Production of nickel nanoparticles from a nickel precursor via laser pyrolysis

Abstract

The present invention discloses a process for producing nickel nanoparticles. The process involves heating a nickel precursor generated in situ in the presence of a carrier gas under conditions effective to decompose the nickel precursor and produce nickel nanoparticles.

Description

This application claims the benefit of U.S. Provisional Patent Application Ser. No. 60/618,288, filed Oct. 13, 2004, which is hereby incorporated by reference in its entirety.

FIELD OF THE INVENTION

The present invention relates to a process for producing nickel nanoparticles.

BACKGROUND OF THE INVENTION

There is an intense and growing interest in the development of nanostructured magnetic materials, motivated primarily by the immense potential of these materials in a broad range of applications including data storage, spintronics, biomedicine, and telecommunications (Pileni, Advanced Functional Materials 11:323 (2001); Leslie-Pelecky and Rieke, Chem. Mater. 8:1770 (1996); Skomski, J. Phys. Condens. Matter. 15:R841-R896 (2003); Prasad, Nanophotonics John Wiley & Sons, New York (2004)). The synthesis and characterization of such materials are also important for basic science in that they can provide insight into the fundamentals of surface chemistry and magnetic interactions at the nanoscale. Magnetism in the transition elements has been investigated for decades, with the ferromagnetic series—Fe, Co, and Ni—being by far the most extensively studied systems. One fascinating discovery, surface-enhanced magnetism, in which the magnetic moments of small clusters of the ferromagnetic series exceeded their bulk values, occurred in the last decade (Apsel et al., Phys. Rev. Lett. 76:1441 (1996); Billas et al., Science 265:1682 (1994)). These magnetic moments in fact, exhibited dramatically higher values at a certain magic number of atoms. Ever since this discovery, an extraordinary amount of attention has been devoted to understanding the dependence of magnetic properties on particle size and surface treatment. In the development of nanostructured magnetic materials, it is important to control the structure, size, composition, surface characteristics, and self-assembly of the nanoparticles, and to understand how these properties impact the bulk magnetic behavior.

The magnetic behavior of a material depends on numerous factors including the sizes, shapes, and orientations of grains within it, the structure of the grain boundaries, the magnitude and direction of internal stress, the crystallographic phases, the concomitant presence of any other phase, and the overall size and shape of the specimen (Vincent and Sangha, GEC J. Res. 13:2 (1996)). All of these factors can be affected by the particle synthesis procedure. Thus, different methods of preparation are expected to result in different overall properties. This is especially true in nanomaterials because of their high surface to volume ratios. Over the years, a variety of procedures have been developed for the preparation of nanomagnetic materials, including hot colloidal synthesis (Murray et al., IBM J. Res. Dev. 45:47 (2001)), microemulsion (Feltin and Pileni, Langmuir, 13:3927 (1997)), sol gel (Leite et al., J. Nanosci. Nanotechnol. 2:89 (2002)), laser ablation (Jonsson et al., J. Appl. Phys. 79:5063 (1996)), and mechanical milling (Gonzalez et al., Euro. Phys. Lett. 42:91 (1998)). Each of these has advantages and disadvantages relative to the key criteria of controlling particle size, shape, dispersibility in desired solvents, yield, production rate, and processability. It is universally agreed upon that a stable dispersion of uniformly sized particles is desirable for many purposes.

Because of their important potential applications (Hyeon, Chem. Commun. pp. 927-934 (2003)) as pigments, catalysts (Weber et al., J. Nanoparticle Res. 5(3-4):293-298 (2003); Guo et al., Phys. Chem. Chem. Phys. 3(9):1661-1665 (2001)), components of magnetic data storage media (Teng and Yang, J. Am. Chem. Soc. 125(47):14559-14563 (2003)), and elements of chemical and biological sensors (Carpenter, J. Magn. Magn. Mater. 225:17-20 (2001)), the synthesis of nickel-based nanophase materials has attracted considerable interest. Ultrafine magnetic particles (Shi et al., Science 271:937-941 (1996); Majetich and Jin, Science 284:470-473 (1999)) can also be used in magnetic inks and other magnetic fluids (ferrofluids) (Raj et al., J. Magn. Magn. Mater. 149:174-180 (1995)) and in biomedical applications (Berry and Curtis, J. Phys. D: Appl. Phys. 36(13):R198-R206 (2003); Tartaj et al., J. Phys. D: Appl. Phys. 36(13):R182-R197 (2003); Halbreich et al., Biochimie 80(5-6):379-390 (1998); Scherer et al., Gene Therapy 9(2):102-109 (2002); Pankhurst et al., J. Phys. D: Appl. Phys. 36(13):R167-R181 (2003)). Magnetic nanoparticles are also ideal systems for fundamental research in several areas including superparamagnetism, magnetic dipolar interactions, and magnetoresistance. As a result, a significant amount of work has been done to study the preparation and magnetic properties of such particles. A number of techniques have been used for the production of metallic magnetic nanoparticles, such as inert gas evaporation/condensation (Choi et al., Mater. Lett. 56:289-294 (2002); Jonsson et al., J. Appl. Phys. 79(8):5063-5065 (1996); Ullmann et al., J. Nanoparticle Res. 4:499-509 (2002); Wu and Xie, Mater. Lett. 57:1539-1543 (2003)), sonochemisty (Kataby et al., Appl. Surf Sci. 201:191-195 (2002); Jung et al., J. Phys. Chem. Solids 64:385-390 (2003)), coprecipitation (Ge et al., NanoStruct. Mater. 8(6):703-709 (1997); Li et al., Ceram. Int. 28:165-169 (2002)), wet chemical methods (Teng and Yang, J. Am. Chem. Soc. 125(47):14559-14563 (2003); Dubios et al., J. Mol. Liq. 83:243-254 (1999); Bermejo et al., Powder Technol. 94:29-34 (1997); Sun et al., Chem. Mater. 11:7-9 (1999); Chen and Hsieh, J. Mater. Chem. 12:2412-2415 (2002); Hou and Gao, J. Mater. Chem. 13:1510-1512 (2003); Wu and Chen, Chem. Lett. 33(4):406-407 (2004)), microemulsion methods (Guo et al., Phys. Chem. Chem. Phys. 3(9):1661-1665 (2001)), the polyol process (Wu and Chen, J. Colloid Interface Sci. 259:282-286 (2003)), and laser-driven thermal methods (Ullmann et al., J. Nanoparticle Res. 4:499-509 (2002); Miguel et al., IEEE Trans. Magn. 38(5):2616-2618 (2002); Veintemillas-Verdaguer et al., Mater. Lett. 57:1184-1189 (2003); Martelli et al., Appl. Surf Sci. 186:562-567 (2002)).

There are very few reports on vapor-phase synthesis of nickel nanoparticles with diameters below 100 nm, though larger (micron diameter and larger) particles are synthesized commercially in tonnage quantities by thermal decomposition of nickel carbonyl. He et al. reported a UV laser-assisted gas-phase photonucleation process to generate nickel ultrafine particles (UFPs) at ambient temperature with Ni(CO)₄ as precursor (He et al., NanoStruct. Mater. 8(7):879-888 (1998)). Ni(CO)₄ has a high vapor pressure and decomposes cleanly to give pure nickel, and is therefore used in vast quantities in commercial nickel refining. However, its high toxicity has limited its use as a gaseous precursor in laboratory studies.

The present invention is directed to overcoming these deficiencies in the art.

SUMMARY OF THE INVENTION

The present invention relates to a process for producing nickel nanoparticles. The process involves heating a nickel precursor generated in situ in the presence of a carrier gas under conditions effective to decompose the nickel precursor and produce nickel nanoparticles.

