This is due to theincreasing strength of the gold−halide interaction relativeto the silver−gold and silver−halide interactions. At lowhalide concentrations, when the halide is bromide oriodide, the decreasing stability of the AgUPD layer results ingreater silver coverage via underpotential deposition andthus more open (higher-index) surfaces form at lowersilver ion concentrations.(5) In the presence of silver ions as a shape-directing additive, theaddition of a large amount of a larger halide (bromide andiodide) greatly decreases the stability of the AgUPD layer andblocks silver deposition, limiting the number of particleshapes that can form. This conclusion is based on experi-ments which show that at high halide concentrations,where the halide is bromide or iodide, this destabilizingeffect results in a lower surface silver coverage, due to thestrong binding affinity of bromide and iodide for the goldsurface, blocking silver deposition. Consequently, forbromide, high-index {730} surfaces can still be formeddue to the mobility of the AgUPD layer, but for iodide, thisenhanced mobility results in surfaces that are not well-defined.(6) The enhanced stability of the AgUPD layer in the presence ofchloride causes growing gold nanoparticles to becomekinetically “trapped” or “locked” into a particular facetstructure early in their growth, enabling the formation of awide variety of shapes as well as concave particles. This issupported by the observation that the shape of theconcave cubes is fixed very early in their growth. As aresult of this kinetic trapping, a greater variety of shapescan form in the case of chloride than in the case ofbromide, where the AgUPD layer is more mobile and canreach more of an energy minimum.
These design considerations are self-consistent and accu-rately describe all of the experimental observations discussed inthis article. In addition to explaining the formation of the 10classes of nanoparticles of different shape, including concavecubes and tetrahexahedra, this study also has implications inexplaining previous results beyond high-index nanostructures.In our previously published syntheses for gold nanoprisms, wehave reported that trace amounts of iodide are vital for theformation of nanoprisms.21Our results here suggest that this islikely due to the decrease in reaction rate caused by iodide inthe absence of silver as a result of a lowered reduction potentialand solubility for [AuI2]− as well as surface binding effects ofiodide. In addition, our study lends some insight to themechanism of formation of gold nanorods in reaction solutionscontaining silver ion in CTA-Br.16In this case, it is observedthat the rods increase in length with the addition of largeramounts of silver ion. Others have suggested that these rods arein fact high-index faceted structures, analogous to tetrahexahe-dra with high aspect ratios.79,80We have shown that the varietyof high-index gold nanostructures stabilized by silver in CTA-Bris limited, and thus, in the case of rods, rather than stabilizinga more open surface facet, the AgUPD layer, which is highlymobile in 100 mM CTA-Br, expands to cover a larger area withthe same facet structure. The observed result of this process isthat an increase in silver ion concentration leads to an increasein rod aspect ratio, as high aspect ratio structures have a largersurface area relative to their volume. Further, such conclusionsare in agreement with work reported by Guyot-Sionnest andco-workers regarding the control of gold nanorod growth viasilver underpotential deposition as well as their synthesis ofpenta-twinned bipyramids.23The silver serves to passivate thesurface in both cases, and it is shown that the controlling factorin the growth of the bipyramids is the use of penta-twinnedseeds, which forces the growth of the penta-twinned bipyramidsrather than single crystalline rods. Despite differences incrystallinity, these two structures both have high-index facets,likely with comparable surface silver coverages as a result of thedestabilizing effects of bromide of AgUPD.23,79,80We have shown in this article that the two major factorscontrolling particle growth are reaction kinetics and surfacepassivation effects (particularly by silver), and these generaldesign considerations likely can also be applied to synthesesother than those which use a CTA-X surfactant. For example,Xu and co-workers have reported the synthesis of rhombicdodecahedra, octahedra, and cubes from single-crystalline seedsin cetylpyridinium chloride (CPC).36The reaction product ischanged from {110}-faceted rhombic dodecahedra to {111}-faceted octahedra by significantly decreasing the ascorbic acidconcentration, leading to a lower-energy facet at a slower re-action rate, as predicted by our design considerations. Cubesbound by {100} facets are generated by adding a large amountof bromide to the slower set of reaction conditions. This initiallyseems contradictory to our work, since in our studies the addi-tion of such a large concentration of bromide to the reactionwould lead to the formation of bitetrahedra, which favor growthalong a twin plane due to slow, kinetically controlled growthconditions. However, it is important to note that the seeds usedto initiate nanoparticle growth in the work by Xu and co-workers are single crystalline and also somewhat cubic or rodlikein shape.36Since the single-crystalline seeds do not contain twinplanes where lateral growth can take place, it is possible that aslow layer-by-layer overgrowth of the seeds occurs instead andthe final cube structure reflects the original shape of the seedparticles. Recently, there has been great interest in goldnanostars, synthesized in CTAB,15,81or in a system containinggold, poly(vinylpyrrolidone) (PVP), and N,N-dimethylforma-mide (DMF).82In both cases, it has been proposed that thegrowth of these stars occurs due to rapid kinetic growth.15,81,82Such extremely rapid growth fits into our scheme at reactionrates faster than those which produce trisoctahedra, wheregrowth becomes less controlled and leads to polycrystalline andstellated structures. A similar situation is also observed in thecase of our reactions involving both silver ion and iodide. Whilesome of the principles we have outlined here, such as kineticcontrol, may also be applicable to the polyol process, thereaction conditions employed in that method are vastly differentfrom those of the seed-mediated synthesis and therefore it isbeyond the scope of this work to speculate on such connections.However, we note that while the polyol process is the preferredmethod for the synthesis of shape-controlled nanoparticlescomposed of silver and other metals, the seed-mediatedsynthesis is one of the most popular, if not the most popular,methods used by the nanoscience community to generate goldnanoparticles.9Thus, the design considerations outlined hereshould be of great utility in advancing the shape-controlledsynthesis of gold nanoparticles.■ 金纳米粒子的形状演化英文文献和中文翻译(11):http://www.751com.cn/fanyi/lunwen_33952.html