Physical Vapor Deposition
KeywordsMolecular Beam Epitaxy Boron Nitride Pulse Laser Deposition Physical Vapor Deposition Thin Film Deposition
Anodic arc deposition; Arc discharge; Boron nitride nanotubes (BNNTs); Carbon nanotubes (CNTs); Catalyst; Cathodic arc deposition; Chemical beam epitaxial (CBE); Electron beam evaporation; Electron beam physical vapor deposition (EBPVD); Gallium arsenide (GaAs); Gas-phase molecular beam epitaxy (gas-phase MBE); Graphene; Ion beam; Magnetron sputtering; Metal-organic molecular beam epitaxy (MOMBE); Molecular beam epitaxy (MBE); Nanorods; Nanotubes; Nanowires; Plasma; Pulsed laser deposition (PLD); Si nanotubes; Sputtering; Thermal evaporation
Physical vapor deposition (PVD) is referred to deposition processes of thin films and nanostructures through the evaporation of solid precursors into their vapor phase by physical approaches followed by the condensation of the vapor phase on substrates. The whole process consists of three stages: (1) evaporation of the solid source, (2) vapor phase transport from the source to the substrates, and (3) vapor condensation on the substrates. Since PVD technique can convert solid materials into vapor phase without chemical processes, they are very convenient and possible to deposit many types of materials into thin films and nanostructures.
PVD techniques can be classified based on the techniques used to evaporate the solid source materials into their vapor phase [1, 2]. These techniques include (1) thermal evaporation, (2) ion sputtering, and (3) arc discharge. In the following subsections, examples of PVD based on these evaporation techniques will be described.
PVD by Thermal Evaporation
Molecular Beam Epitaxy
The major advantage of MBE comes from its controllable deposition rate at the atomic or molecular scale. This can be achieved due to the use of effusion cells that enable the generation of “molecular beam,” where the generated vapors have low interparticle collisions before they reach to the substrate surface. This means the evaporated atoms escape from the orifice of the cells through effusion and have long mean free path. They neither interact with each other nor with the gases in the vacuum chamber. Due to this capability, MBE was recently used for the growth of graphene, but the outcome is debatable [3, 4]. Finally, gases were also used as the precursors instead of solid sources, and such configuration is called gas-phase MBE or more appropriately chemical beam epitaxial (CBE) . In this case, the pyrolysis of either trimethylindium (TMIn) or triethylindium (TEIn) and trimethylgallium (TMGa) or triethylgallium (TEGa) at the heated substrates was used as the In and Ga source, while As2 and P2 were obtained by thermal decomposition of the trimethylarsine (TMAs) and triethylphosphine (TEP). Such approach is more like a combined MBE and metal-organic chemical vapor deposition (MOCVD) and is sometimes called metal-organic molecular beam epitaxy (MOMBE).
In fact, MBE is an advanced form of thermal evaporation technique with highly controllable deposition thickness (and thus with the drawback of slow deposition rate) due to the use of effusion cells and high vacuum. For the deposition of amorphous thin films, much simple setup can be employed with less demanding vacuum level (~10−2 mbar in a glass bell jar chamber). This approach is simply called direct resistive evaporation and can be achieved by passing a large current through a resistive wire or foil containing the solid material to be deposited. The heating element is often referred to as an “evaporation source.” The evaporation source can appear in many forms including tungsten wires in the form of filaments or baskets and thin tungsten, tantalum, or molybdenum foils in the form of boats.
PVD by Electron Beam (E-Beam) Evaporation
In contrast to MBE, electron beam physical vapor deposition (EBPVD) provides very high deposition rate on large deposition areas. EBPVD is referring to thin film deposition initiated by electron beam evaporation. The deposition rate is typically ranging from tens of nanometer/minute to hundreds of micrometer/minute, although a rate as high as 50 μm/s was reported . EBPVD is based on electron beam evaporation of anode materials (usually in ingot form) in high vacuum (10−4 mbar). A typical EBPVD system can have multiple electron guns and sources. The electron beams can be generated by either thermionic emission or field emission of filaments. These beams are then accelerated to a high kinetic energy and focused on the ingot. The energy of these electrons is converted into thermal energy as the beam bombards the surface of the ingot. The high deposition rate of EBPVD without corrosive products meets the needs of industrial applications. A potential drawback of EBPVD is that filament degradation in the electron gun results in a nonuniform evaporation rate. In addition, EBPVD may cost contamination of the deposition chamber as the vapors generated from the source materials are coated everywhere in the vacuum chamber, the so-called line-of-sight process.
