WO2006008175A1 - Rapid flame synthesis of doped zno nanorods with controlled aspect ration - Google Patents

Abstract

Described is a flame spray pyrolysis (FSP) method for the production of nanorods wherein at least one dopant metal precursor is added to a precursor solution of the predominant metal in amounts of at most 15 atom-% referred to the predominant metal, said dopant having an higher valency than the predominant metal, and wherein the crystal lattice of the pure metal oxide formed by FSP is in a hexagonal closed package. In preferred embodiments, the coordination number of the dopant atoms is higher than the coordination number of the predominant metal atoms and the dopant atoms and/or ions and the predominant metal atoms/ions have sizes varying by less than 20 atom-%.

Description

ID FLAME SYNTHESIS OF DOPED ZNO NANORODS WITH CONTROLLED ASPECT RATION

Cross References to Related Applications

This application claims the priority of US provisional application no. 60/590,363, filed July 23, 2004, the disclosure of which is incorporated herein by reference in its entirety.

Technical Field

The present invention concerns a flame spray pyrolysis method (FSP) for the production of metal oxide nanorods, in particular zinc oxide nanorods with controlled aspect ratio. The invention especially concerns a single-step, continuous, and readily scaleable method.

Background Art

Anisotropic materials exhibit non-uniform properties in different characteristic directions. In particular, crystals that exhibit non-uniform morphology- are a particularly interesting class of anisotropic materials. Nanorods, as addressed herein, are nanoscale solid structures with one characteristic dimension larger than the other. This disparity gives an aspect-ratio (ratio of length to width) greater than unity and a dominant linear, 1-dimensional nature to the structure. Nanorods are similar in geometry to nanowires, nanofilaments and nanofibers, the term rod, however, implies a shorter length.

Depending on the particular material composition, nanorods may exhibit many unique electronic, optical and mechanical properties that appeal to a range of applications. The state of the art suggests the use of nanorods in the area of electronics, e.g. in transistors, diodes, interconnects and other nanoscale circuit features (Wang, 2003; Xia et al . , 2003), and due to their high surface-to-volume ratio they are suggested as constituents for electro-chemical sensors (Xia et al . , 2003) . Proposed optical applications are e.g. field- emission displays, photoluminescent materials, optical switches and non-linear optical filters and converters (Xia et al . , 2003) . Nanoscale linear structures such as nanorods are also disclosed to exhibit enhanced strength compared to the analogous bulk materials (Xia et al . , 2003) . Proposed mechanical and electro-mechanical applications include high-strength polymer-composites (Manhart et al . , 2000), actuator devices (Lin and Chang, 2002) and strain sensors. Due to the linearity of the nanorods, Tani et al . suggest their use in novel conducting polymers or thermo-electric applications (Tani et al . , 2001) .

Nanorod synthesis techniques typically include vapor or wet-chemistry routes. Vapor phase methods include evaporation/condensation (Sun et al . , 2004; Wen et al . , 2003) , pulsed-laser deposition (Kawakami et al., 2003), chemical vapor deposition (Liu et al . , 2004), or gas-phase reaction (Yan et al. , 2003) . In particular, the evaporation/condensation technique has been widely used to synthesize anisotropic materials including nanorods, nanowires and nanobelts (Kong et al., 2004; Wen et al., 2003; Yan et al . , 2003) . Wet chemistry methods include solvothermal (Patzke et al . , 2002) , hydrothermal (Cheng and Samulski, 2004; Liu and Zeng, 2003) , and sol-gel techniques (Guo et al., 2002; Vayssieres, 2003) . While these techniques are capable of well-controlled material uniformity and properties, they are limited in their scale-up potential.

A variety of mechanisms have been proposed for formation of nanorods. For nanoscale inorganic crystals such as zinc oxide (ZnO) , the introduction of certain impurities such as indium (Wen et al . , 2003), tin (Gao et al . , 2003) , selenium (Sun et al . , 2004), antimony (Zeng et al . , 2004) , cerium (Cheng et al . , 2004) and other elements was found to induce preferential growth within specific crystal planes (Kong et al., 2004) resulting in asymmetric crystals. Another nanorod growth possibility is through control of the reactant composition ratio. Yan and coworkers made ZnO nanorods (among other interesting morphologies) by carefully tuning the ratio of oxygen to zinc precursor (Yan et al., 2003) . The oxygen stoichiometry has been shown also to influence nanorod growth during evaporation/condensation (Leung et al . , 2004; Tseng et al . , 2003) and chemical vapor deposition (Liu et al . , 2004) syntheses.

Flame synthesis is a technique that can be readily scaled to produce nanoscale materials in high- volume at low-cost (Mϋller et al . , 2003) . Flame-generated materials are generally dominated by spherical primary particles and chain-like agglomerates (Pratsinis, 1998) . Carbon nanotubes have been made using flames (Height et al., 2003; Height et al . , 2004) however such pronounced linear features as characteristic for nanorods have not been observed for flame-synthesized inorganic materials. Short rod-like particles have been observed in flames before (Akhtar et al . , 1992), however their formation was neither controlled nor a dominant feature of the synthesized materials (Tani et al . , 2002a) . Thus, it is still a very interesting challenge to selectively form anisotropic structures in a flame environment.

Disclosure of the Invention

Hence, it is a general object of the inven¬ tion to provide an improved method for the production of inorganic nanorods, in particular a method allowing the scale-up of the production.

It is a further object of the present invention to provide flame made nanorod products.

It is still a further object of the present invention to provide products comprising such nanorods or the use of such nanorods in the manufacture of such products, respectively.

