Electrochemical Investigation of the Growth of Copper Nanowires in the Presence of Ethylenediamine through Mixed Potential

Copper (Cu) nanowires about 90 nm in mean diameter and lengths of about 50 μm were successfully synthesized by electroless deposition at 60◦C in the presence of ethylenediamine (EDA). Without addition of EDA, only large Cu particles were formed in the solution. On the other hand, high quality nanowires were formed in the presence of 176 mM EDA. Below this amount, shorter Cu nanowires were observed. The mixed potential shifted to more negative values without EDA, indicating faster reduction rate. It is possible that EDA does not only act as a structure-directing agent, but also slows down Cu deposition. This condition favors formation of long and thin Cu nanowires. Then again, in all EDA concentrations experiment, the mixed potential is below the oxidation-reduction potential of the Cu(II)/Cu redox pair, suggesting Cu depositions occurs instantaneously at 60◦C. Cu nanowire transparent conducting electrode attained a sheet resistance of about 197 sq−1 at a 61% optical transmittance. © The Author(s) 2017. Published by ECS. This is an open access article distributed under the terms of the Creative Commons Attribution 4.0 License (CC BY, http://creativecommons.org/licenses/by/4.0/), which permits unrestricted reuse of the work in any medium, provided the original work is properly cited. [DOI: 10.1149/2.0491707jes] All rights reserved.

Copper (Cu) nanowire is gaining a lot of attention due to its viability as a material for transparent conductive film commonly used in various optoelectronic devices, such as flat panel display, organic light emitting diodes (OLED) and touch sensor. [1][2][3] Bulk Cu has high electrical and heat conductivity. [4][5] Thus, Cu is traditionally used as wires in mains cables in houses and buildings and as heat exchangers in hot water tanks. Cu is also an abundant element and has excellent recyclability. This makes Cu an attractive alternative for both indium tin oxide (ITO) film and silver (Ag) nanowires for transparent conducting applications in future electronic devices.
In recent reports, surfactants, such as ethylenediamine (EDA), 1,6,8,12,13 hexadecylamine (HDA) 7 and octadecylamine (ODA), 2,10 have played a significant role in preventing oxidation and controlling the dimension of Cu nanowires prepared in water. Other studies have even utilized two surfactants at the same time to further increase length of the Cu nanowires and improve their surface quality. 3,4,11,14 For example, the diameter of Cu nanowires synthesized with EDA was decreased by about 55% when a small amount of cetyltrimethyl ammonium bromide (CTAB) was added into the solution. 14 Additionally, the surface roughness of the Cu nanowires was significantly reduced. 14 However, most of these studies on Cu nanowires prepared in aqueous media are still focused on the optimization of solution parameters for morphological control.
In our previous works, we have applied simple electrochemical measurements, such as in situ monitoring of mixed potential, to understand the growth of cobalt (Co) nanoparticles and nanowires, 15,16 nickel (Ni) nanoparticles and nanowires, 17,18 and Cu nanoparticles [19][20][21] during electroless deposition in water and ethylene glycol solutions. More recently, we have utilized mixed potential measurements, in combination with thermodynamic calculations, to elucidate the ef-fect of temperature on the growth of zinc oxide (ZnO) nanostructures during wet oxidation of Zn foil in hot water. 22 By comparing the kinetically measured mixed potential with the thermodynamically calculated oxidation-reduction potential of the target metal, we can determine the stable chemical species of the target metal in the solution. [19][20][21][22] In effect, we can determine the solution condition that facilitates metal deposition or oxidation. Such simple approach also allows understanding of the kinetics and thermodynamics that influences the growth of metallic nanostructures in solution. [19][20][21][22] More recent work by Kim et al. 23 has presented similar method to explain the formation of Cu nanowires in the presence of surfactant.
Therefore, in this work, we present a simple preparation process for Cu nanowires by electroless deposition in aqueous solution using EDA as surfactant. The effects of EDA concentration on the morphology and surface quality of the Cu nanowires were investigated. In situ monitoring of mixed potential, in conjunction with thermodynamic calculations, was performed to elucidate the formation of Cu nanowire in solution and the influence of solution parameters on nanowire growth. nanowire solution was centrifuged at 1500 rpm for 3 min to collect the Cu nanowires. Then, the Cu nanowires were sonicated in approximately 3 mL of the ink and then centrifuged at 1500 rpm for 3 min. The washing step was repeated three times. After the last washing, a small amount (∼0.5 mL) of ink was finally added to the collected Cu nanowires.

