and the plates. This situation cannot last because both electrons and ions are moving and some will escape to the walls. But the rate at which electrons escape is very much greater so in the first few moments more electrons escape than ions. This leaves some extra ions behind between the plates. The extra positive charge causes a potential difference to develop between the plasma and the walls Since the plasma itself shields out any electric field (the potential  is flat), any changes in potential (an electric field) must arise between the plasma and the wall. This region is the sheath and it contains very few electrons. The field in the sheath is such that electrons are reflected back into the plasma. The effect on the ions is they are accelerated by the sheath field to the plates. The extra ions left behind after the initial loss of electrons are all concentrated in the sheath. In the steady state the ions and electrons must naturally be lost at the same rate. The sheath does not reflect all the electrons. Those that have enough energy can overcome the sheath potential and escape to the wall. Enough electrons are reflected to reduce the random thermal flux (which arrives from the plasma at the sheath edge) to the same flux as that of the ions at the walls.  In contrast to the electrons, ions are actually accelerated by the sheath field. Although they are relatively cold in the plasma ( ni , the potential profile becomes complicated and the density profiles become unstable.  Bohm Sheath Criterion We can therefore say that the ion density at the sheath edge must be dropping at least as fast as the electron density and this give a lower limit on the directed ion velocity at the sheath edge. Total Energy = ½ miv2 = ½ kTe Bohm sheath criterion. In real plasmas vs = vB. Note that we use the ion mass but the electron temperature to get the Bohm velocity. Given these two parameters, we can work out how many ions strike a surface in contact with a plasma. It is not always easy to understand what is happening in a complex physical situation such as the plasma to sheath transition. If a more intuitive picture is needed, we can think of the plasma as a high alpine lake which empties into a waterfall at one end. A small fish can act as a representative ion. In the middle of the lake there are no strong currents and the fish can swim freely. However, as the fish swims closer to the end with the waterfall, it experiences a current which gathers strength nearer the end of the lake. The fish can still swim against the current, but there is a tendency to drift along with it. Finally when the fish gets to the end of the lake, a strong current is pulling it to the edge. When it gets there, it is at the point of no return. The water streams over the edge and the fish goes with it, no matter how it tries to resist. The farther it falls, the faster it goes! RF sheath structure and scaling The sheath in rf plasmas has a structure that is strongly influenced by the oscillating voltage applied to the electrodes. RF excitation forces the voltage across the sheath to change at the rf frequency. In this next part we will examine the details of the sheath structure and how the sheath width scales with plasma parameters. How can electrons from a distribution with an average energy of only a few volts ever overcome a sheath voltage which can be as high as 200 or 500 volts? The solution to this dilemma is to remember that the sheath voltage oscillates. As shown before there are brief moments during the rf cycle when the sheath voltage is very low. At these times the electrons escape from the plasma. The electrons must move fast enough through the sheath to be able to take advantage of these brief moments when the voltage is low and ‘the exit door is open’. Because of their very small mass, the electrons do indeed move fast enough, and we can assume that they respond instantaneously to any change of voltage across the sheath. The ions again behave differently. The ion density is always the same. The electrons move depending on the voltage. When the voltage is low, the electron density is the same as the ion density in most of the sheath region, except very close to the wall. Electrons can escape from the plasma at these times. The motion of the electrons and ions is effectively decoupled by the large difference in mass between the two particles. This simplifies the sheath structure considerably. Since the ions respond only to the time average field, their density is stationary and similar to the ion density profile in a DC sheath. The electron density in the rf sheath changes with the instantaneous voltage. The electrons move in an out of the sheath region. When the voltage is low enough electrons escape to balance the ion current so that the average total current in each rf cycle is zero. As shown before, the ion energy depends on the dc voltage across the sheath. The rf sheath is a little more complicated because the amplitude of the rf voltage is not the same as the dc voltage experienced by the ions. Particle and energy balance Effectively using or optimising a plasma process starts with knowing how the plasma will respond if the external parameters are changed. In most applications it is possible to change or control a number of process parameters: Gas pressure and flow. Type of gas, or the proportion of each gas in a mixture. The applied rf power. Duration of process cycle. Process Optimisation: It is essential to know what results can be expected when external parameters are altered. A basic and powerful principle in physics is the use of conservation laws. We can use two simple conservation laws to discover how a plasma will respond to changes in external parameters. The two most important parameters that we can set are the applied rf power and the gas pressure.  