Now, in order to implement, these and still further objects of the invention, which will become more readily apparent as the description proceeds, the flame spray pyrolysis (FSP) method for the production of metal- oxide nanorods is manifested by the features that at least one dopant metal precursor is added to a precursor solution of the predominant metal in amounts of 1 to 15 atom-% referred to the amount of predominant metal, said dopant having an higher valency than the predominant metal, and wherein the crystal lattice of the pure metal oxide formed by FSP is in a hexagonal closed crystal structure.

Preferably, the coordination number of the dopant atoms within the product crystal lattice is higher than the coordination number of the predominant metal atoms. Also preferred is that the dopant atoms and/or ions and the predominant metal atoms/ions have sizes varying by less than 20 % referred to the atomic radius of the predominant metal as 100 %.

As mentioned above, in the scope of the present invention, flame spray pyrolysis (FSP) is used for the synthesis of metal oxide powders, wherein the metal oxide particles are at least in part, preferably predominantly, in form of nanorods. Preferred nanorod materials are ZnO powders.

FSP is a combustion-based process whereby a liquid solvent containing various dissolved precursor species is injected and atomized into a flame (Madler et al., 2002) . The carrier solvent provides energy for combustion and the flame temperature causes evaporative release of the precursors from the spray droplets into the gaseous environment where reactions proceed to form nanoparticles of the desired composition. The use of liquid precursors gives wide flexibility and accurate control of precursor composition while the high temperature and reactive conditions of the flame provide a favourable environment for the rapid formation of nanosized particles. FSP has been demonstrated in the synthesis of a wide variety of metal and mixed-metal oxides, including ZnO (Tani et al . , 2002a) and mixed ZnO/SiO₂ (Tani et al . , 2002b) .

It has now been found, that the above described FSP method can be further developed such that it allows the rapid and direct synthesis of nanorod particulate metal oxides with close control of the nanorod geometry, in particular the aspect ratio, through addition of suitable dopant precursors (in the case of ZnO e.g. indium and tin dopant precursors) into the precursor solution that is injected and atomized into the flame .

The effect of dopants in the FSP is especially noticeable in the case of metal oxides forming hexagonal close packed (HCP) crystal structures such as ZnO crystallites having wurtzite lattice structure. Addition of indium or tin dopant species to such ZnO dramatically reduced the size of the (002) plane, leading to nanorod features with controlled aspect ratios as confirmed by TEM and BET analysis. Doping with lithium gave no change in the ZnO crystal texture. Nanorod formation in general can primarily be attributed to dopants with higher valency and/or coordination relative to the predominant metal of the metal oxide. In the case of doped ZnO both, indium and tin, have higher valency and coordination compared to zinc. The incorporation of dopants having the above outlined characteristics leads to structural disruption and reduced growth within the suitable lattice planes.

Besides the associated valency and degree of coordination of the dopant atoms, it was assumed that also the atomic or ionic size might have an influence. However, it was found that the size is only relevant in so far as similarly sized dopants are readily incorporated into the metal plane of the metal oxide crystal where they have a disruptive coordination due to the different valency from the predominant metal. This limits the growth in the crystal plane parallel to the metal atoms thereby altering the aspect ratio of the crystal. For example, the atomic radius of indium, tin, and lithium (In 1.55 A; Sn 1.45 A; Li 1.45 A) (Slater, 1964) are similar to zinc (1.35 A) (Slater, 1964) and the atomic radii of Sn and Li are the same. Since the results obtained with these dopants do not reflect this similarity, the size of the dopant atoms is unlikely to be a relevant parameter for alteration of the shape of the crystal to induce the observed rod-like geometry. It has, however, a relevance in so far as similarly sized dopants are more readily incorporated into the crystal lattice and thus bear less risk of influencing the crystal lattice geometry.

While atomic/ionic size is not considered a relevant distinguishing factor between the dopant species, a higher valency of the dopant than the valency of the predominant metal is assumed to be a relevant driving force for nanorod formation as well as coordination effects. The higher valency of dopants relative to the predominant metal lead to greater coordination with surrounding 0 atoms than for the predominant metal's atoms within the same crystal plane. It is assumed - without wanting to be bound by any theory - that the higher valency dopant atoms have a disrupting influence on the metal plane of the metal oxide lattice, hindering crystal growth within the metal layer (Kong et al . , 2004) , i.e. that a dopant with a higher valency than the predominant metal has a greater tendency to strongly bond with oxygen atoms in the crystal and so selectively disrupt the bonding within specific crystal lattice planes leading to the inhibited growth in that plane and an elongated crystal shape for the overall particle. This hindering effect is observed in the decreasing (002) crystal plane size as the dopant concentration is increased. The more pronounced nanorod forming influence of some of the dopants may be attributed to higher valency, and larger disruptive influence, of e.g. tin with regard to indium. Conversely, lower valency means that such dopant atoms have minimal disruptive influence within the metal atom layer and so little effect on the size of the (002) plane.

The elemental composition of the obtained nanorods will be identical to the metal composition in the FSP liquid precursor solution as confirmed by energy dispersive x-ray spectroscopy (EDXS) .

While in general only one dopant, metal will be incorporated, due to the similar sizes preferred, also different dopant metals may simultaneously be incorporated by using a precursor solution comprising more than one dopant precursor.

The FSP synthesis method delivers precursor compounds in liquid, or rather dissolved form to a high temperature flame environment where the precursors vaporize, oxidize, and subsequently form metal-oxide clusters that coagulate into nanometer-scale particles. Particle growth can continue via surface reactions and collisions with other particles (Madler et al., 2002) . Precursor reaction and particle formation occur in the vapor-phase, therefore the initial stage of the nanorod formation process is the formation of primary particles from the vapor. In the early stages of the flame, particles are at a high temperature, for many metal oxides at a temperature higher than their melting point. This is e.g. true for ZnO having a melting point of 1970 ⁰C. Particles formed in the early stages of the flame are spherical in shape and liquid-like. However, in the latter region of the flame, particles cool rapidly, solidify and crystallize to be collected on a filter.