CO₂ laser pyrolysis of different chemical vapor deposition (CVD) precursors has proven to be a successful method for preparation of nanoparticles of a variety of materials (Cannon et al., J. Am. Ceram. Soc. 65(7):324-330 (1982); Cannon et al., J. Am. Ceram. Soc. 65(7):330-335 (1982); Li et al., Langmuir 19(20):8490-8496 (2003); Li et al., Langmuir 20(11):4720-4727 (2004), which are hereby incorporated by reference in their entirety) and suggest the possibility of using this method to produce nanopowders with primary particle diameters from 5 to 50 nm at production rates of hundreds of mg per hour in a small bench-scale reactor system. The present invention discloses a method of preparing nickel nanoparticles by laser-driven decomposition of a nickel precursor, such as nickel carbonyl. In this method, an infrared laser rapidly heats a dilute mixture of nickel carbonyl and a photosensitizer in a carrier gas, to decompose the precursor and initiate particle nucleation. To produce nickel nanoparticles, nickel carbonyl was generated in situ from activated nickel powder and CO at room temperature, in order to avoid maintaining an inventory of the highly toxic Ni(CO)₄. During the synthesis process, laser heating allows for rapid cooling of the freshly nucleated particles by mixing with unheated gas. By varying the precursor flow rate, laser energy, and unheated gas flow rate to change the residence time, precursor concentration, and reaction temperature, the average particle size can be controlled over a range of primary particle diameters from 5 to 50 nm. The particle size and crystalline structure have been characterized using X-ray diffraction (XRD), transmission electron microscopy (TEM), nitrogen physisorption surface area measurement (the BET method), and X-ray photoelectron spectroscopy (XPS). For nanosize nickel, which has a low tendency to oxidize, very high surface area powders with mean particle diameter below 10 nm sometimes oxidize violently upon exposure to air. Therefore, conditions that produce somewhat larger particles, 10˜20 nm in average size, were used for studying the effects of reactor operating parameters on particle size and morphology. Magnetization measurements, on the other hand, are presented for smaller particles, 5-8 nm in diameter, since these are the ones that exhibit superparamagnetism and have the most interesting magnetic behavior.

In addition, the present invention discloses a method of using laser-driven decomposition of nickel carbonyl vapors to produce particles in the form of an aerosol, followed by exposure to a solvent containing an appropriate surfactant to yield a stable dispersion of particles. This method is scalable and yields a substantially monodisperse distribution of particles at a relatively high rate of production. The particles produced by this method are subjected to a detailed characterization using transmission electron microscopy, atomic force microscopy, energy dispersive spectroscopy and dc magnetization. The particles have an average diameter of 5 nm, and the observed magnetization curves show no hysteresis above 200 K. The normalized magnetization curves follow a scaling law proportional to the quotient of the applied field over temperature. This data indicates the presence of randomly oriented superparamagnetic particles. The measured magnetization is significantly smaller than that of the bulk, probably due to an effective surface anisotropy and spin canting. The coercivity is the same in either direction of the applied field which indicates that there is negligible exchange coupling between the nickel particles and any possible antiferromagnetic oxide layer on their surfaces.

The present invention combines the advantages of high purity and high throughput achieved in aerosol synthesis with stabilization by surface treatment, as used in colloidal chemistry, to obtain a reasonably stable dispersion of nickel nanoparticles. The method of the present invention is clean in that there are no side products that could adhere to the surface of the particles. Thus, the particles are devoid of any magnetic dead layers, which may be detrimental to the magnetic properties.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic drawing of the reactor system for producing Ni nanoparticles by laser-driven decomposition of Ni(CO)₄.

FIG. 2 is a schematic drawing of the reactor.

FIG. 3 illustrates the nickel nanoparticle production rate and average size vs CO flow rate.

FIG. 4 shows the TEM image and selected area electron diffraction (SAED) pattern from nickel particles with 100 sccm CO flow rate.

FIG. 5 shows the TEM image and selected area electron diffraction (SAED) pattern from nickel particles with 250 sccm CO flow rate.

FIG. 6 shows the XRD pattern from nickel samples with 100 sccm CO flow rate.

FIG. 7 shows the XRD patterns for nickel samples produced using He and Ar as sheath and purge gas.

FIG. 8 shows the TEM image and SAED pattern from nickel particles produced with He sheath and purge gas.

FIG. 9 shows the TEM image and SAED pattern from nickel particles produced with Ar sheath and purge gas.

FIG. 10 shows the XRD patterns from nickel particles produced by experiments (1) and (3).

FIG. 11 depicts high resolution of TEM image of a nickel sample from experiment (3). Both nickel nanoparticles, with a lattice spacing of about 2 Å, outlined by the three smaller circles on the right, and NiF₂ nanoparticles, with a lattice spacing of about 3.1 Å, outlined by the larger circle on the lower left and the half circle at the bottom of the image, are present.

FIG. 12 shows the XRD patterns from nickel particles produced by experiments (2) and (4).

FIG. 13 shows the TEM image and SAED pattern from nickel particles produced with C₂H₄ as photosensitizer.

FIG. 14 illustrates magnetization measurements on a nickel nanoparticle sample, showing magnetization without hysteresis at 300 K.

FIG. 15(a) shows a TEM image of nickel particles deposited from a toluene dispersion. FIG. 15(b) shows a high-resolution transmission electron microscopy (HRTEM) image showing <200> lattice fringes of single particle. FIG. 15(c) shows an electron diffraction pattern from a selected area on the TEM grid.

FIG. 16 is an atomic force microscopy (AFM) picture of Ni particles drop cast from a pyridine dispersion onto a silicon wafer.

FIG. 17 shows an XRD pattern of nickel nanoparticle powder. No phases other than metallic nickel are apparent from the diffractogram.

FIG. 18 shows an energy dispersive X-ray spectroscopy (EDS) spectrum of nickel powder.

FIGS. 19(a)-(c) illustrate magnetization of nickel nanoparticles. FIG. 19(a) illustrates low-temperature results showing magnetic hysteresis. FIG. 19(b) illustrates higher temperature results showing superparamagnetic behavior above 200 K. FIG. 19(c) shows Langevin function fit of the magnetization curve at 300 K. The normalized plot in FIG. 19(b) shows the expected scaling behavior.

FIG. 20 shows zero-field cooling (ZFC) and field cooling (FC) magnetization curves for nickel nanoparticles. The onset of divergence occurs at ˜150 K indicating spin blocking.

FIG. 21 illustrates shifting of the broad maximum in the blocking curve with the applied field magnitude.

FIG. 22 illustrates that the coercive field approximately scales as T^1/4 at low temperatures.

DETAILED DESCRIPTION OF THE INVENTION

The present invention relates to a process for producing nickel nanoparticles. The process involves heating a nickel precursor generated in situ in the presence of a carrier gas under conditions effective to decompose the nickel precursor and produce nickel nanoparticles. Thus, the process of the present invention involves generating the Ni(CO)₄ in situ and immediately reacting it to form nickel nanoparticles and decompose any residual precursor, so that no inventory of the highly toxic Ni(CO)₄ is maintained. In one embodiment of the present invention, the nickel precursor is Ni(CO)₄.

In another embodiment of the present invention, the nickel precursor, Ni(CO)₄, is generated by reacting activated nickel and carbon monoxide. The activated nickel can be generated by heating nickel powder in the presence of hydrogen under conditions effective to remove oxide on the nickel surface.

The step of heating the nickel precursor can involve heating by a directed energy source, preferably a laser, and more preferably a CO₂ laser.

In another embodiment of the present invention, the step of heating is carried out in the presence of a photosensitizer. Examples of photosensitizer include, but are not limited to, sulfur hexafluoride, ethylene, silicon tetrafluoride, and ammonia. The choice of photosensitizer can depend on various factors. For example, with sulfur hexafluoride, a much smaller amount of photosensitizer can be used, but there is a potential for nickel fluoride formation. With ethylene, a much larger amount of photosensitizer is needed, but there is no risk of nickel fluoride formation and there may be some beneficial effects on the particle morphology as well.

The carrier gas used in the method of the present invention can be helium, hydrogen, argon, nitrogen, carbon dioxide, carbon monoxide, or a mixture thereof.

In another embodiment, the process of the present invention further involves collecting the produced nickel nanoparticles on a filter. Examples of filter include, but are not limited to, a cellulose nitrate membrane filter, a cellulose acetate filter, a polyvinylidene fluoride filter, a polytetrafluoroethylene filter, a nylon filter, and a polypropylene filter.

In another embodiment, the process of the present invention further involves collecting the produced nickel nanoparticles into a solution containing a solvent. Examples of solvent include, but are not limited to, toluene, octane, decane, and other aromatic and aliphatic hydrocarbons. In another embodiment of the present invention, the solution further contains a surfactant. Examples of surfactant include, but are not limited to, oleylamine, oleic acid, hexadecylamine, and hexadecanoic acid.