Electroless deposition of Cu nanowires in aqueous solution in
Before deposition of the Cu nanowire ink, the glass substrates (3 × 1 in) were first sonicated in acetone for 10 min, followed by washing in isopropanol to remove surface contaminants. The slides were wiped and blown dry under nitrogen (N 2 ) gas immediately prior to nanowire deposition. To coat the glass slides, ∼25 μL of the Cu nanowire ink was dropped on the upper part of the glass slide. Using a Meyer rod (#10, RD Specialist), the ink solution was evenly coated onto the glass substrate by one quick (∼1 s) downward sliding movement.
Characterization.-The morphology of the resulting Cu products was observed using scanning electron microscope (SEM, JEOL 5300) and high resolution transmission electron microscope (HRTEM, JEOL JEM-ARM200F). The average diameter and length of the Cu nanowires were determined by measuring the dimensions of over 300 nanowires from several SEM images. Additionally, selected area electron diffraction (SAED) of the Cu nannowires was performed during HRTEM analysis to determine the growth direction. Phase analysis was carried out using X-ray diffraction (XRD, Cu Kα, Shimadzu XRD-7000).
In-situ mixed potential measurements were carried out on a platinum (Pt) sheet with an active surface area of 1 mm 2 using a potentiostat/galvanostat [Metrohm Autolab]. A double junction silver/silver chloride electrode [Ag/AgCl, Metrohm] immersed in 3.3 M potassium chloride (KCl) aqueous solution was used as the reference electrode. In this study, all potentials were reported against the standard hydrogen electrode (SHE).
The sheet resistance (R s ) was measured using a four-point probe [Jandell], at predetermined distances across the glass substrate. All reported sheet resistance values are the average of 10 individual measurements. The sheet transmittance of the Cu nanowire electrode was measured using UV-VIS spectroscopy. All reported values were taken at λ = 550 nm, which coincides with the spectrum of visible light. In order to determine the amount of Cu nanowires deposited on the glass substrate, the film was etched using 1 M HNO 3 . The Cu content was analyzed using microwave plasma -atomic emission spectroscopy (MP-AES). Figure  1 shows the SEM images of Cu nanowires prepared by electroless depositon at 60 • C using increasing amounts of EDA. In the absence of EDA, only spherical particles with diameters ranging from 100-500 nm were formed in the solution as shown in Fig. 1a. When less than 40 mM of EDA is added, both spherical and rod-like morphologies were produced as seen in Fig. 1b. Only Cu nanowires attached to spherical seeds were observed after the addition of 71 mM EDA, which indicates that there is a minimum concentration of EDA required to promote nanowire growth. The Cu nanowires have a mean diameter of about 60 nm and mean length of about 15 μm. As the EDA concentration is increased to 141 mM, longer nanowires with mean length of 33 μm were precipitated as in Fig. 1d. However, the Cu nanowires have thicker mean diameter of about 120 nm. Further increase in the amount of EDA to 176 mM produced nanowires with mean length of about 50 μm and average diameter of about 90 nm. This results in Cu nanowires with large aspect ratio of about 450. However, addition of EDA above 176 mM, specifically 211 mM, resulted in larger Cu nanowires, with mean diameter of about 242 nm. In all concentrations experimented, the wire distribution was random, which is typical of template-free processes for the synthesis of nanowires.

Synthesis of copper nanowires in the Presence of EDA.-
As seen in Fig. 1b, addition of EDA is necessary to initiate the growth of Cu nanowires. Without EDA, only spherical particles are generated in the solution. Spherical is the most thermodynamically stable shape assumed by particles to decrease their total surface energy per unit volume. This suggests that EDA acts as a structure directing agent, which is responsible for the formation of nanowires in the solution. [1][2][3][4][5][6][7][8][9][10][11][12][13] As such, anisotropic growth is observed. However, adding any amount of EDA does not ensure complete nanowire formation as seen in Fig. 1b, where both spherical and rod-like products are observed. It is also evident in Fig. 1b that the rods protrude out of the spherical particles, which may shed light on the growth mechanism of the Cu nanowires. An optimum amount of EDA is possibly necessary to ensure complete capping of the crystallographic planes of the Cu crystals, leading to wire growth. 1,6,8,12,13 Additionally, it was found that the dimension of the Cu nanowires can be controlled to some degree by increasing the EDA concentration. As seen in Figs. 1c-1e, larger amount of EDA led to higher aspect ratio nanowires. However, excessive amount of EDA may result to chaotic deposition. This could prevent deposition of new Cu atoms at the ends of the wires, which results in the thicker and shorter nanowires in Fig. 1f. [11][12][13] The corresponding XRD patterns of Cu nanowires synthesized using 0-211 mM EDA are shown in Fig. 2 The measured mixed potentials were almost constant and no significant change was observed after 1 h reaction. In the absence of EDA, a sharper decrease in the mixed potential was initially observed (−0.902 V) up to about 20 min in the reaction. This period also corresponds to a change in the solution color from royal blue to colorless, indicating reduction. After 20 min, the solution turned colorless, indicating end of the reaction. It is possible that there is an increase in the overall reaction rate when EDA was not added in the solution. EDA possibly forms a complex with Cu(II) ions, thereby regulating the activity of Cu(II) ions in the solution. Without EDA, more Cu(II) ions are available for reaction. Thus, more Cu crystals are generated in the solution at a time, which explains the small crystallite size determined from the XRD pattern of the spherial particles formed without EDA in Fig. 2a. In the same way, the Cu(II) activity may also decrease to a certain extent when excess EDA is added. As such, the deposition may slow down, leading to a larger crystallite size. The increased Cu(II) activity, and consequently faster reaction, may also be deduced from the sharper drop in the mixed potential of the reaction in the absence of EDA as shown in Fig. 3a.
