Preparation of Tetragonal Cu Nanopillars

A summary of growth conditions of Cu nanopillars is listed in Table 4.1. The optimum growth condition (Pillar-3) is described below as a typical example.

Table 4.1 A summary of growth conditions of Cu nanopillars.

Sample CuCl2 (mM)

A glass substrate (5 × 10 × 1 mm³) was ultrasonically cleaned in alcohol and acetone for 10 min sequentially. The substrate was masked with Scotch tape to leave an exposed rectangular surface (1.5 × 4 mm²). Then, a layer of Au with a thickness of 5 nm was deposited onto the substrate by DC sputtering (2.2 kV, 15 mA, 120 s). Finally, the tape mask was removed to offer the Au electrode. A piece of Al metal sheet (1 × 2 × 0.5 mm³) was ultrasonically cleaned in alcohol (5 min), H3PO4(aq) (Riedal-de Haen, 5%, 2 min), and finally, rinsed by deionized water. The Al slice was then attached to one side of the Au electrode by silver paste (Toyobo) and dried on a hot plate (353 K, 1 h). The whole glass substrate was immersed into a limpid aqueous solution (4 mL) containing CuCl2 (5 mM, Strem) and dodecyltrimethylammonium chloride (DTAC)(0.15 mM, Fluka) In a glass vial at 290 K without stirring. The golden electrode turned black rapidly from its edge to the center within 2 min. As the reaction proceeded, it turned into dark red gradually. After a designated period of time (7 h), the substrate was removed and rinsed by deionized water and dried in a

desiccator.

4.2.2 Characterization

The deposited product layer was characterized by SEM (JEOL JSM-7401 at 15 kV), EDS (Oxford Link Pentafet), FETEM (JEOL JEM-2010F at 200 kV and JEOL JEM-4000EX), and XRD (Bruker AXS D8 Advance). Current-Voltage properties from field emission measurements were carried out using a needle-shaped anode with an effective tip to sample distance of 65 µm in a vacuum chamber at 4 × 10^-6 torr at room temperature. A positive voltage swept up to 1 kV with a step of 50 V was applied to the anode using a Keithly 2410 power supply.

4.3 Results

4.3.1 SEM and EDS characterization

Figure 4.1a shows a field-emission scanning electron microscopic (SEM) image of Pillar-1. The product grew into a pine tree-like morphology in 0.075 mM DTAC. A detailed image shows that all stems and branches of the tree-like configuration displayed an analogous tetragonal structure, as shown in Figure 4.1b. Figure 4.1c shows a SEM image of Pillar-2. An enlarged image shows that tree-like products were not observed. Many large clusters were produced in 0.03 mM DTAC, as shown in Figure 4.1d.

Figure 4.2 shows that SEM images of Pillar-3 produced in 0.15 mM DTAC solution.

Straight 1D NRs pointing randomly grow densely on the substrate, as shown in Figure 4.2a.

Figure 4.1 Low and high magnification SEM images of Cu nanostructures on Au/glass substrates at different DTAC concentrations at 7 h. (a), (b) 0.075 mM (Pillar-1) and (c), (d) 0.3 mM (Pillar-2).

Figure 4.2 SEM images of Cu nanopillars (Pillar-3) on Au/glass. (a) Top view (inset: EDS), (b) side view, and (c) side view and (d) top view of a pagoda-shaped tip.

An energy dispersive spectrum (EDS, inset Figure 4.2a) confirms that the NRs are composed of Cu mainly. The Pt signal is due to a sputtered Pt thin layer for clear SEM observation. The C signal may originate from the surfactant while the tiny O signal is the consequence of surface oxidation of Cu during the sample preparation under atmosphere.

Figure 4.2b shows a side view of the product on Au/glass. This displays an array of NRs aiming upward. The NRs are 1 – 6 µm long and 150 ± 25 nm wide. Some of the longest NRs show length-to-width aspect ratios close to 50. Between the array and the substrate, there is a layer (0.6 µm in thickness) of nanoparticles (NPs, 30 – 200 nm in size), presumably formed at the early growth stage. Figures 1c and 1d show enlarged views of a typical NR tip found in the array, revealing its four-side pagoda-like morphology. The pagoda has a height of 400 nm and a bottom perimeter of 600 nm. The images also suggest that the NR’s main body has an apparent tetragonal geometry. This is confirmed by the observation of an idealized tetragonal-shaped NR with a flat top square face in the SEM images. Thus, we conclude that majority of the NRs has a tetragonal pillar-like main body with a pagoda-shaped tip. Consequently, we name this unique type of new structure

“pagoda-topped tetragonal Cu nanopillar”.

Figure 4.3 shows growth evolution of nanopillars at different growth time. Pillar-4 was prepared for 1 h. A layer of cube-like Cu NPs grew on the substrate, as shown in Figure 4.3a.

Figure 4.3b shows that nananopillars with a length of up to 1 µm grew on the NP layer in Pillar-5 after 3 h. Based on experimental observation, the length of nanopillars increase gradually as growth time within 7 h. When growth time is elongated to 12 h, Pillar-6 were obtained. Figure 4.3c shows that pine tree-like dendrites were produced. This was also observed for the product grown in absence or deficiency of DTAC (Pillar-1).