The nickel nanoparticles produced by the process according to the present invention have an average diameter of less than about 50 nm, specifically from about 5 nm to about 50 nm.

EXAMPLES

The following examples are provided to illustrate embodiments of the present invention but are by no means intended to limit its scope.

Example 1

Experimental Methods

The reactor configuration shown in FIGS. 1 and 2 was used to prepare nickel nanoparticles in the experiments described herein. Because nickel carbonyl is highly toxic, it was generated in situ by flowing CO through a tube packed with nickel powder, which was placed inside of a tube furnace. This safe and convenient method was applied to generate a small flow of nickel carbonyl to the reactor without maintaining any inventory of nickel carbonyl. Details of this nickel carbonyl generator were as follows. About 605 g of nickel powder (laboratory grade, reduced powder, Fisher Chemicals, Fairlawn, N.J.) was packed into a 1-inch o.d. stainless steel tube with a packed length of about 14 inches. Glass wool and a porous frit were packed at both ends to prevent the nickel powder from leaving the tube. This nickel powder was heated to 300-350° C. in a stream of flowing H₂ (250 sccm) for at least 60 minutes to remove any oxide on the nickel surface. The tube was then allowed to cool to room temperature under the protection of helium. A parallel line was reserved to purge the system and switch on and shut down the Ni(CO)₄ as required. This line provided a convenient means to avoid toxic nickel carbonyl accumulation in the system. Ni(CO)₄ was generated at the time of use by flowing CO through the activated bed of nickel powder at room temperature or slightly above. Exact flow rates of Ni(CO)₄ were therefore not known. After use, the tube was purged with helium for at least 30 minutes to remove any unreacted Ni(CO)₄ before opening the tube. A similar tube furnace at the outlet of reactor system was used to decompose any unreacted Ni(CO)₄ to nickel film or powder before exhausting the reactor effluent.

The Ni(CO)₄ stream from the generator was typically mixed with additional helium and a small amount of sulfur hexafluoride before entering the reactor. A continuous CO₂ laser beam (Coherent, Inc., Santa Clara, Calif.; Model 42) passed directly above the central reactant inlet, which is made from ⅛-inch o.d. tubing centered within a piece of ⅜-inch o.d. tubing through which a sheath flow of helium enters the reactor. This sheath gas helped to confine the reaction zone to a small region near the axis of the reactor. In all experiments described herein, the laser output was near its maximum value of 60 W. The distance between the inlet nozzle and laser beam can range from 2 to 5 mm and can be changed slightly by sliding the central tube up or down, which provides an additional means for adjustment of the reaction zone. Within a given set of experiments in which some other parameter was varied, the beam position was kept fixed. The ideal dimensions of the reaction zone are almost equal to the beam diameter, which is about 7-8 mm (1/e² power point). Sulfur hexafluoride (SF₆) was added to the precursor stream as a photosensitizer. A photosensitizer was needed because nickel carbonyl does not absorb light appreciably at the laser wavelength of 10.6 microns. SF₆ (technical grade, Aldrich, St. Louis, Mo.) has a large absorption cross-section at the laser wavelength and is stable at high temperatures. The SF₆ absorbed the laser energy and transferred it to the precursor and other gaseous molecules, resulting in very rapid heating of the gas stream. Helium (UHP grade; Irish Welding Supply, Buffalo, N.Y.; passed through an oxygen trap to remove residual O₂ and H₂O) flows confined the reactant and photosensitizer (SF₆) to a region near the axis of the reactor and prevented them from accumulating in the arms of the six-way cross from which the reactor was constructed. These helium streams entering at the ends of the horizontal arms of the reactor were referred to as the purge gas flows, while the gas entering in the outer portion of the concentric nozzle was referred to as the sheath. If the sheath gas flow did not contain any photosensitizer, then it was not heated by the laser. Mixing of the heated particle-containing gas with this cool sheath gas and, eventually, with the purge gas, resulted in rapid cooling of the freshly formed particles and prevented their coalescence into larger particles. This was essential to producing small particles at high throughput. All gas flow rates to the reactor were controlled by mass flow controllers. The resulting particles were collected on cellulose nitrate membrane filters. Particles can also be collected directly into a surfactant solution using a bubbler connected in parallel with the collection filter. The solvent used for particle collection was usually toluene with surfactants (octadecylamine or oleic acid) added to improve particle dispersion. Very high surface area powders, with mean particle diameter below 10 nm, sometimes oxidize violently upon exposure to air. Therefore, larger nickel nanoparticles (10˜20 nm average diameter), which were more stable in air, were used for studying the effects of operating parameters on particle size and morphology. Smaller particles could be protected from oxidation by collecting them directly into surfactant solution in the bubbler. In experiments designed to produce the smallest particles (˜5 nm diameter), the laser beam was focused above the precursor inlet to reduce the size of the reaction zone to 1-3 mm in diameter.

Transmission electron microscopy (TEM) was performed using a JEOL JEM 2010 microscope (Peabody, Mass.) at an acceleration voltage of 200 kV to characterize particle size and morphology. Samples were prepared for TEM imaging as follows. Powder samples that had been collected on membrane filters were dispersed in toluene under an inert atmosphere. Multiple drops of the dispersion were cast onto a carbon-coated TEM grid and the toluene was allowed to evaporate in air. Selected-area electron diffraction was performed in the TEM to determine the crystalline structure. TEM grids were also prepared by evaporating drops of the dispersion of particles produced by direct collection of the particles into solution. There was no significant difference in morphology or degree of agglomeration between samples prepared in these two ways. When oleic acid or oleylamine was added to the toluene dispersions before casting them onto the TEM grid, the degree of agglomeration observed in TEM was somewhat lower. However, residual surfactant reduced the quality of both the images and the small-angle electron diffraction patterns obtained in the TEM. Specific surface area was measured by nitrogen physisorption (the BET method; Micromeretics Model 2010 ASAP Physisorption Apparatus, Norcross, Ga.) and was used to estimate the average particle size. Wide-angle powder X-ray diffraction (XRD; Siemens D500, New York, N.Y.) was used to characterize powder samples, identifying the crystalline phases present and giving an estimate of crystalline domain size. X-ray photoelectron spectroscopy (XPS) (also called Electron Spectroscopy for Chemical Analysis, or ESCA) was performed on a Physical Electronics/PHI 5300 X-ray photoelectron spectrometer (Chanhassen, Minn.) for elemental analysis of surface of the powder samples. DC magnetization measurements were made using a superconducting quantum interference device (SQUID) MPMS C-151 magnetometer from Quantum Design (San Diego, Calif.). The powder sample was dispersed in a high boiling solvent, such as dodecane or hexadecane, and put in a nonmagnetic gel capsule that was then wrapped with layers of Teflon tape to prevent breakage under vacuum. The magnetic field was ramped from 0 to 10,000 G at both 300 K and 5 K.

Example 2

Collection of Nickel Particles

The nickel particles collected on filters were loosely agglomerated and black in color. They were attracted to a permanent magnet, both as a powder and when dispersed in a solvent. They were readily dispersed in nonpolar solvents, such as toluene or hexane, and their dispersion was improved by the addition of oleic acid or oleyl amine. However, they did not form colloidal dispersions with long-term stability. Over a period of hours to days, depending on particle size, presence of surfactant, and particle preparation conditions, particles agglomerated and sedimented out of solution. This process was accelerated in the presence of a magnetic field. Many parameters, such as gas flow rates, the carbonyl generator temperature, and the gases used as the photosensitizer and sheath gas, can affect the nanoparticle production rate, size, and morphology.

Example 3

Effect of Carbon Monoxide (CO) Flow Rate

After regenerating the nickel powder in the nickel carbonyl generator with flowing H₂ at about 300° C., carbon monoxide (CO) can react with the active nickel surface to form nickel carbonyl at room temperature. The concentration and flow rate of Ni(CO)₄ into the laser-driven reactor can be changed by varying the CO flow rate, and this, in turn, affects the production rate and the average size of the nickel nanoparticles. FIG. 3 shows the results that were obtained by fixing all other operating parameters and changing only the CO flow rate. The operating conditions are listed in Table 1. The average size obtained from XRD and BET experiments shown in FIG. 3 are different, since XRD estimates the average crystalline domain size, but nitrogen physisorption (the BET method) measures the specific surface area of the sample. From the surface area, the average particle size was calculated assuming that the particles are spheres with the density of bulk nickel. Both of these methods of estimating average particle size show the same trend in particle size with respect to changes in the CO flow rate.