After 20 min reaction, the mixed potential of the solution without EDA steadily increased up to about −0.50 V as seen in Fig. 3a. At this point, no more Cu deposition possibly occurs in the solution as indicated by the colorless solution and presence of precipitate at the bottom. However, the negative value of the mixed potential can be attributed to other faradaic reaction occurring simultaneously with Cu deposition, i.e. hydrogen generation. It is possible that the remaining N 2 H 4 in the solution continuously reacted with H 2 O since the redox potential was still below the calculated oxidation-reduction potential of H + /H 2 redox pair, E H+/H2 vs SHE = −0.606 V at the same temperature and pH values, thereby producing H 2 gas. [16][17][18] This is consistent with what was observed during the reaction where gas evolution was observed simulatenously with the appearance of Cu nanowires, and continues well even after no more Cu nanowires were precipitated.
Formation mechanism of copper nanowires in the presence of EDA.-In order to understand the growth process of the nanowires, SEM images were taken at different reaction time during the electroless deposition of Cu nanowires with 171 mM EDA at 60 • C as shown in Fig. 4. The reaction started upon heating of the total solution from room temperature to 40 • C. After about 2 min of heating, the solution color changed from royal blue to milky white indicating a reaction. SEM image of the Cu product showed non-uniform spherical seeds as in Fig. 4a. These seeds possibly acted as nuclei upon which the succeeding nanowire grew. When the solution reached 60 • C after about 5 min of heating, the rods started to project from the seeds as shown in Fig. 4b. After about 10 min, short but thin nanorods with diameters of about 73 nm and lengths of a few microns were produced. As the   The corresponding XRD patterns of the Cu products at varying reaction times are shown in Fig. 5. The peaks at 43.5, 51.0, and 74.5 • correspond to the 111, 200, and 220 peaks of metallic Cu. Similar to previous XRD patterns, no other peaks attributed to Cu oxide and/or hydroxide was observed in the diffraction patterns for Cu products at 2, 5, and 10 min. This suggests that the spherical seeds obtained at the early stages of the reaction is purely metallic Cu, which corresponds well with the observed changes in the mixed potential for this sample shown in Fig. 3b. It can be seen from the plot that the mixed potential was well below the calculated reduction potential (E Cu(II)/Cu vs SHE = −0.22 V) even at the early stages of the reaction indicating Cu deposition.