The two conservation laws are: 1) Particle balance. This simply means that ions and electrons cannot disappear. If we count how many are created in ionising collisions, then this must be how many are lost at the walls. 2) Energy balance. This is a form of the Law of Conservation of Energy. Energy is supplied by the rf generator to the electrons. It is transferred through various processes to the other particles, but ultimately it must all be conducted to the walls of the reactor. Welcome & Introductions Welcome back everyone to a new semester in college Dr Catherine Grogan [email protected] Techniques for Nanoscience: PHYS 3831 & PHYS 3004 Vacuum and Semiconductor Technology: PHYS 3706 Electronics and Instrumentation: PHYS3804 Lecture 1 (PHYS 3831, PHYS 3004, PHYS 3706, PHYS3804) Dr Catherine Grogan [email protected]  PHYS 3706 TU754/3 PHYS 3831 DT222/4, TU855/DT227/3C PHYS 3004 TU855/DT227/3P PHYS3804 TU877/3, TU878/3, Lecture delivery: follow timetable  Course outline Exam, CA and lab breakdown TU877/4 & TU855/3C as per module descriptor: PHYS3831 (12 weeks. First 9 weeks with TU855/3 and TU754/3, final 3 weeks with TU855/3) TU855/3 P as per module descriptor : PHYS3004 (12 weeks. First 9 weeks with TU877/4 and TU754/3, final 3 weeks with TU877/4) TU754/3 as per module descriptor for PHYS 3706 (12 weeks. First 9 weeks with TU855/3 and TU877/4, final 3 weeks TU754/3 only) TU877/TU878/3 as per module descriptor for PHYS 3804 (First 3 weeks with TU855/3 and TU877/4, final 9 weeks back with Tim Hogan) CA from first 3 weeks is for this content only Module Content Delivery Lecture Notes and presentations on Brightspace --------------------------------------------------------------------- Module Code on Brightspace:PHYS3831 Module Title: Vacuum Tech for Nanoscience PHYS3831: 2023-24 PHYS 3831 (TU855/3 TU877/4) PHYS3706 (TU754/3) --------------------------------------------------------------------- Module Code on Brightspace:PHYS3804 Module Title: Electronics and Instrumentation PHYS3804: 2023-24 PHYS3804 (TU878/TU877/3) --------------------------------------------------------------------- You will all need to self enroll. Delivery of course Importance of using Brightspace for course material All material content will be provided in pdf documents on-line. The Powerpoint presentations will be provided to discuss and explain the material content. There will be two lectures per week Importance of Engagement in the module There is a great importance on you as students engaging fully in the module. Course description Nanoscience requires a broad range of materials manufacturing and analysis techniques. Many modern production industries from semiconductor fabrication to medical devices employ vacuum technology and vacuum-based processing systems for various reasons.  Course description This course will first introduce the fundamentals of vacuum technology and system design and will then proceed to look at a broad range of applications of vacuum technology for materials processing. Applications include introductory surface physics, materials analysis techniques including scanning probe microscopies, plasma technology in industry and chemical and physical growth techniques. Topical examples of structures are used wherever possible. Vacuum Science and Technology  Vacuum Science and Technology It’s application to create macro and nanostructures Module Content: Part 1 (PHYS 3831 PHYS 3004 PHYS 3706, PHYS3804 ) Gas fundamentals & kinetic theory Classification of vacuum Gas sources in a chamber, gas flow regimes, gas flow in real systems and pipes Vacuum materials and sealing techniques The pumping process Vacuum pumps, pressure measurement Mass spectrometry and system analysis, good vacuum Part 2 Plasma Physics (PHYS 3831 PHYS 3004 PHYS 3706, PHYS3804) Introduction to plasmas in the industrial context RF Power, DC sources, microwave sources and plasma generation Plasma-surface interactions / plasma chemistry Plasma processing, damage, sputtering, etching, ashing, deposition, implantation Specific processing, manipulating friction, material hardness, biomaterial treatment Vacuum Science Need idealised conditions to implement controlled processes and reactions. Thin layers of the order of nanometers and sub nanometers currently being produced in research and commercial devices and processes..... E.g to create a thin metal layer as in interconnect in an electronic circuit. Need to control the properties of this metal film, morphology, thickness, resistivity, hardness, roughness....... Therefore need a controlled environment to do so.... Vacuums give us a controlled environment Therefore we are going to start this module by learning