It is assumed - without wanting to be bound by these theories - that the dopants influence the particle formation and morphology in two regions of the flame. Firstly, during the initial stages of the flame, the presence of the dopants may affect the size of the primary particles by influencing the initial cluster formation and the sintering and growth rate of the particles (Akhtar et al . , 1994) . Secondly, during the late stages of the flame, the gas temperature decreases significantly and the particles crystallize. During this cooling process, the dopant species within the particles are assumed to begin to influence the formation of the crystal structure as described previously, disrupting growth in the (002) plane. The final nanorod shapes therefore could be attributed to the annealing and rearrangement of the particle during the later stages of the flame. An alternative mechanism based on epitaxial growth of the nanorods via selective surface growth of the alternating metal (in particular Zn) and 0 planes is unlikely for the flame system based on the observed relatively short rod lengths, distribution of sizes, and uniformity of crystal structure within the product particles. The short rod lengths (<100 nm) observed, and the absence of reactor wall surfaces during the synthesis, indicate that the growth proceeds exclusively via gas-phase processes, in contrast to vapor- condensation methods (Leung et al . , 2004; Liu et al . , 2004; Tseng et al . , 2003; Wen et al., 2003) . The formation of these structures in the gas-phase makes FSP an appealing process for continuous and scaleable synthesis of nanorod materials. Large scale production of nanorods with this method is straight-forward and has been demonstrated for other FSP systems (see Muller et al . 2003, the disclosure of which is incorporated herein in its entirety) .

In general, the FSP is performed as described by Muller et al . 2003, but in particular within the following parameter ranges:

- Precursor liquid feed rate: 5 to 50 ml/min

- Total metal concentration: 0.1 to 3.0 mol/L

- Dopant concentration in precursor (with respect to total metal atoms) : 1 to 15 at%

- Maximum flame temperature: 1900 to 2500 ⁰C The metal oxide of the present invention will in general be predominantly in the form of nanorods . Predominantly here means that analysis of the XRD pattern for the bulk powder according to the fundamental parameter method (described elsewhere in this document) will yield an aspect ratio (length to diameter) greater than 1, preferably greater than 1.5 and most preferably greater than 2.

Suitable precursors for the metal oxide are e.g. organometallic compounds, in particular zinc naphthenate, and suitable dopant precursors are e.g. also organometallic compounds, in particular indium , acetylacetonate and stanneous-2-ethyl hexanoate .

The metal oxides in nanorod form of the present invention exhibits at least one of the following properties:

Unique electronic, optical and mechanical properties that make them suitable for application in a wide range of applications including electronics, sensors, optical components and displays, polymer- composites and actuator devices. At present, the use in polymer composites is preferred. Applications for nanorods in the area of electronics are, e.g. in transistors, diodes, interconnects and other nanoscale circuit features. Further uses are as constituents for electro-chemical sensors and/or in optical applications such as e.g. field-emission displays, photoluminescent materials, optical switches and non-linear optical filters and converters and/or in providing enhanced strength compared to the analogous bulk materials, whereby proposed mechanical and electro-mechanical applications include high-strength polymer-composites, actuator devices and strain sensors. Further applications are in novel conducting polymers or thermo-electric applications .

Brief Description of the Drawings

The invention will be better understood and objects other than those set forth above will become ap¬ parent when consideration is given to the following de¬ tailed description thereof. Such description makes refer¬ ence to the annexed drawings, wherein:

Figure 1 is a schematic diagram of the flame spray pyrolysis (FSP) nozzle (shown in cross-section) and associated flow control and sample collection systems (Madler et al . , 2003) .

Figure 2 shows the powder X-ray diffraction (XRD) patterns for pure ZnO and 1 to 10 at% In-doped ZnO. Crystal plane indices for each peak are shown in the upper frame.

Figure 3 shows crystallite size analysis. The top panel (Figures 3a,b,c)gives detail of XRD patterns for ZnO doped with a) In, b) Sn and c) Li. The peaks correspond to the (100) , (002), and (101) planes of the wurtzite structure at 31.8, 34.5, and 36.3 2Θ degrees respectively. The lower panel (Figures 3d,e,f) indicates control of ZnO crystal plane sizes (Δ-(100) ; O- (002) ) as a function of dopant concentration for d) In, e) Sn, and f) Li, together with BET equivalent diameter (D) for each of the doped ZnO samples. The Sn-doped ZnO shows a steeper variation than the In-doped ZnO, in agreement with XRD. Annealed samples (filled symbols) and associated temperatures are shown in d) and e) .

Figure 4 shows the aspect ratio control, wherein the left part, designated a, shows the ZnO lattice along with the calculation of XRD lattice aspect ratio from the ratio of (100) to (002) XRD crystal plane size measurements. The right part, designated b, shows the XRD lattice aspect ratio as a function of dopant concentration for In- (0), Sn- (Δ) , and Li-doped (O) ZnO and annealed 6 at . % In-ZnO (filled diamonds) with the corresponding temperatures. The inset graphics illustrate the idealized geometry associated with the aspect ratio for the as-prepared In-doped ZnO samples .

Figure 5 shows transmission electron microscope (TEM) images of a) pure ZnO also showing the scale marker that is also valid for all images except for image i) . b-f) 2-10 at . % In-doped ZnO and 4 at . % g) Sn- and h) Li-doped ZnO. The crystal morphology of the In- doped ZnO, clearly changes from spheroidal (pure ZnO) , to cuboid and to increasingly rod-like as In concentration increases, consistent with XRD (Figure 3a,d) . Image (i) shows a high-resolution TEM of a single 10 at. % In-doped ZnO nanorod crystal .

