KW. Kolasinski Current Opinion in Solid State and Materials Science 10(2006)182-191 b root growth float growth d multiprong single prong Fig. 2. The processes that occur during catalytic growth.() In root growth, the particle stays at the bottom of the nanowire (b) In float growth, the rticle remains at the top of the nanowire. (c) In multiple prong growth, more than one nanowire grows from one particle and the nanowires must cessarily have a smaller radius than the particle. (d) In single-prong growth, one nanowire corresponds to one particle. One of the surest signs of this lode is that the particle and na have very similar radii. be(vi)diffusion of material through the catalytic particle in example, there have been several reports on the growth addition to(ii) diffusion along its surface. Substrate atoms of Ill-V nanowires -such as GaAs, GaP, InAs and might also be mobile. They might (v) enter the particle InP-that have invoked either a solid [41, 42, 44, 69, 70 directly or else (iv) surface diffusion along the substrate or liquid [43, 47, 71] catalyst particle. Similarly, during can deliver them to the surface of the particle. Not shown the growth of carbon nanotubes or fibres both liquid [21 in the diagram is that atoms from the catalytic particle and solid [23] catalyst particles have been reported. What might also be mobile and diffuse along the sidewalls and is clear from these reports is that there are circumstances the substrate under which the catalyst can be liquid and others in which a Four major distinctions in the growth process are illus- it can be solid and the nanowires that result do not appear ted in Fig. 2: root vs. float growth and multiprong vs. to be materially affected. In other words, while it is impor single-prong growth. The particle may either end up at tant for characterizing the growth mechanism, as far as the the bottom(root growth) or top(float growth)of the nano- dynamics of nanowire and nanotube formation are con- wire. In multiprong growth, Fig 2c, more than one nano- cerned, it does not appear to matter whether the catalytic wire grows from a single particle. In this case, the radius of particle is liquid or solid. Therefore, any mechanism that the nanowire rw must be less than the radius of the catalytic relies upon a particular phase for the catalyst cannot be gen- particle rp. In single-prong growth there is a one-to-one erally true. correspondence between particles and nanowires. A natu Likewise, the phase from which the growth material is ral means to exercise control over the nanowire diameter taken is of little consequence. The growth material may in single-prong growth would be if the nanowire radius come from a gas that is unreactive on the substrate and determines this value and w A Tp. Here is should be men- only reactive on the catalyst surface, such as in the case tioned that in single-prong growth, IwArp is usually of SiNW growth from silane. It may come from an atomic observed but that the catalyst particle sometimes is signif- vapor that has unit sticking probability on both the sub- icantly larger and occasionally is somewhat smaller than strate and the catalyst, as in mBe growth of SiNW or the nanowire radius. In multiprong growth rw is not deter- III-V compounds. It may come from a plasma, a solution mined directly by rp but must be related to other structural or even supercritical fluids. Hence, any model that requires factors such as the curvature of the growth interface and a particular phase for the growth material cannot be gener- lattice matching between the catalytic particle and the ally valid. What is required is that the growth material is nanowire mobile and can readily reach the growth interface with a Quite a bit of discussion has revolved around whether low probability of nucleating a crystallite(alternate growth he catalyst particle is liquid or solid. In other front)anywhere other than at the nanowire/catalyst whether the mechanism of growth is VlS or VSS. As an interfacebe (vi) diffusion of material through the catalytic particle in addition to (ii) diffusion along its surface. Substrate atoms might also be mobile. They might (v) enter the particle directly or else (iv) surface diffusion along the substrate can deliver them to the surface of the particle. Not shown in the diagram is that atoms from the catalytic particle might also be mobile and diffuse along the sidewalls and the substrate. Four major distinctions in the growth process are illus￾trated in Fig. 2: root vs. float growth and multiprong vs. single-prong growth. The particle may either end up at the bottom (root growth) or top (float growth) of the nano￾wire. In multiprong growth, Fig. 2c, more than one nano￾wire grows from a single particle. In this case, the radius of the nanowire rw must be less than the radius of the catalytic particle rp. In single-prong growth there is a one-to-one correspondence between particles and nanowires. A natu￾ral means to exercise control over the nanowire diameter in single-prong growth would be if the nanowire radius determines this value and rw  rp. Here is should be men￾tioned that in single-prong growth, rw  rp is usually observed but that the catalyst particle sometimes is signif￾icantly larger and occasionally is somewhat smaller than the nanowire radius. In multiprong growth rw is not deter￾mined directly by rp but must be related to other structural factors such as the curvature of the growth interface and lattice matching between the catalytic particle and the nanowire. Quite a bit of discussion has revolved around whether the catalyst particle is liquid or solid. In other words, whether the mechanism of growth is VLS or VSS. As an example, there have been several reports on the growth of III–V nanowires – such as GaAs, GaP, InAs and InP – that have invoked either a solid [*41,42,44,69,70] or liquid [*43,*47,71] catalyst particle. Similarly, during the growth of carbon nanotubes or fibres both liquid [21] and solid [23] catalyst particles have been reported. What is clear from these reports is that there are circumstances under which the catalyst can be liquid and others in which it can be solid and the nanowires that result do not appear to be materially affected. In other words, while it is impor￾tant for characterizing the growth mechanism, as far as the dynamics of nanowire and nanotube formation are con￾cerned, it does not appear to matter whether the catalytic particle is liquid or solid. Therefore, any mechanism that relies upon a particular phase for the catalyst cannot be gen￾erally true. Likewise, the phase from which the growth material is taken is of little consequence. The growth material may come from a gas that is unreactive on the substrate and only reactive on the catalyst surface, such as in the case of SiNW growth from silane. It may come from an atomic vapor that has unit sticking probability on both the sub￾strate and the catalyst, as in MBE growth of SiNW or III–V compounds. It may come from a plasma, a solution or even supercritical fluids. Hence, any model that requires a particular phase for the growth material cannot be gener￾ally valid. What is required is that the growth material is mobile and can readily reach the growth interface with a low probability of nucleating a crystallite (alternate growth front) anywhere other than at the nanowire/catalyst interface. Fig. 2. The processes that occur during catalytic growth. (a) In root growth, the particle stays at the bottom of the nanowire. (b) In float growth, the particle remains at the top of the nanowire. (c) In multiple prong growth, more than one nanowire grows from one particle and the nanowires must necessarily have a smaller radius than the particle. (d) In single-prong growth, one nanowire corresponds to one particle. One of the surest signs of this mode is that the particle and nanowire have very similar radii. K.W. Kolasinski / Current Opinion in Solid State and Materials Science 10 (2006) 182–191 187