TABLE 1
Typical Reaction Parameters
Operating	SF6 (inlet)	SF6 (sheath)	He (sheath)	Purge He Flow
Pressure	flow rate	flow rate	flow rate	Rate
12.0˜12.1 psia	11.9 sccm	2.6 sccm	145 sccm	1300 sccm

With increasing CO flow rate (from 30, 60, 100, 150, 250, 350 to 450 sccm), the production rate first increased and then decreased. This can be rationalized as follows. The main factor determining the production rate at low CO flow rates is the supply of Ni(CO)₄ to the reactor. When the CO flow rate was lower than 150 sccm, the total supply of Ni(CO)₄ to the reactor increased with increasing flow rate. However, when the CO flow rate increased more, even though the absolute Ni(CO)₄ flow rate may have continued to increase, the concentration of Ni(CO)₄ in the precursor stream decreased, if the kinetics of Ni(CO)₄ generation from Ni and CO in the generator limited the production rate of Ni(CO)₄. With higher flow rates, the residence time in the laser beam was shorter, and a smaller fraction of the Ni(CO)₄ entering the reactor was converted to particles. Therefore, the production rate decreased with further increase in CO flow. For the average particle size, the residence time and Ni(CO)₄ concentration are more important than the total supply of Ni(CO)₄ to the reactor. With increasing CO flow rate, the residence time and Ni(CO)₄ concentration decreased and, as a result, the average particle size decreased. When the CO flow rate was higher than 250 sccm, the average particle size increased slightly with increasing CO flow rate.

Particles were increasingly prone to deposit on the inlet nozzle at higher CO flow rates. Deposition of particles on the nozzle and reactor walls also contributed to the apparent decrease in production rate, since these particles were not collected. TEM was used to characterize the size and morphology of these nickel particles. FIG. 4 and FIG. 5 show TEM images and selected area electron diffraction patterns (inset) from two samples produced with 100 and 250 sccm CO flow rates, respectively. The SAED pattern in FIG. 5 was taken from the larger particles in the TEM image. From the TEM images, it is clear that for the larger CO flow rate, some large, agglomerated, and (partially) fused particles are present. On the basis of these results, the increase in average particle size and the increasing tendency for particles to deposit on the inlet nozzle at high flow rates can be attributed to recirculation of gases in the top of the reactor. As shown schematically in FIG. 2, there is potential for recirculation of gases in the region above the laser beam. Presumably, with increasing CO flow rate, the recirculation of mixed gases already containing nickel nanoparticles was stronger. Reheating of these recirculated particles then led to their growth and (partial) sintering to form larger particles with nonspherical shape. At the same time, some recirculated particles were deposited on the nozzle and reactor walls. The average particle size increased as a result, even as the particle size for non-recirculated particles continued to decrease slightly.

In FIGS. 4 and 5, as well as other TEM images shown herein, the nickel nanoparticles were substantially agglomerated. Because the particles were produced at high number concentrations, estimated to be about 10¹¹ particles per cm³ for typical conditions within the reactor, some coagulation prior to collection downstream is inevitable. Because the TEM grids were prepared from particle dispersions, it was not clear how much coagulation occured in the reactor system and how much occured during solvent evaporation when the TEM grids were prepared. Direct thermophoretic sampling onto TEM grids within or just after the reactor would allow this question to be addressed.

Wide-angle powder X-ray diffraction (XRD) was also used to characterize all of these samples. The results were identical except for the slight changes in peak width reflected in the size estimates shown in FIG. 3. FIG. 6 shows the X-ray diffraction pattern from the sample that was prepared at a CO flow rate of 100 sccm. All of these samples showed the major characteristic peaks for pure crystalline metallic nickel at 2θ values of 44.5 [Ni-111] and 51.8 [Ni-200] degrees. This indicates that there was no significant amount of crystalline NiO or other crystalline material formed, and that there was not a large amount of amorphous material present, since no broad peaks indicative of an amorphous phase were observed.

Example 4

Effect of the Inert Gas

An inert gas was used both as the sheath gas (entering the reactor in the concentric inlet surrounding the precursor inlet) and as the purge gas (entering the reactor at the ends of the four horizontal arms of the six-way cross) during particle synthesis. Experiments to compare helium (He) with argon (Ar) as sheath gas and purge gas in the system are described herein. Table 2 lists the operating parameters used in the experiments and the resulting production rate and mean particle size. The average size is based on XRD peak broadening.

TABLE 2
Effect of Sheath and Purge Gas Properties
Operating	SF6 (inlet) flow	CO flow	Sheath flow	Purge flow
Pressure	rate	rate	rate	rate
12.0˜12.1 psia	4.0 sccm	100 sccm	580 sccm	1300 sccm
	Experiment (1): Using Helium	Experiment (2): Using Argon
Production	360	mg/hr	140	mg/hr
Rate
Average	17.4	nm	19.8	nm
Size

Helium and argon have the same heat capacity, but helium has a higher thermal conductivity of about 140 mW/(m K)) compared to 16 mW/(m K) for argon (at 273 K and 1 bar). Therefore, when the gas used as sheath gas and purge gas was changed, reactor conditions were significantly affected. The reaction zone temperature was increased when argon was used rather than helium. For these conditions, using argon, a “flame” could be clearly observed above the inlet nozzle, due to thermal emission from the particles produced. At the beginning of the experiment, the emission was the brightest. With increasing reaction time, the “flame” turned weaker and faded. By the end of the experiment, it had almost totally disappeared. This observation is discussed in more detail below. As shown in Table 2, using helium as the sheath and purge gas led to a higher production rate and smaller average particle size than the same nominal conditions using argon as the sheath and purge gas. However, the apparent difference in production rate is based only on particles collected on the filters downstream of the reactor. When argon was used as the carrier gas, there was greater deposition of particles on the inlet nozzle and reactor walls than when helium was used. This may be the source of the apparent difference in production rate, rather than any genuine change in the amount of Ni(CO)₄ converted to particles.

XRD and TEM were also used to characterize the particles. FIG. 7 shows the XRD results for the two samples. The sample produced using helium as sheath and purge gas showed only the characteristic peaks for nickel (˜44.5 [Ni-111], and ˜51.8 [Ni-200] degree) particles. The sample prepared using argon as the sheath and purge gas showed not only characteristic peaks for nickel but also four more characteristic peaks at ˜27.0 [NiF₂-110], 34.9 [NiF₂-101], 41.2 [NiF₂-111], and 53.2 [NiF₂-211] degrees, which suggested that nickel fluoride was present in the sample. FIGS. 8 and 9 are TEM images and SAED patterns from nickel particles produced with He or Ar as sheath and purge gases, respectively. In FIG. 8, most particles were smaller than 20 nm while in FIG. 9 almost all the particles were larger than 20 nm, which confirmed the trend shown in the XRD results. The electron diffraction rings of the samples shown in FIG. 8 and FIG. 9 were consistent with the XRD results. The rings in FIG. 8 could all be indexed to the cubic lattice of nickel, while, in FIG. 9, additional, less intense rings were present that could be indexed to elemental sulfur and to NiF₂.

These samples were also characterized using X-ray photoelectron spectroscopy (XPS or ESCA). XPS is an extremely surface-sensitive technique that samples only the topmost 1-2 nm of the sample. As a result, it measures the composition at the particle surface and not the overall composition, even for nanoparticles of the size considered herein. These measurements confirmed that the sample prepared using argon as the sheath and purge gas had a small amount of sulfur and fluorine contamination from the photosensitizer (SF₆) but that the sample made using helium as the sheath and purge gas did not. Thus, using argon as the sheath and purge gas can result in a higher temperature in the reaction zone, which leads to the presence of NiF₂ and sulfur in the products.

