BRIEF DESCRIPTION OF THE DRAWING
FIG. 1. is a three-part graphic diagram where part (a) is a plot of ion current versus laser fluence where thresholds. Part (b) illustrates particle density dependence on fluence, and part (c) illustrates particle size dependence on fluence. Filled triangles represent mesoparticles; filled circles represent nanoparticles in parts (b) and (c). Fth1 and Fth2 are indicated by the two vertical dashed lines.
FIG. 2. illustrates the system of the invention including a vacuum chamber (and related pumps, not shown), a target manipulator, an ion probe (Langmuir probe), a gas inlet, and a substrate manipulator, where the laser beam is focused onto the target surface through a fused silica window.
FIG. 3. is a two part diagram wherein part (a) is an AFM image of Ni nanoparticles generated at a fluence of 0.4 J/cm2 and part (b) shows the particle size distribution.
FIG. 4. illustrates the conventional Gaussian intensity distribution of a laser beam, where thresholds are indicated as horizontal lines. Above Fth2, at the center of a focal spot, the laser is intense enough to fully vaporize the material, which reduces the yield of nanoparticles. The tall and short curve illustrate an intense and a weak beam profile, respectively. The dotted square line illustrates a flat-top beam profile.
FIG. 5. is a two part diagram proving a comparison of the background gas effect: in part (a) CoO particles are generated in vacuum; in part (b) CoO particles are generated in 30 millitorr Argon.
FIG. 6. is a three part HRTEM image of a NiO nanoparticle obtained in 30 millitorr oxygen by ultrafast laser ablation of a metal Ni target, where parts (a) and (c) are high resolution images which show the single crystal structures and part (b) is a nanoelectron beam diffraction pattern which shows the NiO(100) diffraction.
FIG. 7. is a two part diagram showing HRTEM images of a nanoparticle with a structure that has a Ni core and NiO shell. This is obtained by first ablating a metal Ni target in 30 millitorr argon and then exposing the sample to oxygen at room temperature. The core-shell structure is evident in part (b). FFT analysis confirms the core shell structure.
DETAILED DESCRIPTION OF THE INVENTION
FIG. 2 illustrates the system used in this invention. The system includes a vacuum chamber 1 pumped by a turbo pump and a mechanical pump (not shown), a target manipulator 2 which provides for rotational and lateral movement for four targets of different materials, a substrate manipulator 3 which provides heating and rotational and lateral movements for the substrate 10, a gas inlet 4, and an ion probe 6 (Langmuir probe) to measure the ion current of the ablation plume. When measuring the ion current, the ion probe is biased −50 V relative to the ground to collect the positive ions in the plume (the number of negative ions in plasma is negligible). An ultrafast laser (not shown in the figure) is positioned outside the chamber and the laser beam 20 is focused onto the target surface through a fused silica window 21. The laser has a pulse duration between 10 fs-50 ps, preferably between 10 fs-1 ps; a pulse energy between 100 nJ-10 mJ; and a repetition rate greater than 1 kHz. The metal nickel and metal oxide CoO are used as example materials in this invention, but this invention is not limited to these materials, because the physics behind particle formation during ultrafast pulsed laser ablation applies similarly to both metals and insulators (including semiconductors, metal oxides, and polymers).
FIG. 1 (a) shows the collected ion current (i) versus the laser fluence (F). It is clearly recognizable that above a fluence of about 1.0 J/cm2, the ion current gains significantly with the fluence, while below this value, the ion current is nearly vanishing. This behavior of the ion current can be understood by considering ultrafast pulsed laser ablation of metals in several generalized stages: (1) electrons gaining energy from the intense laser irradiation, (2) electron thermalization with the lattice (here lattice means the atoms that constitute the solid target), (3) breakdown of the chemical bonds and formation of plasma, and (4) cooling and expansion of the plasma into a vacuum. For ultrafast pulsed laser ablation, stages (1) and (2) occur within a few picoseconds and have been described by the so-called two-temperature model, which predicts that after electron-lattice thermalization, the system temperature T is proportional to the fluence, and the ablation depth (L) is related to the fluence through
L˜(Dτ)1/2 ln(F/Fth1),
where D is the thermal diffusivity and r is the pulse duration. (See, S. Nolte, C. Momma, H. Jacobs, A Tunnermann, B. N. Chichkov, B. Wellegehausen and H. Welling, Journal of the Optical Society of America B, Vol 14, No. 10, 2716, 1997.) For stage (3) of ionization and plasma formation, for simplification, the Saha-Boltzmann equation can be applied, which predicts that for a low temperature plasma (which holds for most laboratory plasma), the ionization fraction a can be estimated as
α2˜noG(T)3/2 exp(−U/T),
where no represents the neutral density, G(T) is a slow varying function of the temperature, and U is the first ionization energy of the constituent element. Combining the expressions for the ablation depth and the ionization fraction, the amount of ions can be estimated as i˜ALNα, where A is the area of the focal spot and N is the target material density. It is therefore understood that the fast increase of the ion current at high fluences is a result of a high level of ionization, which comes from a high temperature (after electron-lattice thermalization). If the temperature is close to the ionization energy U, intense plasma can be formed. Therefore an i-F plot provides an easy approach for estimating the threshold Fth2 of plasma generation, which is about 1.0 J/cm2 in the case of FIG. 1(a).
FIGS. 1(
b) and 1(c) display the density and particle size dependences on the fluence. It is seen that at high fluences, both types of particle acquire stabilized size and density, while at low fluences (<1.0 J/cm2), the size and density of the mesoparticles vanish quickly, and the size and density of the nanoparticles vanish very slowly. The threshold Fth1, below which no material removal occurs, can be found by extrapolating the size and density of both particles asymptotically to the horizontal (fluence) axis.