Figure 6 shows the interplanar d-spacing of (100) , (002), and (101) peaks for ZnO doped with various concentrations of indium. The (101) spacing is normalized to the undoped value and serves here as an internal reference. Relative to the (101) plane, the (100) spacing remains constant while the (002) spacing shifts towards higher values as dopant concentration increases . The selective expansion of the (002) plane separation indicates dopants associate with the (002) plane consistent with the incorporation of the dopant atoms within the Zn wurtzite lattice planes.

Figure 7 shows HRTEM images of 10 at . % In- doped ZnO showing lattice planes for cross-sectional (a) and axial (b) directions of nanorod crystals. The measured spacings are consistent with the (002) (ca. 2.6A; a) and (001) (ca. 2.8A; b) planes respectively, confirming the nanorod length along the c-axis .

Figure 8 shows the dopant coordination, wherein a) shows a Raman spectroscopy analysis (632 nm excitation) of Sn-doped ZnO powders. The peaks at 570 cm^~l and 670 cm^~l are associated with different coordination states of the Sn dopant with the latter peak reflecting higher coordination, b) shows the ratio of the areas (circles) below the peak at 670 cm^~l to that below 570 cm^"1. This ratio is consistent with the XRD lattice aspect ratio (diamonds) (Figure 4b) .

Modes for Carrying out the Invention

The invention is now further described for doped ZnO nanorods .

Synthesis :

All samples were made by flame spray pyrolysis (FSP) - a single-step, continuous, and readily scaleable process .

A diagram of an experimental apparatus suitable for the FSP method for the production of metal oxide nanorods is shown in Figure 1. For details of the FSP process and associated experimental setup reference is made to the work of Madler and coworkers (Madler et al . , 2003), the disclosure of which is incorporated herein in its entirety.

For the following experiments, a dispersion gas flow rate of 5 L/min and a pressure drop of 1.5 bar was maintained across the nozzle during FSP operation. A sheath gas flow of 5 L/min of oxygen was issued concentrically around the nozzle to stabilize and contain the spray flame . The precursor liquid feed was supplied at 5 ml/min using a rate-controlled syringe pump (Inotech R232) and all gas flows (Pan Gas, >99.95%) were metered using mass flow controllers (Bronkhorst) . A water-cooled, stainless-steel filter housing supported a glassfiber sheet (Whatman GF/D; 25.7 cm diameter) for collection of the flame-produced powder with the aid of a vacuum pump (Busch) . The basis liquid precursor solution was composed of toluene (Fluka, 99.5%) and zinc naphthenate (STREM Chemicals, 65% in mineral spirits) . Dopant species were indium (indium acetyl acetonate; Aldrich, 99.99%) , tin (stannous 2-ethyl hexanoate; Aldrich 95%), and lithium (lithium tert-butoxide; Aldrich, 1. OM in tetrahydrofuran) . Li-doped ZnO samples were also prepared using lithium acetylacetonate (Aldrich, 97%) resulting in identical powders to the tert-butoxide precursor. Dopant concentrations ranged between 1 and 10 atom percent with respect to the Zn metal. The total metal concentration for each precursor solution was 0.5 mol/L. Thermal stability of the powders was investigated by annealing powders in a Carbolite CWF1300 temperature programmed oven. Samples were heated at 5 °C/minute to either 700 ⁰C or 900 °C, maintained at that temperature for 5 hours, followed by cooling at 10 °C/minute.

Characterization of the products prepared as described above:

The flame-generated powders were characterized using powder X-ray diffraction (XRD) with a Bruker AXS D8 Advance spectrometer at 2Θ (Cu-Ka) 20 to 70°, step size of 0.03°, and scan speed of 0.6 °/min (source 40 kV, 40 mA) . XRD patterns were analyzed using the Fundamental Parameter (FP) method to match the profile of individual peaks within each XRD pattern, allowing extraction of crystal size information (Cheary and Coelho, 1992) . BET adsorption isotherms and specific surface area analysis was performed using a MicroMeritics TriStar 3000 system after degassing in nitrogen for 1.5 hours at 150 °C. The specific surface area (SSA) was measured using 5-point nitrogen adsorption at 77 K. The BET equivalent diameter was evaluated from the measured SSA for each sample, assuming a spherical primary particle geometry and a composition-corrected density. Transmission electron microscopy analysis was carried out with a Phillips CM30ST microscope (LaB6 cathode, 300 kV) . A confocal Raman microscope (Labram, Jobin Yvon, ex DILOR Instruments SA) was used to acquire Raman spectra. An internal HeNe laser at 632.8 nm with approximately 2.5 mW power was used for ex situ measurements. Raman spectra were recorded in the spectral range from (200 to 1050 cm^{" ■}"^■) with a line resolution of 4 cm^'1. Raman band positions were calibrated against the spectrum of a neon lamp (Penray, Oriel) . With the confocal hole at 500 micrometers the spatial resolution was between 3-5 μm³.

As produced powders were investigated with the following

Results

Figure 2 shows X-ray diffraction (XRD) patterns for pure and indium doped ZnO (1 to 10 atom %) . The XRD-pattern of pure ZnO (bold line) shows the typical reflections of ZnO with the crystal plane assignments indicated in the upper portion of the frame. Patterns for indium doped ZnO were arranged upwards from the ZnO pattern with increasing dopant concentration. All patterns were normalized relative to the intensity of the

(101) peak at 36.3°. A main feature of the patterns is the diminishing intensity and broadening of the (002),

(102), (103) and (112) peaks as indium dopant concentration is increased. In contrast, the (100) ,

(101) , (110) , (200) , and (201) peaks remained relatively unaffected by the dopant. Comparison of the (100) and

(002) peaks in particular yielded insight into two crystal planes that are orthogonal to each other and so were expected to yield information on crystal geometry and texture .