4.3.2 XRD Analysis

XRD pattern of nanopillars grown on Au/glass is shown in Figure 4.4. The peaks at 2θ =

Figure 4.3 Side-view SEM images of Cu nanostructures on Au/glass substrates at different growing time. (a) 1 h (Pillar-4), (b) 3 h (Pillar-5), and (c) 12 h (Pillar-6). The insets show top view images enlarged.

43.0^o, 50.1^o, and 74.0^o are assigned to Cu (111), (200), and (220) reflections, respectively (JCPDF 89-2838). Lattice constant a is estimated to be 0.362 nm, is consistent with the reported value of Cu (JCPDF 89-2838). Reflections at 2θ = 38.0^o, and 43.8^o are from residual Ag paste after Al was removed (JCPDF 04-0783). The remaining peak at 54.0 are from a XRD holder. XRD patterns confirmed that the nanopillars are composed of a face-centered cubic (fcc) structure.

Figure 4.4 XRD of Cu nanopillars on Au/glass. Reflections of Ag were from residual Ag paste after Al was removed. (★: peaks from holder.).

4.3.3 TEM Characterization

Figure 4.5 shows a low magnification transmission electron microscopic (TEM) image of a representative nanopillar with an overall length of 3.6 µm. The regions selected from the tip, the body, and the root, as shown in Figure 4.5a, have side widths 100, 130, and130 nm, respectively. Their corresponding selected area electron diffraction (SAED) patterns are displayed in Figures 4.5b - d, respectively. All three images present identical square-shaped spot patterns, revealing the single crystalline nature of the nanopillar. The d spacing estimated from the spots closest to the beam center is 0.18 nm. This is consistent with the d spacing of Cu (200) planes (JCPDF 89-2838). Thus, the lattice parameter a is determined to be 0.36 nm, consistent with the reported value of fcc Cu. From the patterns, the crystallographic zone axis can be was corresponded to [001]. In addition to the spots of Cu, there are extra dim spots, as shown by the one pointed by an arrow in Figure 4.5b, indicating the presence of a tiny quantity of Cu2O. From this set of spots, a d spacing of 0.25 nm is estimated and assigned to be the d spacing of Cu2O (111) planes, 0.246 nm (JCPDF 78-2076). Figure 4.5e shows an enlarged image of the lower half of tip shown in

Figure 4.5 TEM studies of a Cu nanopillar. (a) Low magnification image; (b)-(d) SAED patterns of the rectangular marks in (a) indicating tip, middle, and bottom of the nanopillar (from left to right), respectively; (e) enlarged image of the lower half of the tip in (a); (f) HR image of the tip in (a).

Figure 4.5a. The complicated patterns and Moiré fringes are the result of uneven side walls of the pagoda shaped structure, as indicated in Figures 4.2c and d. Figure 4.5f shows a high-resolution TEM (HRTEM) image of the tip. The fringes are spaced 0.180 nm apart.

These are consistent with the d spacing of Cu {200} planes (JCPDF 89-2838). The dihedral angle of 90° is also consistent with the theoretical value of Cu. Using the information discussed, growth direction of the nanopillars is determined to be along [100] while the four side walls are bounded by {100} planes.

4.3.4 Proposed Growth Mechanism

In Scheme 4.1, the overall growth steps are illustrated to account for the deposition of

Scheme 4.1 Proposed growth mechanism of pagoda-topped tetragonal Cu nanopillars on Au/glass electrode. For clarity, only two side faces of the nanopillar are shown to be covered by DTAC.

pagoda-topped nanopillar. We attribute the growth of these unusual structures to the following reasons. First of all, the over all reduction of Cu²⁺(aq) by Al(s) to form Cu(s) is thermodynamically favored, as suggested by the equation 3 Cu²⁺(aq) + 2 Al(s) → 3 Cu(s) + 2 Al³⁺(aq), E0 = 2.00 V.^[18] This is especially favored at the initial growth stage. In the reaction, a layer of cube-like Cu NPs grew on the substrate after 1h. Apparently, Cu²⁺(aq) ions near the electrode surface were reduced rapidly to deposit these NPs and served as the seeds of growth of the nanopillars. Secondly, diffusion limited conditions in electrochemical deposition system is important for anisotropic growths.^[19] As the Cu²⁺(aq) ion concentration near the electrode surface depleted, a concentration gradient of the ion was formed. This diffusion layer would favor anisotropic growth of the crystals. Finally, the presence of an adequate amount of DTAC is necessary. It probably acted as a capping reagent to assist shape control of the crystals. It is likely that the surfactant molecules self-assembled into a bilayer structure and co-adsorbed selectively on Cu {100} crystal planes to confine the branching growths.^[1,16-17] In absence or deficiency of DTAC, the product grew into a pine tree-like morphology. This was also observed for the product grown at an extended

deposition time. All stems and branches of the tree-like configuration also displayed an analogous tetragonal structure closely related to the nanopillars. Clearly, capping of the DTAC surfactant molecules on {100} stabilized the facets and reduced further extrusion of branches from the stems. On the other hand, the nanopillar tip with the step-edge structure, which was near the bottom of the diffusion layer, could expose more active growth sites and allow more Cu²⁺(aq) reduction on them for further nanopillar extension.^[20] When near perfect {100} side facets were formed, DTAC would passivate them and confine transverse growth.

When excess DTAC was added, large clusters were produced. Obviously, too much DTAC suppressed the anisotropic crystal growth condition because the direct galvanic displacement reaction was carried out at near equilibrium condition. This caused isotropic growth of polyhedron or sphere shaped crystals.^[19]