Pulsed Laser Deposition
PLD can be conducted in vacuum or in a low-pressured gas ambient. When PLD is conducted in reactive gas ambient (say oxygen or nitrogen), the technique is called reactive PLD. In this case, the expanding laser plume will form shock waves and cause chemical reaction between the target vapor and the gas molecules. PLD can also be modified into various configurations. For example, plasma can be applied to enhance the interaction of the vapors with the gas ambient, and the technique is called plasma-enhanced or plasma-assisted PLD. For example, such an approach was shown to control the phases of nitride thin films [9, 10]. Sometimes, an auxiliary ion beam is irradiated on the surface of the substrates during thin film deposition as shown in Fig. 2b, and the technique is called ion beam-assisted PLD.
In fact, the mechanism of laser ablation and the quality of samples deposited by PLD depend strongly on the wavelength and energy of the pulsed laser. This can be understood by the absorption of the laser light by the laser plasma when the plasma index of refraction, n(ω), becomes complex. According to the relation n2 (ω) = 1 − [ωp/ω]2, a complex n(ω) will occur when ωp > ω . This means, shorter laser wavelength (larger ω) will reduce the chances of such plasma heating effect. However, one should also consider that the laser plasma frequency, ωp = [N e2 me ε0]1/2, where me is the mass of electron and N is the electron density. In this case, to avoid plasma heating, N must be minimized, which can be obtained by keeping the laser energy or irradiance low. Plasma heating will lead to the so-called photothermal ablation. For example, photothermal ablation of h-BN target will transfer h-BN target to the substrates and prevent the formation of cubic phase (c-BN) . Photothermal ablation of graphite target at longer laser wavelength will also prevent the formation of diamond-like carbon film  as the sp 2 species from the graphite are transferred to the substrates.
PVD by Sputtering
Sputtering is a process where atoms are released from a solid target material due to bombardment of the target by energetic particles such as ions and atoms. The atoms released from the target can then be deposited as thin film on substrates and is commonly referred as PVD by sputtering. The physics behind the process is momentum transfer from the incident ions/atoms to the target materials through collisions. The average number of atoms ejected from the target per incident particle is called the sputter yield. The yield depends on the ion incident angle, the energy of the ion, the masses of the ion and target atoms, the surface binding energy of atoms in the target, and the crystalline orientation of the target when crystals are used as the target. There are many approaches to generate sputter including the use of DC, AC, and RF plasmas or external ion beams.
Sputtering offers several advantages over other PVD techniques including (1) large area deposition as compared to PLD, (2) convenience for thin film deposition of alloy/composite and materials with high melting points, and (3) avoidance of device damaging from X-rays generated by electron beam evaporation. However, sputtering also has the following disadvantages: (1) more complicated setup than PLD, (2) contamination more likely to occur due to the use of plasma as well as relatively lesser vacuum level, and (3) the film morphology that may be rougher and even damaged due to bombardment of energetic growth species and clustering of the growth species in the relatively high deposition pressures.
PVD by DC, AC, RF, and Magnetron Sputtering
PVD can be obtained by applying plasma in various configurations. The simplest one is to apply a DC potential across the substrate (anode) and target (cathode). In this case, plasma can be generated when gases are introduced (10–100s of mTorr) at sufficient electric field strength (~ a few kV/cm). For example, argon gas is often used, as argon ion can lead to high sputter yield and uniform glow discharge. These Ar ions will bombard the target surface to generate vapors that will be deposited to the substrates. Despite the simplicity, the DC configuration will have the following issues: (1) positive charges will build up on the target surface and cause sparking within the plasma and (2) not suitable for insulating target. To avoid these issues, AC and RF configurations are often used. In these cases, the potential between the target and substrate is alternating so that the target will be negative in potential more often (to generate sputtering) than positive potential. These configurations will also allow lower operation pressure (~10–30 mTorr).