Therefore a combination of the data from FIGS. 1 (a), (b), and (c) provides convenient and practical guidance for controlling the particle size distribution by setting the fluence below the plasma generation threshold Fth2 and above the fundamental breakdown threshold Fth1. Within this region, the large mesoparticles can mostly be eliminated, as exemplified in FIG. 3, and significant plasma formation is avoided, which results in a good yield of nanoparticles.
FIG. 3(
a) is an AFM image of a sample obtained by placing a substrate 5 cm away from the target and collecting the particles for 1 min. The laser used in this example has a 1 kHz repetition rate; the fluence is 0.4 J/cm2. It is seen that the particle size distribution peaks significantly at 1.6 nm, as shown in FIG. 3(b). In FIG. 3(b), the particle height is used to represent the particle size, assuming that the particle is a hard sphere. This assumption is valid because of the effects of cooling rates and the background gas, which will be discussed later.
That the appearance of mesoparticles is coincident with the beginning of strong plasma formation suggests a role played by the intensity distribution of the laser beam. The TEM00 mode used by most ultrafast lasers has a Gaussian type intensity distribution. FIG. 4 schematically illustrates an intense and a weak beam profile of Gaussian type. The two thresholds Fth1 and Fth2 are indicated by the two horizontal lines. For the intense beam profile (the taller curve), most of the center portion is above the plasma formation threshold Fth2, as a result, the materials exposed to the center of the focal spot will be mostly vaporized and transformed into plasma, and only the edge of the focal spot (the hatched area) is within the two thresholds and contributes to particle formation. Therefore, with the intense beam, first, the nanoparticle yield is low. Second, the vaporized center portion applies a strong recoil force to the melts and causes mesoscale liquid droplets to be splashed out, which is the source of the mesoparticles. On the other hand, for the less intense beam (the shorter curve), the center portion falls within the two thresholds and contributes to nanoparticle formation without causing much plasma generation. Optimally, a ‘flat-top’ beam profile (as illustrated by the dashed thick line) improves the control of particle size distribution and nanoparticle yield. Transformation of a Gaussian profile into a flat-top beam profile can be achieved with a number of known approaches, for example using diffractive optics.
Supplying a background gas of sufficient pressure can also help to improve the nanoparticle size distribution. In vacuum, the particles (after being formed during ablation) cool down through black-body radiation, which is an inefficient way of losing heat. It can be shown that for black-body radiation, the cooling rate dT/dt of a small spherical particle satisfies
dT/dt=−(3M/rρCp)eσ(T4−T04),
where M, r, ρ, and Cp represent the molar mass, radius, and heat capacity of the material, T0 is the ambient temperature (room temperature), e is the emissivity, and σ is the Stephen-Boltzmann constant. It can therefore be estimated that for a Ni particle of a diameter of 10 nm, it takes on the order of 0.1 μs for the particle to cool from 2500° C. (a typical temperature reached by ultrafast pulsed laser ablation) to its melting point (1455° C.). This time scale may be shorter than the time it takes for the particles to reach the substrate, depending on the target-substrate distance. However, liquid-solid phase transition requires the release of the latent heat. Also, solidification also requires presence of nucleation centers for solidification to be initiated; otherwise the liquid will remain super-cooled even below its melting point. Therefore, it is possible that the particles can maintain the liquid state when flying in vacuum. And it is further possible in this case that the liquid particles can easily change shape or even break into pieces when smashing onto a hard substrate surface. Supplying the system with gas helps to cool down the flying particles by heat exchange during collision with the gas molecules. Collision can also provide density disturbance on the liquid droplet surface, which can introduce nucleation centers. These effects help the nanoparticles to solidify before they reach the substrate.
FIG. 5 compares the shape of the nanoparticles obtained in vacuum [1×10−7 torr, FIG. 5(a)] and in 30 millitorr argon [FIG. 5(b)] under otherwise identical conditions. In this case, metal oxide CoO is used as the target material. It is evident that the background argon gas changes the shapes from shallow and smashed droplets to round ping-pong-ball-like hard spheres, and the particle sizes are also smaller because of the shape change.
When the background gas is reactive, for example oxygen, there are additional benefits that can help to bring in new chemical and structural properties. First, the ablated metal can react with oxygen during ablation to form metal oxide nanoparticles. Second, by simply exposing the metal particles to oxygen after ablation, nanoparticles with oxide shell and metal core structures can also be formed. Two examples are shown in FIG. 6 and FIG. 7.
FIG. 6 shows HRTEM images of a Ni nanoparticle obtained by ultrafast pulsed laser ablation in 30 millitorr oxygen. It is evident that single crystal cubic NiO nanoparticles are formed, as clearly seen FIG. 6(c). This is also confirmed by electron beam diffraction in FIG. 6(b), which displays the NiO(100) diffraction.
Another example is shown in FIG. 7. By first forming Ni nanoparticles in 30 millitorr argon and then exposing the nanoparticles to oxygen, nanoparticles with Ni-core-NiO-shell structures are obtained. The core-shell structures are especially evident in FIG. 7(b). Fast Fourier transformation analysis (not shown) performed at the shell and core regions also confirms the core-shell structures.
The scope of protection of the invention is not limited to the examples given hereinabove. The invention is embodied in each novel characteristic and each combination of characteristics, which particularly includes every combination of features which are set forth in the claims, even if this feature or this combination of features is not explicitly mentioned in the specification, the claims or in the examples.
Patent Application
20080006524