about vacuum science and vacuum systems Introduction A Brief History A vacuum is defined as an absence of free matter In practice by this we mean a reduction in the number of free gas or vapour molecules in motion in a specified system or chamber. In the laboratory, vacuum physics is more general again in that it covers any environment which has a different composition and/or a different pressure than atmospheric air. The technology we will introduce therefore serves to isolate a chamber volume in order to control its gas composition and/or its pressure. Need to understand gases. The primary aim in understanding gases then is to comprehend their properties as best as possible so that we may be able to exploit weaknesses in their behaviour. That will allow us develop equipment or techniques to move or divert gas from the chamber we are trying to isolate. We need to think on the atomic level to successfully deal with gas expulsion and atmospheric control. The pressure exerted by the atmosphere The pressure exerted by the atmospheric gas around us is extremely large. A classic experiment was carried out in the mid 1600’s to illustrate the enormity of atmospheric pressure. In 1650 Guericke invented the air pump, which he used to create a partial vacuum. Guericke placed two copper bowls (Magdeburg hemispheres) together to form a hollow sphere about 35.5 cm in diameter. After he used the air pump to evacuate the sphere. Von Guericke’s Magdeburg sphere experiment A team of 16 horses were unable to pull the bowls apart. Air pressure around us was at least equivalent to 16 horsepower! The pressure gradient between inside (low pressure) and outside (atmosphere) created a force imbalance that kept the Magdeburg sphere intact. Measurement of Pressure One of the first barometers or pressure measuring devices was devised by Evangelista Torricelli, a student of Galileo Galilei, in 1643. Torricelli succeeded in creating a vacuum in the top of a tube of mercury by inverting it into a dish of mercury. Torricelli also noticed that the level of the fluid in the tube changed slightly each day and concluded that this was due to the changing pressure in the atmosphere pushing down on the mercury surface in the dish. He wrote: “We live submerged at the bottom of an ocean of elementary air, which is known by incontestable experiments to have weight”. Uses of Vacuum or Controlled Atmospheres Why do we need vacuum? Vacua, or rarefied atmospheres, are produced for any number of reasons: To aid in the removal of contaminants during the processing of food, drugs, materials, integrated circuits. To remove absorbing gaseous media (usually the air and its constituents) To create idealised conditions for particle movement as used in electron guns (found in television tubes, cathode ray oscilloscope tubes etc) or particle accelerators (cyclotrons, synchrotrons, linear accelerators etc.). To control of environmental situations such as the creation of pure atmospheres (necessary for fundamental physics studies like surface physics, plasma physics in the creation of nanostructures and devices etc) To remove unwanted force or momentum transport due to gas bombardment (sometimes necessary or at least desirable in medical physics and surface physics).  We will therefore require methods of vacuum production and methods of vacuum measurement……………… Gas Fundamentals Scales and number To either produce or to measure vacuum, we must first understand gas behaviour. To do that, we must look at gases on an atomic or molecular scale. We first recap on some basics. The masses of individual atoms and molecules are minute but are certainly not negligible. For convenience, the masses are usually expressed in atomic mass units or amu: Atomic mass unit (amu) 1amu = 1.66  10-27kg To state something obvious - this, by the way is an extremely small number. It is on a scale that we cannot easily appreciate. We can determine the relative molecular masses of the main gases of interest: We can determine the relative molecular masses of the main gases of interest: Hydrogen Water Nitrogen Oxygen Carbon Methane Ammonia Carbon monoxide dioxide H2 H2O N2 O2 CO CH4 NH3 CO2 2 18 28 32 28 16 17 44 When we deal with gases (or solids and liquids) in the world around us, we are dealing with millions and millions of molecules. Need a Convenient scale for specifying amount of molecules of gas A convenient measure of gases is the mole, which is defined so that one mole of any substance contains 6.022  1023 constituent atoms or molecules (Avogadro’s number ) A mole of a substance is also known as the gramme molecular equivalent of the substance. If we express the relative mass of any molecule or atom in grammes, then that quantity of the material consists of 6.022  1023 constituent atoms or molecules. Avogadro’s Law When measured at the same temperature and pressure, molar amounts of different gases will have the same number of molecules. Avogadro’s Number NA = 6.022  1023 One mole is the molecular mass in grammes contains 6.022  1023 molecules occupies 22.414 litres or 22.41410-3 m3 at STP. Example How many molecules are there in 34g of ammonia and what is the volume of this mass of ammonia gas at STP? The ammonia molecule has a molecular mass of 