Test materials included anhydrous copper sulfate (CuSO₄, 99.0%, Tianjin Fengchuan Chemical Reagent Technology Co., Ltd.), acacia gum (Select delicate level by hand, Shanghai Aladdin Biochemical Technology Co., Ltd.), Hydrazine hydrate solution (N₂H₄∙H₂O, 80%, Xilong Science Co., Ltd.), acetone (C₃H₆O, 99.5%, Beijing Chemical Plant), absolute ethanol (C₂H₆O, 99.5%, Beijing Tongguang Fine Chemical Company), acetaldehyde solution (C₂H₄O, 40%, Tianjin Damao Chemical Reagent Factory), silver nitrate (AgNO₃, 99.0%, Beijing Hengye Cosco Chemical Co., Ltd.), glycerol (C₃H₈O₃, 99.0%, Beijing Hengye Cosco Chemical Co., Ltd.), Spredox D-364 (Taiwan Texas Chemical Company), and Surfynol 465 (Air Products Inc.), BYK-028 (Germany Bike Company).

Copper nanoparticles were prepared by a chemical reduction method. Copper sulfate was reduced in deionized water with hydrazine hydrate as the reductant and gum arabic as a dispersant. Gum arabic (0.6 g) and CuSO₄∙5H₂O (0.3 g) were weighed out, and 25 mL of deionized water were added with stirring until fully dissolved. The mixture was stirred with a magnetic stirrer for 10 min at 70˚C, then 2.6 g hydrazine hydrate solution (80%) was added and a high-speed disperser was used to stir the mixture for one h. After the reaction, a purple-red colloid was obtained. The prepared nano-copper sol was centrifugally separated (12,000 rpm). The precipitate was washed three times with ethanol and acetone in sequence. A red-brown powder was obtained by drying in a vacuum drying chamber at 40˚C for 12 h.

The purple-red nano-copper sol obtained from the first step was reacted for 10 min with 20 mL acetaldehyde solution and an excess of hydrazine hydrate in the reaction system. Then 0.2 g of AgNO₃ was dissolved in 20 mL deionized water, and silver nitrate solution was slowly added. A dark red dispersion was obtained after stirring at room temperature for 20 h at high speed. The nano Cu@Ag particle dispersion was placed in a high-speed centrifuge, and the precipitate obtained by centrifugal separation (12,000 rpm) was washed three times with ethanol and acetone in sequence. Finally, Cu@Ag powder was obtained by drying in a vacuum drying chamber at 40˚C for 12 h.

With deionized water as solvent, conductive ink with a solids content of 10 wt% was obtained by high energy ball milling with 0.5 wt% predox D-364, 3 wt% glycerol, 0.3 wt% Surfynol 465 and 0.1 wt% BYK-028 as additives. Ink-jet printing was done on polyimide (PI) film by an Epson L485 printer and the product was sintered in a vacuum tube furnace for 30 min.

The phase of the synthesized particles was detected by AXS D8 Advance X-ray diffractometer (Brooke Instruments, Germany). Ultraviolet-visible spectrum analysis was obtained by a Cary 5000 Ultraviolet-visible Spectrophotometer (Agilent Technology Co., Ltd.). The morphology and particle size of the particles were characterized by JEM-2100F transmission electron microscopy (TEM) (JEM-2100F) or by a Nanozetasizer-ZS laser particle size analyzer (Malvin Company, UK). The conductivity of the printed layer was tested by a TECH RTS9 four-probe resistor (Guangzhou Four-probe Technology Co., Ltd.). A Nova NanoSEM 450 Scanning Electron Microscope (FEI) was used to observe the surface morphology of conductive films.

In this experiment, hydrazine hydrate was used as a reductant to quickly reduce copper ions in solution to elemental copper, and no new impurities were introduced into the reaction system, as shown in Equation (1). As the reaction proceeds, the initially formed copper reacts reversibly with unreacted Cu²⁺ (2). In order to inhibit the occurrence of reaction (2), an excess of hydrazine hydrate was used in this experiment. This also protects nano-copper from oxidation in solution.