Even though SF₆ is a very stable compound, NiF₂ formation from Ni and SF₆ is a thermodynamically favorable process. Overall reactions leading to the formation of NiF₂ from Ni and SF₆ include
Ni(s)+SF₆(g)NiF₂(s)+SF₄(g)  (1)
Ni(s)+⅓CO(g)+⅓SF₆(g)NiF₂(s)+⅓SCO(g)  (2)

Reactions (1) and (2) are both exothermic, with ΔH₁^rxn(298 K)=−194 kJ/mol and ΔH₂^rxn(298 K)=−254 kJ/mol. Equilibrium calculations show that, with at least a 1 to 3 SF₆ to Ni ratio, these reactions would result in complete conversion of Ni(s) to NiF₂(s) under all attainable reaction conditions if equilibrium were achieved. For example, an initial mixture representative of experiment (2), containing 93.3% Ar, 6.1% CO, 0.29% SF₆, and 0.27% Ni(CO)₄, equilibrated at 0.5 bar and 1000 K gave 92.4% Ar, 0.18% SF₆, 0.02% SF₄, 0.08% SCO, 7.0% CO, and 0.27% NiF₂. At higher or lower temperatures, only the amounts of SF₄ and SCO formed changed Thus, if the reaction mixture were to reach temperatures high enough for these reactions to occur on a timescale of milliseconds, only NiF₂ and no metallic nickel would be expected to form. However, SF₆ alone is stable on this time scale to temperatures of at least 1500 K, and therefore is usually considered to be inert. It appears that in this system, there was some SF₆ decomposition, possibly catalyzed by the nickel nanoparticles themselves or by some gas-phase nickel-containing species.

Example 5

Effect of Photosensitizer

The basic principle of this reaction system is that the laser energy is used to heat the precursor to decompose it and induce particle nucleation. Ni(CO)₄ does not absorb at the operating wavelength of the CO₂ laser. Therefore, a photosensitizer gas must be added to the precursor stream. In addition to SF₆, the most common sensitizer used in this method, ethylene (C₂H₄, 99.5+%, Aldrich, St. Louis, Mo.) was also used as a photosensitizer. Ethylene and SF₆ are both strong absorbers at 10.6 microns, however, the absolute absorbance of SF₆ is much greater than that of ethylene in this range. Other potential photosensitizers for use with a CO₂ laser include silicon tetrafluoride (SiF₄) and ammonia. SiF₄ absorbs very strongly near 9.6 microns, and CO₂ lasers can be made to operate at that wavelength. Ammonia absorbs strongly near 10.6 microns, but the wavelength of the CO₂ laser falls between two absorption lines of ammonia. However, it should be possible to use ammonia as a photosensitizer with a CO₂ laser tuned to one of the many other CO₂ laser lines near 10.6 nm. Unfortunately, Ni(CO)₄ does not have any strong absorption matching the wavelength of CO₂ lasers or other lasers that can be economically used at high (tens of Watts to kilowatt) powers. Based on the IR absorbance spectra of SF₆ and C₂H₄, nickel particle synthesis experiments were designed. Table 3 lists the reaction conditions that were used.

TABLE 3
Reaction Parameters Used to Study the Effect of the Photosensitizer Gas
system pressure: 12˜12.1 psia, CO flow rate: 100 sccm
(1) SF6 (inlet): ˜4 sccm,
    He (sheath): ˜580 sccm, He (purge): ˜1300 sccm
(2) SF6 (inlet): ˜4 sccm, Ar (sheath): ˜580 sccm, Ar (purge): ˜1300 sccm
(3) SF6 (inlet): ˜11.9 sccm, SF6 (sheath): ˜2.6 sccm,
    He (sheath): ˜580 sccm, He (purge): ˜1300 sccm
(4) C2H4 (inlet): ˜150 sccm, Ar (sheath): ˜364 sccm,
    Ar (purge): ˜1300 sccm

Experiments (1) and (2) have been discussed in Example 4, in the context of the effect of the sheath and purge gas used. Comparing experiments (1) and (3) demonstrates the effect of SF₆ flow rate. With higher SF₆ concentration, more laser energy was absorbed, resulting in higher temperature in the reaction zone. Thus, it might be expected that NiF₂ and S would be present in the product, as was the case when argon was used as the sheath and purge gas. XRD (FIG. 10) and TEM (FIG. 11) analyses confirmed this expectation. In FIG. 10, there were four peaks at ˜27.0, 34.9, 41.2, and 53.2 degrees, in addition to the nickel peaks at 44.5 [Ni-111] and 51.8 [Ni-200] degrees. These peaks correspond to NiF₂ and possibly also crystalline sulfur. XPS experiments were again carried out to identify the elemental concentrations at the surface of the samples. Sulfur and fluorine were detected in the products from experiments (2) and (3), but not in the products of experiments (1) and (4). The fluorine concentration in sample (3) was substantially higher than in sample (2). These results confirmed that using higher reaction temperature and SF₆ flow rate can lead to fluorine and sulfur contamination. FIG. 11 shows a high-resolution TEM image of a sample from experiment (3), for which XRD showed that both cubic nickel and NiF₂ were present. By measuring the lattice spacing in the TEM image, it can be determined which particles in the picture are nickel (lattice spacing of 2.0˜2.1 Å) and which are NiF₂ (lattice spacing of 3.1˜3.2 Å), thus confirming that discrete Ni and NiF₂ particles are formed, rather than Ni/NiF₂ composite particles or core-shell structures.

In experiment (4), nickel particles were produced using ethylene (C₂H₄) as the photosensitizer. In this case, of course, sulfur (S) and fluorine (F) were not found in the products. In order to compare experiment (4) with experiment (2), XRD and TEM results are shown in FIGS. 12 and 13. In FIG. 12, there were a total of nine characteristic peaks in the two spectra shown, for which the 2-θ values were 27.1, 34.9, 39.1, 40.1, 41.5, 44.6, 51.8, 53.1, and 58.4 degrees (from left to right). All of these peaks clould be indexed to four crystal structures: NiF₂ (27.1 [NiF₂-110], 34.9 [NiF₂-101], 40.06 [NiF₂-111], and 53.06 [NiF₂-211] degrees), cubic Ni (44.507 [Ni-111] and 51.8 [Ni-200] degrees), hexagonal Ni (39.1 [Ni-010], 41.5 [Ni-002], 44.5 [Ni-011], and 58.4 [Ni-012] degrees), and sulfur (27.08 [S-222] degree). Combining the discussion of experiment (4) with that in Example 4, it can be concluded from XRD results that, when C₂H₄ was used as photosensitizer, both hexagonal and cubic nickel were produced. This conclusion was also supported by the SAED pattern in FIG. 13. The ring diameters were measured as 9.6 mm (only several dots shown in the picture with low intensity), 10.2 mm (a clear ring with many bright dots), 11.8 mm (a ring with many dots), and 13.1 mm (several dots). This indicated that hexagonal nickel (9.6, 10.2, and 13.1 mm) and cubic nickel (10.2 and 11.8 mm) were present in the sample from experiment (4). Comparing FIG. 13 with other TEM images, it was found that, when C₂H₄ was used as the photosensitizer, the particles are relatively spherical, showing less sintering and necking compared with samples produced using SF₆ as the photosensitizer. It was also observed that the particles produced with C₂H₄ as the photosensitizer were much more readily dispersed in nonpolar solvents, such as toluene or hexane, than the particles produced using SF₆. Both the decreased sintering and improved dispersibility may have resulted from the presence of carbon on the particle surface. No graphite phase was seen in the XRD or SAED results, but it is still possible that a small amount of amorphous carbon was present on the particle surface.

Example 6

Effect of Precursor Concentration

To supply Ni(CO)₄ to the reactor, reaction of CO with activated nickel powder was used to generate Ni(CO)₄ just upstream of the particle synthesis reactor. Therefore, in addition to the CO flow rate, there are also other factors that influence the Ni(CO)₄ supply to the reactor, such as the amount of nickel powder inside the generator, the state of the nickel powder surface, and the temperature in the generator. The effects of temperature and of hydrogen activation of the nickel surface were studied qualitatively. Silicone rubber extruded heating tape (SRT Series, Omega Engineering, Stamford, Conn.) was used to heat the bottom of the generator tubing (outside the furnace) until the temperature at that location stabilized at 77° C. while Ar gas was flowing through the tubing. The furnace itself was not heated, but simply served as insulation to maintain the temperature inside the generator. The experimental details and results are listed in Table 4.

