We have discovered typographical errors in the discussion of TEM results for the Ni nanoparticles synthesized with and without the capillary injector shown in figure 5. The text on p 6 should read (with corrections in italics): We collected the Ni nanoparticles under a selection of growth conditions, particularly with and without the capillary injector, to verify the aerosol mobility results and assess the composition and crystallinity of the material. Figure 5 shows representative TEM images of Ni nanoparticles collected by filtration, dispersed in solution, and drop cast onto substrates. An image of Ni nanoparticles synthesized with a 180 μm capillary injector at a nickelocene vapor concentration of 0.43 ppm and total volumetric flow rate of 100 sccm is shown in figure 5(a). The nanoparticles are relatively uniform in size, with an approximate diameter of 5 nm, in agreement with aerosol mobility measurements (see figure 3). The nanoparticles are crystalline, with a measured average lattice spacing of 0.21 nm (see figure S8, supporting information) (stacks.iop.org/JPhysD/48/314003/ mmedia), corresponding to the Ni(1 1 1) crystal plane (figure 5(b)). From TEM, we could not assess the presence of carbon, which could have been incorporated in the nanoparticle product from the organometallic precursor, since the TEM substrates themselves are made up of amorphous carbon. Previously, we have shown that for dc microplasmas, carbon contamination can be characterized by x-ray photoelectron spectroscopy (XPS) and correlates to the formation of the C2 Swan band detected by optical emission spectroscopy (OES) . We did not observe any C2 Swan band in the OES spectra collected from the DBD microplasma, suggesting that solid carbonaceous species were not present in our grown material (see supporting information, figure S9) (stacks.iop.org/JPhysD/48/314003/mmedia). An image of Ni nanoparticles synthesized with a 510 μm capillary at the same nickelocene vapor concentration and total volumetric flow rate of 100 sccm is shown in figure 5(c). The nanoparticles are found to be larger in size, with an approximate diameter of 8 nm, and slightly agglomerated, which is mostly consistent with aerosol mobility measurements (see figure 4(a)). An image of Ni nanoparticles synthesized without a capillary at the same nickelocene vapor concentration and total volumetric flow rate of 800 sccm is shown in figure 5(d). In this case, the nanoparticles are found to be approximately 5-10 nm in diameter, and there is significant agglomeration. The particle size is significantly smaller than that measured by aerosol mobility measurements, which indicated ∼50 nm; this is most probably because the particles were measured in the gas phase as aggregates and then partially broken up by sonication during preparation of the TEM grid samples. Some aggregates still remain, and overall, this confirms that without a capillary, there is more agglomeration, particularly at the higher gas flow rates. We note that the crystallinity of the synthesized particles does not appear to be affected by the method of precursor injection, as all samples exhibit lattice fringes, although particle imaging is more difficult for samples prepared with a 510 μm capillary or without a capillary because of the agglomeration (see figure S10, supporting information) (stacks.iop.org/JPhysD/48/314003/mmedia). The formation of crystalline particles in a DBD, which is typically characterized by a relatively low gas temperature, may be related to particle heating via interactions with plasma species, as has been previously shown for silicon nanocrystals by modeling of charging and surface reactions [26, 27].
Bibliographical notePublisher Copyright:
© 2018 IOP Publishing Ltd.