17amu. Therefore, 17g ammonia has 6.022  1023 molecules 1g ammonia has (1/17)  6.022  1023 molecules 34g ammonia has (34/17)  6.022  1023 = 1.2044  1024 molecules STP volume is (34/17)  22.414 = 44.828 litres Vacuum Science Lecture 2 (PHYS 3831, PHYS 3004, PHYS 3706, PHYS3804) Dr Catherine Grogan [email protected] Previous Lecture: Introduced Vacuum Chambers and identified their application in the creation of controlled environments. Controlled environments: more control over gas/atom movement and interaction with a surface in a vacuum chamber. Discussed gas molecules in a vacuum chamber, fewer molecules, lower pressure. Using Vacuum Chambers as processing tools: Consider a vacuum chamber with gas molecules moving around. Gas molecules moving around and are colliding with the chambers walls Sample/surface the controlled reactions are happening is placed inside the vacuum chamber. Need to always consider correlation between chamber walls and sample inside vacuum chamber. Energy Distribution – Maxwell Boltzmann Statistics Pressure is defined as the force imparted per unit area. Gas molecules collide with a chamber wall in such quantities that they constitute a force, so we need to quantify this force on the wall area in order to determine the measurable pressure. We need to study the motion of molecules, that is, the kinematics or kinetic theory of the molecules. We deal only initially with ideal gases, consisting of independent particles which undergo elastic collisions. As we add energy to a substance through heating, the substance transforms from solid to liquid to gas. Maxwell Boltzmann Theory Clearly temperature and energy are related here, but since the atoms in a gas are in motion and the atoms in a solid are relatively still, temperature and motion (or speed) are also related. However if we raise a gas to a temperature T, not all the molecules get the same energy. The temperature we are measuring in a laboratory is a macroscopic statistical expression of the behaviour of the gas (what does the temperature of an atom mean?). The gas is a flux of huge quantities of atoms or molecules in constant motion and its particles will all have differing velocities which will change with each collision. Depending on the temperature, and therefore the kinetic energy, the velocity distribution in the container will be different.  Maxwell Boltzmann Theory If we heat a container and allow it to come to equilibrium at a fixed temperature T, a snapshot of the number of particles at each velocity might look like that on the right. The distribution of velocities in the gas is not symmetric but is skewed with larger numbers at low speeds, while a long tail at high speeds is evident. This is known as the Maxwell-Boltzman Distribution (MB) function. The physics of a real gas system is too complex to deal with properly, since the gas is a collective phenomenon, each particle interacting with all the others. However, we can learn a lot about the system by dealing only with the average particle and considering this MB distribution function. For an ideal gas system consisting of N (distinguishable) particles, Maxwell-Boltzmann statistics indicate that the probability f(E) of finding a particle at any energy E in a system at equilibrium (with respect to some constant energy o f the system is 1 f MB ( E) =  E −  exp   kT  Which is a famous result in statistical physics. When considered in terms of the ideal gas where energy E is kinetic energy only, we can deduce from this a distribution formula giving the number of molecules in a gas at any particular velocity v for a constant temperature T:  m  3 2 2 f (v ) = 4   2  e−mv / 2 kT  v  2  kT  Number of molecules at a particular velocity. MB Theory….. People who are used to the principle of equal a priori probabilities, which says that all microstates in a system are equally probable, are often surprised when they come across the MB distribution which clearly says that high energy microstates are markedly less probable than low energy states. But there’s no contradiction here. The principle of equal a priori probabilities applies to the whole system, whereas the MB distribution only applies to a small part of the system. The two results are perfectly consistent. If the small system is in a microstate with a comparatively high energy then the rest of the system ( i.e., the reservoir) has a slightly lower energy than usual (since the overall energy is fixed). On average, when a collection of ideal gas particles is heated to a temperature T, the amount of energy available is partitioned off equally to each degree of freedom (DOF). This is known as the equipartition of energy. A DOF is defined as any independent direction in which movement is possible. For a simple atom that can move in the x, y and z directions in an ideal gas there are 3 possible DOFs that can absorb energy.  1k T Under MB theory, equipartition gives 2 B Joules to each DOF, where k is the Boltzmann constant. (If we assume movement of molecule in 3 directions x,y,z) Then heating a collection of atoms to temperature T gives a total energy to each atom of : 3 12 kB T. As the ideal gas is a collection of independent particles, there is no potential energy and any energy given is purely kinetic, 1 mv 2