TABLE 4
Effect of Carbonyl Generator Temperature and Time Online
Experiment (1)
with carbonyl generator	Experiment (5)
at room temperature	with carbonyl generator heated
1st hour	2nd hour	1st hour	2nd hour	3rd hour
357	mg/hr	235	mg/hr	506	mg/hr	394	mg/hr	344	mg/hr
17.4	nm	15.4	nm	17.8	nm	18.6	nm	18.1	nm

If particles were made with the generator at room temperature, the production rate decreased with time (starting from freshly regenerated powder in the generator), which indicated that activation state of the nickel surface could affect the Ni(CO)₄ concentration delivered. With increasing time online, after regeneration, the Ni(CO)₄ production rate decreased. This explains why the flame was weaker and faded with increasing the reaction time. Lower Ni(CO)₄ concentration also resulted in smaller nickel particles. When the precursor tubing was heated, the rate of conversion of nickel powder to Ni(CO)₄ in the generator increased. Therefore, the production rate with the generator slightly heated was greater than that with the generator at room temperature. Again, the production rate decreased with time online. However, the average size from XRD measurements did not show any meaningful trend in particle size accompanying the change in production rate caused by changing the temperature of the carbonyl generator.

Example 7

Other Factors

There are several other factors that affect the synthesis process, including laser position (relative to the inlet nozzle), sheath and purge gas flow rates, and operating pressure. Lower laser beam position tended to make larger particles because there was less diffusion of Ni(CO)₄ into the surrounding gas before reaching the laser beam. This effectively increased the Ni(CO)₄ concentration in the reaction zone. Larger sheath and purge gas flow rates led to shorter residence time in the reaction zone, which resulted in decreased particle size. Decreasing the operating pressure lowered the concentration of both precursor and photosensitizer, decreased the residence time, and increased diffusion of precursor and photosensitizer out of the laser-heated region. Thus, it tended to decrease the average particle size. Table 5 shows the effect of decreasing total pressure on the particle size, as determined by XRD peak broadening.

TABLE 5
Reaction Parameters and Results (Effect of Operating Pressure)
He (sheath) flow			Purge He flow
rate	SF6 (inlet) flow rate	CO flow rate	rate
580 sccm	4.0 sccm	100 sccm	1300 sccm
Experiment (1):	Resulting particles: ˜17.4 nm
Operating pressure 11.9˜42.0 psia
Experiment (6):	Resulting particles: ˜15.4 nm
Operating pressure 7.9˜8.0 psia
Experiment (7):	Resulting particles: ˜14.2 nm
Operating pressure 5.9˜6.0 psia

Example 8

Methods to Measure Ni(CO)₄ Concentration

As described above, the exact Ni(CO)₄ concentration entering the reactor from the carbonyl generator was not known. However, the nickel carbonyl concentration and delivery rate is a key parameter in determining particle size, production rate, and efficiency of the system in converting nickel carbonyl to nickel powder. One method used to obtain an approximate measurement of the Ni(CO)₄ delivery rate is described herein. A sample flow of the nickel carbonyl containing gas was diverted into two bubblers in series containing a bromine/alcohol solution. The nickel carbonyl dissolved in the alcohol and quickly reacted with the bromine to form NiBr₂. The bubbler solution was collected and analyzed to determine the nickel content. The carbonyl concentration in the original gas sample was then calculated using the amount of nickel collected, the gas flow rate, and the gas collection time. The configuration was like that shown in FIG. 1, except that only the collection bubbler line was open, and the powder collection line was closed. During sampling, the laser was shut down and the stream containing Ni(CO)₄ and CO was mixed with He to control the operating pressure at around 11.7 psi. The flow rate of CO through the precursor tubing was 100 sccm with 580 sccm He as the sheath gas and 1300 sccm He as purge gas (nominally the same conditions as experiment (1), but with the laser turned off). After using He gas to carefully purge the system, contents of both bubblers were mixed together in a flask, and the solvent and unreacted bromine were evaporated. Nitric acid (˜70%) was added to dissolve the NiBr₂ salt remaining in the flask, and a clear green NiBr₂ solution was obtained. The resulting sample contained 0.6145 g of nickel. This corresponded to a Ni(CO)₄ delivery rate of 1.43 g/hr, or 0.008 mol/hr, or 3.1 sccm. About 13% of the CO entering the generator was converted to nickel carbonyl for these conditions (room temperature). Comparing this delivery rate to the production rate shown in Table 4 during an experiment using nominally the same conditions with the laser turned on, it was estimated that 60-70% of the Ni(CO)₄ delivered to the reactor was converted to particles.

Example 9

Magnetic Properties of Nickel Nanoparticles

Nickel is, of course, ferromagnetic as a bulk material. However, the smallest particles produced here are expected to be superparamagnetic. A single result is described herein, demonstrating the superparamagnetic behavior of one sample of nickel nanoparticles. A more complete and detailed analysis of the magnetic properties of the nickel nanoparticles is described in Example 11. Measurements of the magnetization of particles produced in experiment (10) are shown in FIG. 14. The particles showed typical superparamagnetic behavior, with no hysteresis at 300 K and a high saturation field of about 2 Tesla.

TABLE 6
Typical Reaction Parameters for Making Small Nickel Nanoparticles
Experiment (10)
			He	He
Operating	SF6 (inlet)	CO flow	(inlet)	(sheath)	Purge He
Pressure	flow rate	rate	flow rate	flow rate	flow rate
8.4˜8.5 psia	4.0 sccm	30 sccm	55 sccm	320 sccm	2545 sccm

In conclusion, laser driven pyrolysis of nickel carbonyl was used to produce nickel nanoparticles. A safe, convenient method was implemented to supply a small flow of nickel carbonyl for laser-induced synthesis of nickel particles. Hydrogen-regenerated nickel powder and carbon monoxide were used to prepare nickel carbonyl in a stream that flowed directly to the reactor and was immediately decomposed back to nickel and CO. While this carbonyl generation method was convenient for producing small amounts of very small nickel nanoparticles, it limited the experiments to relatively low Ni(CO)₄ concentrations. The present invention considered factors that affect the nickel particle size, production rate, and morphology. TEM images, selected area electron diffraction patterns, XRD, BET, and XPS (ESCA) analysis were used to characterize particles. Particles ranging from 5 to 50 nm in average diameter were obtained. CO flow rate affected particle size as well as production rate in this system by changing the Ni(CO)₄ concentration and delivery rate and, at high flow rates, by altering the flow patterns in the reactor and inducing recirculations. The choice of inert gas used for sheath and purge flows affected the temperature in the reaction zone via the thermal conductivity and heat capacity of the gas. Higher temperature led to appearance of a visible flame in the reaction zone. Both SF₆ and C₂H₄ can be used as photosensitizers. When using a relatively high SF₆ flow rate, which led to relatively high temperature in the reaction zone, F and S were present in the sample as NiF₂ and elemental sulfur. Use of C₂H₄ as the photosensitizer changed the particle morphology and crystal structure as well as eliminating S and F contamination. Laser position and power, inert gas flow rate, and operating pressure could also be used to vary particle size and production rate. In magnetization measurements, powder samples of nanoparticles about 8 nm in diameter showed magnetization without hysteresis at room temperature. That is, they were superparamagnetic at room temperature, as would be expected for such small particles.

Example 10

An Aerosol-Mediated Magnetic Colloid: Study of Nickel Nanoparticles

Ni(CO)₄ vapor, generated by flowing CO through a bed of nickel powder, was mixed with helium and ethylene and flowed into a reactor through a concentric nozzle, surrounded by a flow of helium. A CO₂ laser (Coherent, Model 42 laser emitting up to 60 W) was focused onto the flowing mixture. The laser energy was absorbed by ethylene, which served as a photosensitizer, resulting in very rapid heating and decomposition of Ni(CO)₄. Since the surrounding flow of helium was unheated, the hot gas rapidly cooled by mixing with this unheated gas, quenching particle growth after a few milliseconds. The Ni particles were collected on cellulose nitrate membrane filters. The collection filters were opened in a nitrogen glove box to prevent rapid oxidation of the nickel nanopowders. The particles were dispersed in toluene, using oleic acid (97%, Aldrich) as a surfactant. Ultrasonication was used to aid in dispersion of the particles.