The use of magnetron can further enhance the deposition rate at even lower pressures by the use of DC, AC, or RF electric fields. RF potential is of advantage, since it can be used for insulating targets and is applicable in an antenna configuration, that is, without positive bias voltages on the substrates (less electron bombardments and less substrate heating). In this case, the target is placed on the magnetron surrounded by strong electric and magnetic fields. Electrons in the RF plasma will follow helical paths around the magnetic field lines and generate more ionizing collisions near the target surface with the surrounding gas. Thus, more ions will be created and will lead to a higher sputtering and deposition rates. Because of higher ionization rates, plasma can be sustained at a lower pressure.
PVD by Ion Beam Sputtering
Ion beam sputtering (IBS) is a PVD technique in which the target is evaporated by an external ion source (ion beam). In a typical ion source, ions are generated by collisions of neutral gas atoms/molecules with electrons. They are then accelerated by the electric field by a grid electrode toward the target. Before these ions leave the source, they are neutralized by electrons from a second external filament. The major advantage of PVD by IBS is that the energy and flux of ions can be controlled independently. In addition, the IBS approach is applicable to insulating and conducting targets since the flux of “ions” (which are neutralized as they are emitted from the ion beam) that strikes the target is composed of neutral atoms/molecules.
Like the PLD process, reactive sputtering can be achieved by mixing reactive gases (N2, O2, etc.) with the inert gas (Ar, Kr, etc.). While the inert gas is responsible for the sputtering of the target materials, the reactive gases are used to initiate chemical reactions with the target vapors to form oxide or nitride films. Like PLD, reactive sputtering can also be achieved by various configurations. For example, plasma can be applied to enhance the interaction between the target vapors and the reactive gas ambient, and the technique is called plasma-enhanced or plasma-assisted deposition. Sometimes, an auxiliary ion beam is irradiated on the surface of the substrates during thin film deposition, and the technique is called ion beam-assisted deposition (IBAD).
PVD by Arc Discharge
In addition to PVD by thermal evaporation and sputtering, the desired vapors for thin film deposition can be obtained by arc discharge of the target. This is sometimes called arc vapor deposition , where a high-current (104–106 A/cm2), low-voltage (tens of voltages) discharge was applied on the target. For example, cathodic arc deposition refers to arc vapor deposition using the target as the cathode. In this case, a high-voltage “trigger arc” will be ignited between the cathode and a sharp auxiliary anode. This will form “seed” electrons and ions to initiate a low-voltage, high-current discharge between the target cathode and the adjacent anode.
On the other hand, target materials can be evaporated by using the target as the anode of the arc discharge. This PVD process is referred as anodic arc deposition. In fact, this process is similar to electron beam physical vapor deposition (EBPVD) discussed earlier, where arc discharge is generated by irradiating an unfocused electron beam on the target anode. The electrons can be made to spiral in a magnetic field to further ionize the evaporated materials from the target prior to deposition. Like most other PVD processes, reactive arc vapor deposition can be performed by both anodic arc deposition and cathodic arc deposition using reactive gases like nitrogen and oxygen. This will lead to the formation of nitride and oxide films.
The deposition rates of arc vapor deposition are usually higher than that of PVD by sputtering. This is one of the advantages of anodic arc deposition and cathodic arc deposition in addition to the low operation voltage. On the other hand, contamination is the potential drawback of most plasma-based PVD processes. In addition, particulates or the so-called macros are undesired for thin films deposited by anodic arc deposition and cathodic arc deposition. Macros are formed by ablation of molten or solid particles by thermal shock during arc discharges.