The XRD patterns for each of the In-, Sn-, and Li-doped ZnO samples exhibited the same characteristic pattern of the pure ZnO with wurtzite crystal structure (Figure 2) . For the dopant levels considered there were no additional XRD peaks associated with the dopant oxides, indicating that no new crystal- phase formation occured. The structural agreement between the doped and pure ZnO patterns indicates that each of the dopant atoms are fully incorporated into the ZnO crystal lattice without significantly altering the packing structure of the parent lattice. The preservation of the wurtzite lattice structure could also be confirmed by electron diffraction and Raman spectroscopy measurements (not shown) .

Figure 3, top panel, i.e. a, b and c, shows the (100) , (002), and (101) peaks in detail (2Θ = 28 to 38°) for a) In-doped, b) Sn-doped, and c) Li-doped ZnO samples for 1 - 10 at% dopant concentration. For both In- and Sn-doped samples, the (002) peak exhibits a dramatic reduction in intensity and a peak broadening with increasing dopant concentration. In contrast, the Li- doped ZnO pattern showed to be relatively insensitive to dopant concentration. The decreasing intensity and increased broadening of the (002) peak found is consistent with the (002) crystal size becoming smaller, while the (101) crystal size appeared to remain relatively constant.

To quantitatively investigate the observed effect of dopant concentration on crystal plane dimensions, the XRD patterns were analyzed using the TOPAS software (Bruker, 2000) . The Fundamental Parameter (FP) method was used to fit the profile of individual peaks within each XRD spectrum, allowing extraction of crystal size information (Cheary and Coelho, 1992) . The FP method was used to determine the average crystal size associated with the (100) and (002) peaks in each pattern.

Figure 3, lower panel, i.e. d, e and f, shows the average crystal plane size associated with the (100) and (002) peaks in each pattern as a function of dopant concentration. For the In-doped ZnO, the (002) crystal size decreased by a factor of five from 27 nm to 5 nm while indium concentration was increased from 0 to 8 at .% (Figure 3d) . The (001) size increased slightly from 18 to 20 nm before gradually declining to 16 nm. The Sn-doped ZnO showed a similar, yet more steep trend than the In- doped ZnO with the (002) crystal size decreasing from 27 nm to 6 nm with the addition of only 4 at . % . The corresponding change in size^' for the (100) plane was 18 to 14 nm with 4 at . % tin dopant (similar effect for 8% In-doped ZnO) . Both In and Sn selectively reduced the size of the (002) plane while only slightly decreasing the size of the (100) plane. In contrast, the crystal sizes for the Li-doped ZnO remained unaffected even for dopant concentrations up to 10 at . % .

The specific surface area (SSA) increased steadily from 53 to 77 m²/g as indium concentration increased from 0 to 10 at . % and from 53 to 85 m²/g as tin concentration increased from 0 to 6 at . % followed by a decrease to 75 m²/g at doping up to 10 at.%. The Li-doped ZnO exhibited the same SSA as the pure ZnO (53 m²/g) for dopant concentrations up to 6 at.% and then decreased to 43 m²/g at doping up to 10 at.%. The BET equivalent diameter (dBET) for the In-, Sn-, and Li-doped ZnO samples are shown also as a function of dopant concentration in Figure 3d,e,f (squares) .

The thermal stability of the In-doped nanorods was investigated by annealing for 5 hours at either 900 or 700 ⁰C (Figure 3d) . At both temperatures, the XRD size of the (100) and (002) planes of the pure ZnO powders increased dramatically from 23 ± 5 nm (as prepared) to 65 ± 2 nm for 700 ⁰C and 133 ± 2 nm for 900 °C. For the 6 at . % In-doped ZnO samples the size of the (100) plane increased from the as-prepared value of 16.7 nm to 24.7 nm (700 ⁰C) and 33.0 nm (900 ⁰C) (Figure 3d) . In contrast, the size of the (002) plane remained relatively close to the as-prepared value of 6.0 nm, increasing to 6.9 nm (700 ⁰C) and 7.8 nm (900 ⁰C) (Figure 3e) . The 6 at . % Sn-doped ZnO samples showed a similar increase in the (100) plane size from 12.1 nm (as prepared) to 16.0 nm (700 ⁰C) and the (002) plane size changed from 4.4 nm (as-prepared) to 5.9 nm (700 ⁰C) . Furthermore, for annealing at 900 ⁰C the size of both the (100) and (002) planes increased dramatically to 98.1 nm and 35.3 nm respectively (not shown in Figure 3e) .

Zinc oxide has a hexagonal-close-packed wurtzite structure in which the crystal is composed of alternating planes of Zn atoms and 0 atoms (Figure 4a) . The (002) plane lies in parallel to the 0 and Zn planes while the (100) plane lies perpendicular to the (002) plane, intersecting alternating layers of Zn and O. The crystallite size of the (002) and (100) planes can essentially be considered as metrics of 'Diameter' (D) and 'Length' (L) respectively (Figure 4a) . The crystal aspect ratio (L/D) can therefore be estimated by the ratio of (100) and (002) crystal plane sizes, designated here as the XRD lattice aspect ratio. Figure 4b shows the variation of XRD lattice aspect ratio with dopant concentration for the In-, Sn-, and Li-doped ZnO as calculated from the crystal plane size measurements. For the In-doped ZnO the XRD lattice aspect ratio was found to increase steadily by a factor of 5 from 0.6 to 3.1 as dopant concentration was increased from 0 to 10 at.%. Similarly, the Sn-doped ZnO was also found to increase in XRD lattice aspect ratio as more dopant is added, however tin appeared to have a stronger influence than indium at lower dopant concentrations. The XRD lattice aspect ratio of the Li-doped ZnO remained essentially the same as the one of pure ZnO as the Li-dopant concentration was increased up to 10 at . % .