TEM was performed on a JEOL JEM 2010 microscope at an acceleration voltage of 200 kV. Samples were prepared for by casting a drop of toluene dispersion on a carbon-coated TEM grid. Selected-area electron diffraction (SAED) was performed in the TEM to determine the crystalline structure of the particles. Atomic force microscopy (AFM) was performed in tapping mode on a Dimension 3100 AFM (Digital Instruments, Veeco Metrology Group, Santa Barbara, Calif.). Imaging was done with Veeco NanoProbe Tips model #RTESP14 with resonant frequencies of 210-280 kHz and amplitude set point 1.4-1.6 V. Wide-angle powder X-ray diffraction was done on a Siemens D500 using the Cu K_α line as the X-ray source. The 2θ angles ramped were from 20° to 90° to cover all the major peaks expected from nickel. Energy dispersive X-ray spectroscopy (EDAX) was obtained with a Hitachi S-4000 field emission scanning electron microscope (Ibaraki, Japan) operating at an acceleration voltage of 20 kV. The sample was cast as a thin film from the dispersion or simply analyzed as a powder on a silicon or graphite substrate. The X-ray fluorescence beams were collected with an X-ray collection unit IXRF 500 system. Magnetization measurements (DC) were made using a superconducting quantum interference device (SQUID) MPMS C-151 magnetometer from Quantum Design Corporation. Magnetization hysteresis scans were recorded with a DC magnetic field ramped to 2 Tesla in both directions.

The Ni nanoparticles generated by laser pyrolysis of Ni(CO)₄ were spherical with an average diameter of 5 nm, as indicated in the bright field TEM image in FIG. 15(a) and the AFM image in FIG. 16. The particles, treated with surfactants such as oleic acid or oleyl amine, can form a clear dispersion in nonpolar solvents, such as hexane and toluene. These dispersions were found to be stable without sedimentation for at least several days. The particles also formed clear dispersions in pyridine, which acted as a coordinating solvent, without the need of any additional surfactant. As shown in the TEM image, the particles tended to agglomerate as the solvent evaporated from the TEM grid and they self-assembled. This may have been due to strong dipolar interactions between the particles. The high-resolution transmission electron microscopy (HRTEM) showed clear lattice fringes from the <111> planes of a spherical particle with a lattice spacing of ˜2 Å (FIG. 15(b)). The SAED pattern clearly showed diffraction rings that can be identified with the lattice planes of the Ni crystal (FIG. 15(c)). The concentric rings <111>, <200>, <220>, <222>, <420>, <333, 511>, and <442> bore the radii ratios 1, 1.15, 1.68, 2.0, 2.63, 2.96, and 3.48, respectively, very close to the theoretical values (Cullity, Elements of X-Ray Diffraction, 2nd ed., Addison Wesley (1978), which is hereby incorporated by reference in its entirety), suggesting that the fcc lattice of the bulk Ni was well maintained at this particle size.

The crystallinity of the particles was corroborated by a clean XRD pattern with peaks at 2θ values of 44.09°, 51.70°, and 76.09° that are associated with the <111>, <200>, and <220> lattice planes of nickel, respectively (FIG. 17). The presence of nickel oxide phases could not be ruled out because the position of nickel <111> coincided with the strongest peak from nickel oxide, and hence there could be contribution from such a phase. But the absence of any distinguishable peak at other characteristic positions such as 2θ=37°, expected for nickel oxide, indicated that the presence of an oxide phase was insignificant. The Ni lattice constant was calculated using the Rietveld Analysis program (Young et al., J. Appl. Cryst. 28:366 (1995), which is hereby incorporated by reference in its entirety) DBWS-9807 and found to be 3.546 Å, 0.4% larger than the bulk value of 3.532 Å. This was consistent with the general trend of lattice expansion in small nanocrystallites (Wu et al., Phys. Rev. B 69:125415 (2004), which is hereby incorporated by reference in its entirety). The average particles size estimated from all of the XRD peaks using the Scherrer formula was 4.9 nm, which was in good agreement with the AFM and TEM images. The composition of the particles was also confirmed by energy dispersive x-ray spectroscopy (EDS). The EDS spectrum, shown in FIG. 18, exhibited only peaks characteristic of nickel.

The magnetization curves taken at different temperatures are shown in FIGS. 19(a) and 19(b). No hysteresis was observed at temperatures above 200 K indicating that the particles were superparamagnetic in this range. As the sample was cooled, hysteresis curves systematically developed and widened, approaching a coercivity of 640 Oe at the lowest temperature of 5 K. The temperature dependence of the zero field cooling (ZFC) and field cooling (FC) curves taken at an applied field of 400 Oe are shown in FIG. 20. These data showed a rounding up of the ZFC curve at ˜200 K, which correlated with the onset of superparamagnetism. In the superparamagnetic phase, the magnetization curves for an isotropic sample scaled with temperature such that M/M_s plotted against H/T superimposed over each other. This scaling law was fairly followed with minor deviation (≦10%) probably due to interparticle interaction, in the temperature range 200-325 K (FIG. 19(b)), emphasizing their superparamagnetic character.

The Langevin function for a classical Heisenberg spin under the influence of a field H is used to determine the effective magnetic moment μ per particle, i.e., M/M_S=L(a)=coth(a)−(1/a), where a=μH/kT, and μ=M_sV_avg is the particle magnetic moment. The fitting of the Langevin function to the superparamagnetic curves is shown in FIG. 19(c). By extrapolating the high-field end of the magnetization curve and normalizing for the total particle mass in the 300-K data, the saturation magnetization was found to be 31.40 emu/g as compared to a bulk value of 55.0 emu/g. The observed value translated into an effective magnetic moment of 189 μ_B per particle, which was calculated using an average particle diameter of 5 nm, and a bulk density for Ni of 8.9 g/cm³. The lower value of saturation magnetization can arise from factors, such as lattice defects that depend on the fast kinetics of Ni(CO)₄ decomposition, nanoparticle nucleation, and crystallization that occur on a time scale of milliseconds during particle synthesis. The surface anisotropy and spin canting in the particles also usually contribute to the decrease in the total magnetization (Morales et al., J. Phys. Condens. Matter. 9:5461 (1997); Kodama et al., Phys. Rev. Lett. 77:394 (1996), which are hereby incorporated by reference in their entirety).

It is noteworthy that the coercivities of the magnetization curves had the same magnitude in either direction of the applied field (FIG. 19(a)). If the surface of the Ni particles had a NiO coating layer with an appreciable thickness, then the particles could be viewed as ferromagnet-antiferromagnet core-shell structures. If the oxide coating is very thin, it will have a negligible impact on the magnetization, as was observed in Co nanoparticles with a thin CoO shell embedded in nonmagnetic matrix (Skumyrev et al., Nature 423:850 (2003), which is hereby incorporated by reference in its entirety). On the other hand, for thicker coatings, the exchange coupling at the core-shell interface could result in a shift of the hysteresis loop along the field axis when the system is cooled below the Néel temperature of the antiferromagnetic phase. Previous studies have found exchange bias fields (asymmetry in the coercive fields) on the order of 700 Oe in partially oxidized Ni particles (Kodama et al., Phys. Rev. Lett. 79:1393 (1997), which is hereby incorporated by reference in its entirety). In NiO particles, Kodama et al. discovered large loop shifts (>10 kOe) which were attributed to a weak coupling between antiferromagnetic sublattices (Fonseca et al., Phys. Rev. B 66:104406 (2002), which is hereby incorporated by reference in its entirety). The findings disclosed herein closely resemble those of Fonseca et al. in which no exchange-coupled behavior was observed when Ni particles were formed in a silica matrix through a sol-gel method (Kodama et al., Phys. Rev. Lett. 79:1393 (1997), which is hereby incorporated by reference in its entirety). Thus, this data suggests that the sample in the measurement was devoid of any significant oxide coating, probably because the particles were prepared and then dispersed into a surfactant-containing solution in an oxygen-free environment, and the surfactant may have provided some protection against subsequent oxide formation.

The ZFC and FC curves, measured at a bias field of 400 Oe, overlapped at temperatures above 200 K as shown in FIG. 20. Moreover, no hysteresis was observed above this temperature. The FC showed a steady decrease with the rise of temperature, expected from progressive randomization of the particle magnetic moments as the anisotropy barriers were overcome thermally. The ZFC exhibited a broad cusp implying a distribution of blocking temperatures. The blocking temperatures systematically decreased as the bias field was increased from 50 G to 800 G in FIG. 21.