N 2 H 5 + + 2Cu 2 + → 2Cu + N 2 + 5H + (1)

Cu 2 + + Cu → 2Cu + (2)

The XRD and UV-vis diagrams of synthesized nano-copper are shown in Figure 1. In Figure 1(a), only three distinct diffraction peaks appear at 43.34˚, 50.48˚, and 74.18˚, which correspond to the (111), (200), and (220) crystal planes of face-centered cubic copper. Figure 1(b) is the ultraviolet-visible absorption spectrum of nano-copper sol. There is an obvious plasma resonance absorption peak at 580 nm in the visible region, which is the characteristic peak of copper. No other impurity peaks appeared, which proves that nano-copper was prepared. Figure 2 shows a TEM image and the DLS particle size distribution of the prepared nano-copper powder. TEM images show that the prepared nano-copper samples have small particle size and spherical shape, and their size distribution ranges from 11 to 34 nm, which indicates that gum arabic has a good protective effect on nano-copper.

When AgNO₃ solution is added to the prepared nano-copper sol, the substitution reaction Cu + 2Ag⁺ = Cu²⁺ + 2Ag takes place to obtain a silver coating on the surface of the nano-copper. Because of the difference in standard redox potential between silver (Ag⁺ + e⁻ → Ag, E⁰ = 0.779 V) and copper (Cu²⁺ + 2e⁻ → Cu, E⁰ = 0.46 V), electrons are transferred from Cu⁰ to Ag ions, and silver is replaced. In the reaction, one copper atom can produce two silver atoms, so theoretically a dense silver-coated layer can be obtained. The morphology of the product is similar to that of the previously prepared copper core, and the particle size increases slightly. When an excess of hydrazine hydrate is used to prepare nano-copper sol, the standard reduction potential difference between hydrazine hydrate (N₂H₄ + 4OH⁻ → N₂ + 4H₂O + 4e⁻, E⁰ = −1.16 V) and silver is about 1.96 V,

which is much higher than that between silver and copper. Therefore the products obtained by directly adding silver nitrate solution are silver particles and a small amount of silver-coated copper powder. In order to prevent silver nitrate from being reduced directly by excessive hydrazine hydrate, the excess hydrazine hydrate is consumed by adding acetaldehyde solution (C₂H₄O + N₂H₄ → C₂H₆N₂ + H₂O) to the reaction system before adding the silver nitrate solution.

Figure 3 is the ultraviolet-visible absorption spectrum of Cu@Ag sol. Not only does the plasma resonance peak of nano-copper appear near 590 nm in the ultraviolet region, but also the characteristic absorption peak of silver appears near 450 nm. It can be seen that the characteristic peaks of silver in the figure are significantly stronger than those of copper. A possible reason is that silver is coated on the surface of the copper. TEM images of the above-mentioned Cu@Ag particles are shown in Figure 4(a). The morphology and particle size of the Cu@Ag particles are not significantly different from those of the previously prepared nano-copper particles. As shown in Figure 4(b), the silver (111), (200), and (220) rings and the copper (220) and (311) rings are obtained by selective electron diffraction of the sample, but the rings look blurred. The reason for this is that the lattice constants of copper and silver are different, the distribution of the interface is discontinuous, and the optimum diffraction conditions for copper and silver are different. Therefore, some diffraction rings are not clear when we see the copper and silver diffraction rings at the same time, and some crystal planes do not have diffraction rings [10].

In order to explore the effect of different reaction conditions on the silver coating effect, Cu@Ag particles were prepared by controlling the reaction temperature and the mass ratio of copper to silver. The products were placed in air for one week and then analyzed by XRD. The XRD pattern is shown in Figure 5.

