13DB Chrisey and GK Hubler Pulsed Laser Deposition of Thin Films Wiley 1994 General Description

In PVD of multicomponent thin films, the preparation of a stoichiometric vapor is often difficult to achieve. When singleelement sources are used, the arrival rate of each individual component must be calibrated and inter-regulated. In PLD, the composition of the vapor is the same as that of the target. Starting with the correct composition of the vapor greatly facilitates the growth of a desired phase in the depositing film. In addition, the absence of filaments or charged particles allows films to be grown by PLD in the presence of reactive gases. These gases can promote the growth of the desired thin film material through gas-phase reactions with the evaporated target material and gas-surface reactions with the growing film.

The pulsed laser ablation of materials can be viewed as taking place in several stages (Ref 6). The laser deposits energy into the target. Absorption of the radiation by the target material leads to surface heating and defect formation. The energy absorbed per unit volume depends on the optical penetration depth, the thermal diffusivity, and the rate at which the energy is deposited (Ref 14, 15, 16). The rate at which the energy is deposited is determined by the laser pulse width. The rise in surface temperature can be calculated from the ratio of the rate at which energy is deposited into the material to the rate at which the heat is conducted away. Typically, the laser is focused to a small spot (e.g., a few square millimeters) to minimize the volume of material being heated and to achieve surface temperatures that are greater than the melting temperature. If the optical penetration depth is small compared to the thermal diffusion length, surface heating is confined to the thermal diffusion length, and the change in temperature can be calculated from the energy absorbed, the volume of material irradiated, and the heat capacity. If the optical penetration depth is long, the rise in temperature is at a maximum at the target surface but decays exponentially as a function of depth below the surface (Ref 14, 15, 16).

Surface heating is followed by melting and evaporation. For most materials, there is a threshold laser fluence for the macroscopic removal of material from the surface (Ref 17). The high temperatures generated at the surface cause the emission of many species from the target, including ions and electrons generated by thermionic emission as well as atoms and molecules. Continued interaction of the laser pulse with the evaporated material causes the vapor to be ionized via nonresonant multiphoton processes, creating a plasma above the target. As the density of electrons increases, the laser radiation is absorbed preferentially in the plasma by inverse bremsstrahlung scattering (Ref 6). The absorption further heats the plasma and at the same time screens the laser pulse from further interaction with the target.

The high-pressure, high-temperature plasma expands rapidly into the vacuum. The dynamic behavior of the expansion has been compared to a supersonic or free-jet expansion (Ref 18, 19). In a free-jet expansion, the random motion of gas-phase particles in a high-pressure plasma is converted to a directed mass flow by expanding through a pinhole into a low-pressure region. Free-jet expansions are characterized by narrow, energetic velocity distribution about a center of mass velocity (Ref 20). A similar phenomenon is observed in PLD. A small high-pressure region is created at the target surface and then expands into the relatively low pressure of the vacuum system. The laser-generated plasma is a complex mixture of neutral and charged particles, each with different kinetic energies. Average neutral energies of » 10 eV are observed (compared to thermal evaporation particle energies of »0.1 eV). Charged particle kinetic energies of as much as 10 to 100 times the neutral particle energies are also observed. The fraction of ionized material in the plasma is difficult to measure. Estimates vary between 1 and 10% (Ref 21).

The high energy content of the depositing vapor is believed to be one of the reasons for the success of PLD over other PVD techniques. The added energy can benefit film quality in several ways. First, gas-phase reactivity increases with atom kinetic energy. Enhanced oxidation cross sections are reported for laser-evaporated materials (Ref 22). For some materials, a correlation has been reported between the maximum concentration of metal oxides formed in the gas phase and optimum film properties (Ref 21, 29). Secondly, PLD films can typically be grown at a reduced substrate temperature in comparison to other PVD techniques. Substrates are usually heated to provide the arriving vapor with enough surface mobility for the atoms to build up on the surface in thermodynamically stable sites. Adding the energy to the vapor, as is done in PLD, means the substrate temperature can be lowered. This is desirable in cases in which the vapor can undergo a chemical reaction and the substrate or volatile components can be lost from either the film or the substrate.