The annealed pure ZnO samples showed a dramatic jump in size of both the (100) and (002) planes upon annealing, giving an XRD lattice aspect ratio of unity for both temperatures. The In-doped ZnO samples showed a selective increase in size of the (100) plane yielding an increase in XRD lattice aspect ratio from 2.8 (as-prepared) to 3.6 (700 ⁰C) and 4.2 (900 ⁰C) (Figure 4b) . For Sn-doped ZnO the XRD lattice aspect ratio remained constant at 2.7 for (700 °C) and slightly increased to 2.8 for (900 ⁰C) . However this finding could be associated with a resultant particle size similar to the one of annealed pure ZnO, consistent with the assumption that tin dopant might be leaving the ZnO lattice at higher annealing temperatures (Figure 3e) . This is in contrast to Indium where the XRD lattice aspect ratio was dramatically enhanced for the two annealing temperatures considered, namely 700 ⁰C and 900 ⁰C.

Figure 5 shows a series of TEM images for (a) pure ZnO along with (b to f) showing 2 to 10 at . % In- doped ZnO. Pure ZnO was found to contain mainly spheroidal particles typically with diameters of 20 nm with occasional rod-like structures, consistent with Tani and coworkers (Tani et al . , 2002a) . These short rod-like features seen among the mainly spheroidal particles are assumed to arise from a slight oxygen deficiency for some particles as they form in the flame, in-line with size effects caused by oxygen stoichiometry variation (Tseng et al., 2003; Yan et al . , 2003) . The image of 2 at. % In- doped ZnO (b) shows particles with cuboid morphology, while for 4 at .% doping (c) the particles begin to exhibit an elongated morphology. Between 6, 8, and 10 at.% (d,e,f) the particles were found to exhibit increasing rod-like appearance. The 10 at.% In-doped ZnO sample (f) in particular was found to be dominated by nanorod features with diameters between 10 and 20 nm and lengths up to about 100 nm. The XRD measurements (Figure 4), showing an increased XRD lattice aspect ratio with increased dopant concentration, are consistent with these TEM images.

The TEM images in Figure 5c,g,h also demonstrate the influence of the 4 at.% dopant concentration for the In-, Sn- and Li-doped ZnO. Compared to the pure ZnO sample (a) , the In-doped ZnO (c) and Sn- doped ZnO (g) , both exhibit rod-like structure. It is important to note that these TEM micrographs are essentially a cross-sectional view of the sample, possessing random alignment, and that they show only a limited aspect of the true geometry. Rod-like features aligned parallel to the image plane, however, are clearly evident in these images. Li-doped ZnO (h) is dominated by spheroidal particles similar to pure ZnO (a) . Figure 4i shows a high-resolution TEM image of an individual nanorod with the single-crystalline morphology of each structure clearly evident.

Visual inspection of the nanorods in the TEM images of Figure 5 reveals features with aspect ratios at least between 5 and 10 while the aspect ratios calculated from the XRD are all less than 4. This discrepancy may be due to differences in the averaging of the characterization methods. The TEM images give a representation of the sample essentially in terms of number, while the XRD signal averages according to particle mass, and BET analysis averages according to surface area. The characterisation methods therefore sample different moments of the particle distribution and the averaging caused by this not surprisingly led to the variation between the methods. The key point, however, is that together the methods give a consistent picture of the trends of aspect ratio increase with indium or tin dopant level, and the more drastic aspect ratio change for tin than for indium at low dopant concentration. Figure 6 shows a plot of the displacement of the (100) and (002) peaks from the (101) peak as a function of indium dopant concentration. For dopant concentrations up to 6 at %, the (100) and (002) peaks remained at identical positions to the undoped sample. However, for dopant concentrations of 8 and 10 at %, the maxima of the (002) peak shifted towards higher interplanar d-spacing (smaller 2Θ) values while the (100) peak remained at the same d-spacing. This peak shift was noticeable only for higher dopant concentrations, the specificity of the shift to the (002) peak indicates that the dopant atoms are influencing the layers parallel to the planar layer of O atoms. This is consistent with the dopant atoms being incorporated within the planar Zn layer.

Although the dopants are incorporated within the wurtzite lattice, it could be shown that the lattice geometry is altered for higher indium and tin concentrations with respect to the 2Θ position of the (002) peak (see Figures 3a, 3b) . Comparing the relative displacement of the maxima of the (100) and (002) peaks relative to the position of the (101) peak, treated here as an internal reference, it was possible to gain an insight of the distortion in crystal geometry. For dopant concentrations up to 6 at.%, the (100) and (002) peaks remained at identical positions with those of pure ZnO. However, for dopant concentrations of 8 and 10 at.%, the maximum of the (002) peak shifted towards higher interplanar d-spacing (smaller 2Θ) values while the (100) peak remains at the same d-spacing. This peak shift was noticeable only for higher dopant concentrations. The specificity of the shift to the (002) peak was found consistent with the assumption that the dopant atoms are incorporated within the cross-sectional plane. Further evidence in this respect provides Figure 7. Figure 7 shows high resolution TEM images of 10 at.% In-doped ZnO nanorod crystals. The approximate spacing between the lattice planes in the cross-sectional and axial directions that was found is consistent with the (002) and (100) planes, respectively.

As already mentioned above, the incorporated dopants may influence the surrounding lattice either by their atom size, their valency or their degree of coordination. The atomic radii of In (1.55 A) , Sn (1.45 A) and Li (1.45 A) are very similar to the atomic radius of Zn (1.35 A) (Slater, 1964) so the size of the dopant atoms is unlikely to induce the observed rod-like geometry. Similar dopant sizes can create a substitutional defect as with Sn and Al doping on Tiθ2 (Akthar et al . , 1994) . The peak shift described in the above paragraph was most pronounced for indium, the dopant species with the highest atomic radius relative to zinc and even then the shift occured only at high dopant concentrations where the size of the dopant is likely to increase the d-spacing for the (002) planes.