The results show that the products have diffraction peaks at 43.34˚, 50.48˚, and 74.18˚, corresponding to the (111), (200), and (220) faces of face-centered cubic copper, respectively. Diffraction peaks appeared at 38.12˚, 44.36˚, 64.43˚, and 77.41˚, corresponding to the (111), (200), (220), and (311) faces of face-centered cubic silver, respectively. In Figure 5(a), when an excess of silver nitrate is added, the diffraction peak of Cu₂O appears when the reaction temperature reaches 60˚C, and the diffraction peak of Cu₂O increases gradually with an increase in the reaction temperature. This is because when the reaction temperature is high, the reaction rate is very fast, and silver ions are rapidly reduced to silver, but they cannot be well deposited on the surface of the copper core, resulting in an incomplete and compact coating. The uncoated copper atoms react with the Cu²⁺ produced by the displacement reaction: Cu²⁺ + Cu = 2Cu⁺, while free Cu⁺ reacts with OH⁻ in the solvent to produce Cu₂O, which results in the formation of impurities. Therefore, the reaction is more suitable for our purposes at 20˚C. As indicated in Figure 5(b), when the reaction temperature was 20˚C and the ratio of copper to silver was 5:1, CuO was detected one week after the product was produced, which indicates that the silver coating cannot protect the copper core well and the coating effect is not good. However, when the ratio of copper to silver was 4:1, 3:1, or 2:1, no oxide was detected after one week, which indicates that the prepared Cu@Ag nanoparticles had a good coating structure. Particles with a good coating effect can be obtained when the atomic ratio of copper to silver is 4:1, but the cost will increase if the amount of silver nitrate is increased. Therefore, the appropriate amount of silver nitrate to add is a ratio of copper to silver of 4:1.

Conductive ink was printed on PI film by inkjet printing. After vacuum heating for 30 min, the resistivity of the conductive lines formed at different temperatures of the coating was measured by a four-probe tester. As shown in Figure 6, with an increase of post-processing temperature, the resistivity of conductive lines decreases continuously, and the conductivity increases. The resistivity is 682.1 μΩ∙cm when the post-treatment temperature is 100˚C. When the post-treatment temperature is 150˚C, the conductivity of the sample is greatly improved and the resistivity decreases to 98.6 μΩ∙cm; at 250˚C the resistivity is only 14.8 μΩ∙cm. The resistivity decreases slightly with increasing sintering temperature, reaching 10.6 μΩ∙cm at 300˚C. The reason for the increase in conductivity of the conductive circuit may be that the organic additives in ink volatilize during sintering of the nano-Cu@Ag particles, and the adjacent particles form a continuous conductive path in the melted state. With an increase in temperature, the conductive path increases gradually, and the electrons can be conducted more freely in the conductive circuit.

According to the first law of diffusion, the definition of diffusion coefficient D, which is the key factor affecting diffusion rate, is as follows: D = exp ( − Q R T ) D 0 ,

where Q is the diffusion activation energy, R is the gas constant, T is the thermodynamic temperature, and D₀ is the diffusion constant. It can be seen that the diffusion coefficient increases with an increase in temperature. That is to say, with an increase of temperature, the diffusion coefficient increases and the degree of fusion between particles increases. In order to verify that nano-Cu@Ag particles melt into continuous thin films after heat treatment, scanning electron microscopy (SEM) was used to characterize the micro-morphology of conductive circuits at different temperatures. As shown in Figure 7, after high temperature treatment at 150˚C, the particles begin to melt and sintering neck appears. Metal atoms diffuse from unstable grain boundary positions to form sintering neck through diffusion.

After sintering at 200˚C, the sintering neck size increases, and the number increases, volume shrinkage is not obvious, the atom diffusion channel widens, and more continuous seepage paths are generated in the circuit, which is consistent with the trend of resistivity reduction shown in Figure 6. When the sintering

temperature reaches 250˚C, the size and density of the sintering neck increase greatly, the particles connect with each other, the volume shrinks, and the resistivity decreases significantly. When the sintering temperature reaches 300˚C, the particles fuse with each other to form a highly dense conductive layer. It is also shown that the conductivity of the conductive film cannot reach the level of bulk materials because of the large number of voids generated by gaps between the original particles and inhomogeneous volume shrinkage after the temperature reaches 250˚C.