Under optimized conditions, oriented thin films can be prepared by PLD in situ, minimizing the production of impurity phases, grain boundaries, and random orientations that might be formed during a postdeposition anneal. Although average film deposition rates for PLD are comparable to those in many other PVD techniques, the pulsed arrival of the vapor at the substrate results in instantaneous deposition rates of »1000 to 10,000 A/s. The instantaneous deposition rate minimizes the incorporation of background impurities into the depositing films, further improving the film quality.

Plume Diagnostics. A number of diagnostics have been used to characterize the laser-generated plasma (Ref 23): optical emission (Ref 24, 25, 26), absorption (Ref 27), laser-induced fluorescence (Ref 28, 29), resonance-enhanced multiphoton ionization (Ref 30), mass spectrometry (Ref 31, 32, 33 , 34), and ion probes (Ref 35). These diagnostics are being used to address several mechanistic issues relevant to film growth (Ref 23):

• The role of thermal versus nonthermal evaporation

• The extent to which the laser radiation interacts with the evaporated material

• The effect of the laser-generated plasma on the target

• The expansion mechanism responsible for the high initial particle energy

• The fraction of ionized material from the target surface out to the substrate surface

• The importance of clusters

• The role of the ambient in scattering, diffusion, shock front formation, and gas phase chemistry

The answers to these questions will come not from a single technique, but from a combination of the information provided by each of the diagnostics (Ref 23).

In the laser-generated plasma, the distribution of particles and the particle energy depend on the material, laser wavelength, laser fluence, and ambient composition and pressure. The characteristics are both temporally and spatially dependent. Presumably, this is the reason that optimized deposition conditions vary from system to system. An ideal vapor arrives at the substrate surface compositionally correct and with a kinetic energy that is high enough to allow adequate surface mobility, but not so high as to introduce defects or sputter the depositing vapor. The goal of these studies is to make a direct correlation between the vapor phase properties and the film properties.

Among the plasma diagnostics mentioned above, optical techniques have dominated the research efforts. These techniques are noninvasive and can be performed at the pressures under which films are deposited. Optical techniques have been used to identify the dominant components of the plume as being atomic and small molecular species. Dispersed emission spectra have been measured as a function of distance above the target from the vacuum ultraviolet through the visible (Ref 23, 24, 25, 26). Close to the target a continuum in the emission spectrum is observed, indicative of bremsstrahlung emission and a high density of free electrons. Several centimeters above the target, emission spectra are more characteristic of the isolated atomic species. The presence of electronically excited atoms and molecules at this distance is due to collisional, not optical, excitation. Secondary electrons generated from background gas ionization recombine with highly charged particles to form electronically excited species.

The dispersed spectra have been used to estimate the plasma temperature, which varies from 5,000 to 15,000 K (Ref 26, 36). Particle velocities (v) have been modeled using Maxwell-Boltzmann distributions f(v) ~exp(-mv2/2kT) and shifted distributions about a common center of mass f(v) ~exp(-m(v-vcm)2/2kT) (Ref 30). The shifted Maxwell-Boltzmann arises as a result of a Knudsen layer formed at the target surface (Ref 37, 38). Peak velocities are reported as ~105 to 106 cm/s (Ref 21). Optical techniques have also monitored the formation of gas-phase species from the reaction of the ambient with the ablated material. In these studies, maximum gas phase production has been correlated with optimum film growth pressures. For example, in YBCO, the concentrations of YO and CuO are at a maximum at the optimum deposition pressure for high-quality films (Ref 21, 29).

Optical imaging techniques have also been used to detect the presence of a shock front (region where the plume and background gas pressure are the same) (Ref 39, 40, 41). The position of the shock front depends on the plasma temperature and pressure and the total system pressure. In vacuum, the shock front is relatively far from the target surface (several times the target substrate distance in typical deposition systems). As the ambient pressure is increased, the shock front moves in toward the target. The position of the shock front relative to the substrate is important for optimizing the properties of the deposited film.

0 0

Post a comment