While size was found not to be a large distinguishing factor between the dopants, their valencies are quite different than zinc (+2) : with Li of lower valency (+1), and In and Sn each of higher valency (+3 and +4, respectively) . Multivalent dopants have also been reported to influence the morphology of other crystals such as calcium oxide monohydrate (Touryan et al . , 2004) . Here, the higher valency of In and Sn relative to Zn is assumed to lead to greater coordination with surrounding O than for Zn atoms within the same crystal plane. The In and Sn atoms therefore are assumed to have a disrupting influence, hindering crystal growth within the Zn layer (Kong et al . , 2004) . This hindering effect was observed in decreasing the (002) crystal plane size as the In and Sn dopant concentrations were increased (Figure 3) . The more pronounced nanorod forming influence of Sn relative to In at low dopant concentrations may be attributed also to higher Sn valency and so larger disruptive influence. Conversely, the lower valency of Li is expected to mean that this atom has only minimal disruptive influence within the Zn plane giving the observed invariance of the (002) crystallite plane size (Figure 3f) .

Figure 8a shows Raman spectra for 0 to 10 at.% Sn-doped ZnO. Each spectrum was normalized to the amplitude of the E2 mode (437 cm^"1) of the wurtzite. All Sn-doped powders were found to exhibit two additional peaks at 570 cm^"1 and 670 cm^"1. These peaks are indicative of vibrations associated with the tin dopant atoms within the crystal while the peak at 670 cm"¹ was associated with a more highly coordinated structure than that of the 570 cm^"1 peak (Porotnikov et al . 1983) . Figure 8b shows the ratio (diamonds) of areas below the 670 cm^"1 peak to that below the 570 cm^"1 peak as a function of dopant concentration. The increasing proportion of highly coordinated dopant, as indicated by that ratio, correlated well with the XRD lattice aspect ratio (circles) trend (Figure 4b) further supporting the effect of dopants on the crystal geometry.

While there are shown and described presently preferred embodiments of the invention, it is to be dis¬ tinctly understood that the invention is not limited thereto but may be otherwise variously embodied and prac¬ ticed within the scope of the following claims .

References

Akhtar, M.K., Pratsinis, S.E. and

Mastrangelo, V.R., 1992. Dopants in vapor-phase synthesis of titania powders . Journal of the American Ceramic Society, 75 (12) : 3408-3416.

Akhtar, M.K., Pratsinis, S.E. and

Mastrangelo, V.R., 1994. Vapor phase synthesis of Al- doped titania powders . Journal of Materials Research, 9 (5) : 1241-1249.

Bruker, 2000. TOPAS: General profile and structure analysis software for powder diffraction data. Bruker AXS GmbH, Karslruhe, Germany, 173-190 pp.

Cheary, R.W. and Coelho, A., 1992. A fundamental parameters approach to x-ray line-profile fitting. Journal of Applied Crystallography, 25: 109-121.

Cheng, B. and Samulski, E.T., 2004. Hydrothermal synthesis of one-dimensional ZnO nanostructures with different aspect ratios. Chemical Communications, 8: 986-987.

Cheng, B., Xiao, Y., Wu, G. and Zhang, L., 2004. The vibrational propoerties of one-dimensional ZnO:Ce nanostructures. Applied Physics Letters, 84 (3) : 416-41.

Gao, P.X., Ding, Y. and Wang, Z.L., 2003. Crystallographic orientation-aligned ZnO nanorods grown by a tin catalyst. Nano Letters, 3 (9) : 1315-1320.

Guo, L., Ji, Y.L., Xu, H., Simon, P. and Wu, Z., 2002. Regularly shaped, single-crystalline ZnO nanorods with wurtzite structure. Journal of the American Chemical Society, 124: 14864-14865.

Height, M.J., Howard, J.B. and Tester, J.W., 2003. Flame synthesis of carbon nanotubes. Materials Research Society Symposium Proceedings, 772 (Nanotube- based devices) : 55-61.

Height, M.J., Howard, J.B., Tester, J.W. and Vander Sande, J.B., 2004. Flame synthesis of single- walled carbon nanotubes. Carbon, 42 (11) : 2295-2307. Kawakami, M., Hartanto, A.B., Nakata, Y. and Okada, T., 2003. Synthesis of ZnO nanorods by nanoparticle assisted pulsed-laser deposition. Japanese Journal of Applied Physics, 42: L33-L35.

Kong, X.Y., Ding, Y., Yang, R. and Wang,

Z.L., 2004. Single-crystal nanorings formed by epitaxial self-coiling of polar nanobelts. Science, 303: 1348-1351.

Leung, Y.H., Djurisic, A.B., Gao, J., Xie, M.H. and Chan, W.K., 2004. Changing shape of ZnO nanostructures by controlling Zn vapor release: from tetrapod to bone-like nanorods. Chemical Physics Letters, 385: 155-159.

Lin, M. and Chang, F.-K., 2002. The manufacture of composite ceramics with a built-in network of piezoceramics . Composites Science and Technology, 62: 919-939.

Liu, B. and Zeng, H.C, 2003. Hydrothermal synthesis of ZnO nanorods in the diameter range of 50 niti. Journal of the American Chemical Society, 125: 4430-4431.

Liu, X., Wu, X., Cao, H. and Chang, R. P.H., 2004. Growth mechanism and properties of ZnO nanorods synthesized by plasma-enhanced chemical vapor deposition. Journal of Applied Physics, 95 (6) : 3141-3147.