The blocking temperature T_B in the ZFC curve depends on various factors including the size of the nanoparticles and their intrinsic magnetic anisotropy, as well as interparticle interactions. Recently, Kechrakos et al. observed an increase in T_B derived from the ZFC curve of a dense dispersion of ferromagnetic particles, and attributed it to the anisotropic and ferromagnetic character of the dipolar interaction (Kechrakos and Trohidou, Appl. Phys. Lett. 81:4574 (2002), which is hereby incorporated by reference in its entirety). Also, Liu et al. have shown systematic changes in T_B with variation in size of the nanoparticles (Liu and Zhang, Chem. Mater. 13:2092 (2001), which is hereby incorporated by reference in its entirety).

In the present invention, the broadness in the ZFC curve, which reflects a distribution of blocking temperatures, was most likely due to interactions between particles. The blocking temperature T_B is defined as the temperature that renders a magnetic relaxation time of τ=10² sec. As discussed by Dormann et al., the relaxation time, τ, for a superparamagnetic particle is given by the modified Brown formula (Dormann et al., J. Mag. Mag. Mater. 183:L255 (1998), which is hereby incorporated by reference in its entirety), $\begin{array}{cc}\tau ={\tau }_0\exp \left[\frac{\left(E_B\right)}{\mathrm{kT}}\right],& \left(1\right)\end{array}$
where E_B=E_Ba+E_Bint; E_Ba=K is the energy barrier of the individual particles and E_Bint is the additional barrier due to interactions. Here, K is the anisotropy constant and V is the volume of the particle. The preexponential factor τ₀ also depends on E_Ba and E_Bint, i.e., $\begin{array}{cc}{\tau }_0=\frac{\sqrt{\pi }}{4}\frac{\mu \left(0\right)}{E_B{\gamma }_0}\left[\frac{1}{{\eta }_r}+{\left(\frac{{\mu }_{\mathrm{nr}}\left(T\right)}{{\mu }_{\mathrm{nr}}\left(0\right)}\right)}^2\right]\times \sqrt{\frac{\mathrm{kT}}{E_B}}\left(1+\frac{\mathrm{kT}}{E_B}\right),& \left(2\right)\end{array}$
where μ(0) is the particle magnetic moment at 0 K, μ_nr(0) the corresponding nonrelaxing magnetization, γ₀ is the gyromagnetic ratio, and η_r is the reduced damping constant (η_r=ηγ₀μ_nr(0)). Thus τ, and hence T_B, depends on both the size (V) of the particles, as well as the interparticle interactions, and in the latter case with a more complex functional dependence. In the samples studied herein, the particles were fairly monodispersed as indicated by the TEM and AFM images in FIG. 15(a). Therefore, since the particle size was relatively constant, it is likely that the distribution of blocking temperatures was due to the variation in interparticle interactions E_Bint that results from the random spacing of nearest neighbors across the sample. It is notable that Puntes et al., Appl. Phys. Lett. 78:2187 (2001), Hyeon et al., J. Am. Chem. Soc. 123:12798 (2001), and Sun and Murphy, J. Appl. Phys. 85:4325 (1999), which are hereby incorporated by reference in their entirety, also found broad blocking peaks on a regularly arranged two dimensional superlattice of various monodispersed magnetic nanocrystals, where the size polydispersity would hardly contribute to the peak broadness.

Further, according to the theory of superparamagnetism (Fonseca et al., Phys. Rev. B 66:104406 (2002); Bean and Livingstone, J. Appl. Phys. 30:120S (1959), which are hereby incorporated by reference in their entirety), rotation of the superparamagnetic spins in the presence of a field is thermally activated, provided kT is greater than E_Ba, the anisotropy energy. For a system of randomly oriented non-interacting identical particles, the coercivity is expected to follow the relation $H_c\left(T\right)=H_c\left(0\right)\left[1-{\left(\frac{T}{T_B}\right)}^{1/2}\right],$
where H_c(0)=0.64 K/M_s, and K is the anisotropy constant. The plot of H_c, vs T is given in FIG. 22 where it is noted that the above law, applicable for randomly oriented independent particles, was not exactly followed. It was rather found that exponent was ˜¼ up to a temperature 30 K, as shown in the inset. This was to be expected because of the interparticle interactions in the sample, as reflected in the ZFC curve. Several authors have reported that the T^1/2 power law only applies in a limited low-temperature range for interacting nanoparticle systems. For example, Fonseca et al., Phys. Rev. B 66:104406 (2002), which is hereby incorporated by reference in its entirety, observed that the law holds for temperatures up to ˜16 K in a silica matrix with doped Ni particles. Sun and Dong, Mater. Res. Bull. 37:91 (2002), which is hereby incorporated by reference in its entirety, found that it applied to graphite encapsulated Ni/NiO particles up to ˜120 K, and McHenry et al., Mater. Sci. Eng. A 204:19 (1995), which is hereby incorporated by reference in its entirety, found it to hold for carbon coated SmCo particles for temperatures up to ˜25 K. Brunsman et al., J. Appl. Phys. 79:5293 (1996), which is hereby incorporated by reference in its entirety, discovered that for the interacting Nd—Fe—B—C nanoparticles, the power law was obeyed with an exponent of 0.3 instead of 0.5. It appears from the above-cited reports that the non-interacting character of the particles was enhanced by a robust solid surface coating, such as encapsulation of a graphite layer, rather than a labile surfactant layer, because the latter still can not shield the long-range dipolar forces among the particles.

In conclusion, Ni nanoparticles produced by laser driven pyrolyisis of Ni(CO)₄ were, with proper parametric controls, very monodispersed and had clean surfaces with no evidence of oxidation. The dipolar interaction among the particles may be responsible for suppressing the independent character of the particles. This may entail the broadness of the blocking behavior while the field-temperature scaling in the superparamagnetic regime remains intact. In view of the fact that the crystal field effects of ligands on the surface of the nanoparticles significantly affect the electronic structure of the former and alter the magnetic properties, these particles can serve as a good model system where different chosen ligands can be bound on the surface and magnetic properties systematically studied.

Although the invention has been described in detail, for the purpose of illustration, it is understood that such detail is for that purpose and variations can be made therein by those skilled in the art without departing from the spirit and scope of the invention which is defined by the following claims.

Claims

1. A process for producing nickel nanoparticles comprising:

heating a nickel precursor generated in situ in the presence of a carrier gas under conditions effective to decompose the nickel precursor and produce nickel nanoparticles.

2. The process according to claim 1, wherein said nickel precursor is Ni(CO)4.

3. The process according to claim 2, wherein said nickel precursor is generated by reacting activated nickel and carbon monoxide.

4. The process according to claim 3, wherein said activated nickel is generated by heating nickel powder in the presence of hydrogen under conditions effective to remove oxide on the nickel surface.

5. The process according to claim 1, wherein said heating comprises heating by a directed energy source.

6. The process according to claim 5, wherein said directed energy source is laser.

7. The process according to claim 6, wherein said laser is a CO2 laser.

8. The process according to claim 1, wherein said heating is carried out in the presence of a photosensitizer.

9. The process according to claim 8, wherein the photosensitizer is selected from the group consisting of sulfur hexafluoride, ethylene, silicon tetrafluoride, and ammonia.

10. The process according to claim 1, wherein the carrier gas is selected from the group consisting of helium, hydrogen, argon, nitrogen, carbon dioxide, carbon monoxide, and a mixture thereof.

11. The process according to claim 1 further comprising:

collecting the produced nickel nanoparticles on a filter.

12. The process according to claim 11, wherein the filter is a cellulose nitrate membrane filter, a cellulose acetate filter, a polyvinylidene fluoride filter, a polytetrafluoroethylene filter, a nylon filter, or a polypropylene filter.

13. The process according to claim 1 further comprising:

collecting the produced nickel nanoparticles into a solution comprising a solvent.

14. The process according to claim 13, wherein the solvent is selected from the group consisting of toluene, octane, and decane.

15. The process according to claim 13, wherein the solution further comprises a surfactant.

16. The process according to claim 15, wherein the surfactant is selected from the group consisting of oleylamine, oleic acid, hexadecylamine, and hexadecanoic acid.

17. The process according to claim 1, wherein the nickel nanoparticles have an average diameter of less than about 50 nm.

18. The process according to claim 17, wherein the nickel nanoparticles have an average diameter of from about 5 nm to about 50 nm.