Examples of PVD Approaches for Nanotechnology
Physical evaporation techniques like arc discharge  and laser ablation  are important on the discovery of carbon nanotubes (CNTs). However, PVD techniques discussed here are not popular for the growth of nanomaterials in comparison to chemical vapor deposition (CVD). The major reason is the tendency of forming thin films and amorphous by-products on the surfaces of catalysts which cause “poisoning” effect and prohibit the formation of nanotubes or nanowires . However, there have been some pioneering works reported on the growth of nanotubes and nanowires by PVD techniques such as sputtering, PLD, and MBE. For example, vertically aligned carbon nanotubes (CNTs) were first successfully deposited by RF plasma-assisted pulsed laser deposition . In this approach, two RF plasmas were employed. The first one was applied on the substrates to form a negative bias voltage during deposition. The second one was applied on a ring electrode located between the graphite target and the substrates. Hydrogen gas was used for the plasma generation, and Fe powders (20 nm in diameter) were employed as the catalysts. Although the density of these CNTs is low, these nanotubes have high structural order with internal nanotube diameter as small as 0.4 nm. The formation of these smallest possible CNTs was confirmed by transmission electron microscopy (TEM). Later, CNTs were deposited by magnetron sputtering with a dc bias voltage on substrates . Prior to the growth of CNTs, Ni films (10 nm thick) were deposited by e-beam evaporation to form nanoparticles that later function as the catalyst for the vapor–liquid–solid (VLS) process. However, these CNTs are defective with bamboo-like structures. Later, RF plasma-assisted pulsed laser deposition was used to achieve low-temperature growth of boron nitride nanotubes (BNNTs) [14, 17]. In this case, pyrolytic boron nitride pallets were used as the targets, and a RF plasma was applied on the substrates to create bias voltages, which are important to eliminate the formation of boron nitride thin films. Again, these BNNTs have internal nanotubes of 1 nm in diameter with high structural order.
Various PVD approaches were also employed for the synthesis of nanowires/nanorods. For example, ZnO and Zn1−x Mg x O nanorods were grown by MBE on oxidized Si substrate coated with Ag catalyst . In both cases, an ozone/oxygen mixture was used as the oxidizing source. The Zn and Mg cation fluxes were provided by a Knudsen effusion cell using high purity Zn and Mg metals as the sources. The growth of vertically aligned ZnO  and MgO  nanowires can also be obtained by PLD using ZnO or MgO targets, respectively. Gold particles were used as the catalysts for both cases.
Yoke Khin Yap acknowledges the support from the National Science Foundation (Award number DMR-1261910).
- 1.Mattox, D.M.: Handbook of Physical Vapor Deposition (PVD) Processing: Film Formation, Adhesion, Surface Preparation and Contamination Control (Hardcover). Noyes, Westwood, N.J (1998)Google Scholar
- 2.Mahan, J.E.: Physical Vapor Deposition of Thin Films. Wiley-Interscience, New York (2000)Google Scholar
- 7.Chrisey, D.B., Hubler, G.H. (eds.): Pulsed Laser Deposition of Thin Films. Wiley-Interscience, New York (1994)Google Scholar
- 15.Yap, Y.K., Yoshimura, M., Mori, Y., Sasaki, T., Hanada, T.: Formation of aligned-carbon nanotubes by RF-plasma-assisted pulsed-laser deposition. Special issue for Tsukuba symposium on carbon nanotubes in commemoration of the 10th anniversary of its discovery; S. Ijima et al. eds. Phys. B. 323, 341–343 (2002)Google Scholar
- 18.Heo, Y.W., Kaufman, M., Pruessner, K., Norton, D.P., Ren, F., Chisholm, M.F., Fleming, P.H.: Optical properties of Zn1−x Mg x O nanorods using catalysis-driven molecular beam epitaxy. Solid-State Electron. 47, 2269–2273 (2003); Heo, Y.W., Varadarajan, V., Kaufman, M., Kim, K., Norton, D.P., Ren, F., Fleming, P.H.: Site-specific growth of Zno nanorods using catalysis-driven molecular beam epitaxy. Appl. Phys. Lett. 81, 3046–3048 (2002)Google Scholar