Madler, L., Kammler, H.K., Mueller, R. and Pratsinis, S.E., 2002. Controlled synthesis of nanostructured particles by flame spray pyrolysis. Aerosol Science, 33: 369-389.

Madler, L., Stark, W.J. and Pratsinis, S.E., 2003. Simultaneous deposition of Au nanoparticles during flame synthesis of TiO₂ and SiO₂. Journal of Materials Research, 18 (1) : 115-120.

Manhart, J., Kunzelmann, K.-H., Chen, H.Y. and Hickel, R., 2000. Mechanical properties of new composite restorative materials. Journal of Biomedical Materials Research, 53 (4) : 353-361.

Muller, R., Madler, L. and Pratsinis, S.E., 2003. Nanoparticle synthesis at high production rates by flame spray pyrolysis. Chemical Engineering Science, 58 (10) : 1969-1976.

Patzke, G.R., Krumeich, F. and Nesper, R.,

2002. Oxidic nanotubes and nanorods - anisotropic modules for a future nanotechnology. Angewandte Chemie International Edition, 41: 2446-2461.

Porotnikov, N.V., Savenko, V.G., Sidorova, O.V., 1983. Vibrational-spectra of the Zn₂SnO₄ and Mg₂SnO₄ spinels. Russian Journal of Inorganic Chemistry, 28: 932- 934.

Pratsinis, S.E., 1998. Flame aerosol synthesis of ceramic powders. Progress in Energy and Combustion Science, 24: 197-219.

Slater, J.C, 1964. Atomic radii in crystals. Journal of Chemical Physics, 41 (10) : 3199-3204.

Sun, X., Zhang, H., Xu, J., Zhao, Q., Wang, R. and Yu, D., 2004. Shape controllable synthesis of ZnO nanorod arrays via vapor phase growth. Solid State Communications, 129: 803-807.

Tani, T., Isobe, S., Seo, W.-S. and Koumoto, K., 2001. Thermoelectric properties of highly textured (ZnO) ₅In₂O₃ ceramics. Journal of Materials Chemistry, 11 (9) : 2324 - 2328.

Tani, T., Madler, L. and Pratsinis, S.E., 2002a. Homogeneous ZnO nanoparticles by flame spray pyrolysis. Journal of Nanoparticle Research, 4: 337-343.

Tani, T., Madler, L. and Pratsinis, S.E., 2002b. Synthesis of ZnO/Siθ₂ composite nanoparticles by flame spray pyrolysis. Journal of Materials Science, 37: 4627-4632.

Tourian, L.A., Lochhead, M.J., Marquardt, B.J., Vogel, V., 2004. Sequential switch of biomineral crystal morphology using trivalent ions . Nature Materials, 3: 239-243.

Tseng, Y.-K., Hsu, H. -C, Hsieh, W.-F., Liu, K.-S. and Chen, I.-C, 2003. Two-step oxygen injection process for growing ZnO nanorods . Journal of Materials Research, 18 (12) : 2837-2844.

Vayssieres, L., 2003. Growth of arrayed nanorods and nanowires of ZnO from aqueous solutions. Advanced Materials, 15 (5) : 464-466.

Wang, Z.L., 2003. Nanowires and nanobelts : ^' Materials, properties and devices. Metal and semiconductor nanowires (VoIl), 1. Kluwer, Norwell MA.

Wen, J.G., Lao, J.Y., Wang, D. Z., Kyaw, T.M., Foo, Y.L. and Ren, Z. F., 2003. Self-assembly of semiconducting oxide nanowires, nanorods, and nanoribbons . Chemical Physics Letters, 372: 717-722.

Xia, Y., Yang, P., Sun, Y., Wu, Y., Mayers, B., Gates, B., Yin, Y., Kim, F. and Yan, H., 2003. One- dimensional nanostructures : synthesis, characterization, and applications. Advanced Materials, 15 (5) : 353-389.

Yan, H., He, R., Pham, J. and Yang, P., 2003. Morphogenesis of one-dimensional ZnO nano- and microcrystals . Advanced Materials, 15 (5) : 402-405.

Zeng, D.W., Xie, CS. , Zhu, B.L., Jiang, R., Chen, X., Song, W.L., Wang, J.B. and Shi, J., 2004. Controlled growth of ZnO nanomaterials via doping Sb. Journal of Crystal Growth, 266: 511-518.

Claims

Claims

1. A flame spray pyrolysis (FSP) method for the production of metal-oxide nanorods wherein at least one dopant metal precursor is added to a precursor solution of the predominant metal in amounts of 1 to 15 atom-% referred to the amount of the predominant metal, said dopant having a higher valency than the predominant metal, and wherein the crystal lattice of the pure metal oxide formed by FSP is in a hexagonal closed crystal structure .

2. The FSP method of claim 1 wherein the coordination number of the dopant atoms within the product crystal lattice is higher than the coordination number of the predominant metal atoms .

3. The FSP method of claim 1 or 2 wherein the dopant atoms and/or ions and the predominant metal atoms/ions have sizes varying by less than 20 % referred to the atomic radius of the predominant metal as 100 %.

4. The FSP method of anyone of the preceding claims, wherein the amount of the dopant is from 1 to 10 %, preferably from 4 to 10 atom-%.

5. The FSP method of anyone of the preceding claims, wherein the predominant metal is zinc (Zn) .

6. The FSP method of anyone of the preceding claims, wherein the dopant is indium (In) .

7. The FSP method of anyone of claims 1 to 5, wherein the dopant metal is tin (Sn) .

8. The FSP method of anyone of the preceding claims, wherein the dopant is a combination of two or more metals, in particular a combination of In and Sn.

9. A metal oxide predominantly in the form of nanorods and producible by the method of anyone of the preceding claims.

10. The metal oxide of claim 9 that is Sn and/or In doped ZnO.