High resolution TEM (HRTEM) images were taken from Cu nanowires produced using 171 mM EDA at 60 • C. The Cu nanowire diameter is not uniform along its length as seen in Fig. 6a, suggesting a rough surface. However, the mean diameter was found to be in the range of 90-120 nm, which agrees well with apparent diameter measured from the SEM images. It is clear from the inset of Fig. 6a that the Cu nanowire has high lattice perfection. Lattice fringes were observed and were found to have a spacing of about 2.06 Å. This corresponds to the interplanar spacing of 111 planes. The lattice constant, a 0 , assuming a cubic phase was then calculated using this value to be about 3.615 Å. This is comparable to the lattice spacing of fcc Cu, which agrees with the XRD results that only metallic Cu is produced in the solution. Minute spherical Cu nanoparticles were visible on the surface on the Cu nanowires. Selected area electron diffraction (SAED) at the center of the Cu nanowire showed a spot diffraction pattern, indicating high crystallinity. The growth direction was also determined to be along the [110] direction. Presumably, EDA attaches preferentially at the 200 planes and stabilizes these planes. 1,4,24,25 As such, it leaves the 111 planes available for Cu deposition. The 111 plane has the lowest surface energy for fcc structures and they tend to be exposed on the crystal surface to minimize the total surface energy. Thus, Cu atoms deposit on the 111 planes as the reaction progresses. However, the rough nanowire surface implies imperfect capping of EDA along the 200 planes, which leads to the uneven nanowire surface. Finally, growth occurs along the [110] direction. [24][25][26] Copper nanowire ink and transparent conducting electrode.-To investigate the oxidation stability of Cu nanowires in the ink formulation, successive UV-Vis samples were taken. Figure 7 shows the UV-Vis of Cu nanowire ink taken at 0, 20, and 45 days. After ink preparation, a peak at about 590 nm is observed which corresponds to metallic Cu. A slight shift to the right, up to about 600 nm is observed after 20-45 days possibly due the agglomeration of the Cu nanowires. 27 Then again, no other peak related to Cu oxides was observed even after 45 days. This shows that the Cu nanowires are oxidation resistant in the ink solution. This is critical to ensure high conductivity of the Cu nanowire ink. Figure 8 shows the AFM images of Cu nanowires deposited on glass. The wires were able to retain their length even after cleaning,   Figure 9a shows the plot of sheet resistance with increasing nanowire density. The sheet resistance was about 9617 sq −1 for 30.3 μg/cm −2 and decreases to about 197 sq −1 as the number of Cu nanowire coatings is increased to 151.5 μg/cm −2 . Regression analysis indicates that the sheet resistivity is a logarithmic function of Cu nanowire density. 28 This agrees well with the widely accepted theory on percolation networks which explains that the probability of nanowires to form junctions drastically increases at higher concentrations of nanowires. [28][29][30][31] This is because a single nanowire can connect several other nanowires across the surface.
On the other hand, the transmittance linearly decreases with higher nanowire density as seen in Fig. 9b. A single coating of the Cu nanowire ink on glass can transmit 84.5% of visible light. However, the optical transparency decreases to 61% after 5 coatings (151.5 μg/cm −2 ). It is known that metallic materials like Cu nanowires cannot transmit light and the optical transaprency of the Cu nanowire glass electrode can be attributed to the spaces on the electrode not occupied by the nanowires. Thus, it is expected that as the Cu nanowire density is reduced, there are more spaces between the Cu nanowires. This leads to higher transmittance. Several studies have also shown that the aspect ratio of the nanowire largely influences the overall performance of the electrode. [29][30][31] Simply put, longer nanowires are able to connect to more junctions than shorter wires, but by maintaining a very small nanowire diameter, the amount of light which is allowed to pass through remains unchanged. In addition, the relationship between transmittance and sheet resistance of the Cu nanowire glass electrode is clearly seen in Fig. 9. As the Cu nanowire density is increased, more junctions among the nanowires are formed. This results to a decrease in both sheet resistance and transmittance of the Cu nanowire glass eletrode.
Finally, as a proof of concept, the charge carrying capactiy of the Cu nanowire electrode was tested by using it to light a 5 mA LED in a series connection as shown in Fig. 10a. The bulb was successfully lit indicating that the nanowires had good electrical connections across the 2 × 1 in substrate. Increasing nanowire density leads to brighter light from the LED at the same applied voltage as shown in Fig. 10c. This is due to lower sheet resistance values for the electrode. Then again, even at a higher nanowire concentration, the electrode still retains its transparency, as the university logo is still visible.

Conclusions
In summary, Cu nanowires were successfully grown through a low temperature electroless deposition in aqueous solution for 1 h. Ethylenediamine (EDA) was necessary in order to promote anisotropic reduction of Cu(II) by hydrazine. Without EDA, only Cu particles with diameters of around 100-500 nm were formed. Cu nanowires with mean diameters around 90 nm and lengths exceeding 50 μm were synthesized using 171 mM EDA at 60 • C, giving an effective aspect ratio of about 450. However, excess EDA promotes chaotic deposition leading to shorter and thicker nanowires. In situ monitoring of mixed potential, in conjunction with thermodynamic calculations, was able to provide insight on the role of EDA in the formation of Cu nanowires. It is possible that EDA, a structure-directing agent, also influences the reduction rate to some extent, which results to a more favorable condition for nanowire growth. Lastly, a transparent conducting electrode was made from Cu nanowire ink-coated glass substrate. The sheet resistance of the Cu nanowire glass electrode was measured to be about 9617 sq −1 at 84.5% transparency for 30.3 μg/cm −2 of Cu nanowires coated on a 3 × 1 in glass. The sheet resistance decreases to around 197 sq −1 at 61% transparency as the number of nanowires was increased to 151.5 μg/cm −2 . Finally, a 5 mA LED was powered in a series connection using the Cu nanowire glass electrode in order to prove its viability.