Comparing low-temperature thermal and plasma sintering processes of a tailored silver particle-free ink

Silver nitrate (AgNO₃, ACS reagent, ≥ 99.0%), sodium hydroxide (NaOH, reagent grade, 97%, powder), oxalic acid (H₂C₂O₄, anhydrous, ≥ 99.0%) , and 2-amino-2-methyl-1-propanol (C₄H₁₁NO, AMP, for synthesis) were purchased from Sigma-Aldrich. Methanol (CH₄O, MA, gradient grade, ≥ 99.9%) was purchased from Honeywell. All chemicals were used as received without further purification. Glass substrates were used. Before ink deposition, 20 mm × 15 mm of glass sheets were ultrasonically cleaned in ethanol for 5 min to remove the particles and organic contaminations on the surface and then dried with a nitrogen gun. To obtain a hydrophilic surface for ink deposition, the substrates were further treated with a 5-min O₂-plasma process (Diener Femto PCCE plasma system, 0.35 mbar, 500 W, 40 kHz).

The silver particle-free ink was formulated by simply dissolving the synthesized silver oxalate powder into alcohol solvents. The silver oxalate powder was prepared via an ion exchange method [34]. For ink preparation, 0.152 g of silver oxalate was first dispersed in 0.45 ml of methanol (MA). After stirring for 1 min, 0.3 ml of AMP was added. The mixture was further stirred for 60 min to produce a transparent ink. The solution was filtered through a 0.45 μm PTFE filter and directly used for the experiments.

The silver conductive films on pre-cleaned glass substrates were obtained by spin coating (3000 rpm, 30 s) with a subsequent thermal or plasma sintering process. 75 μl of the silver ink was used in a dynamic spin-coating process to obtain a thin layer that can cover the whole substrate. The thermal sintering was performed in air with a hotplate at 120 °C, 130 °C, 140 °C, 150 °C, 160 °C, and 170 °C for 5–60 min. The plasma sintering experiments were carried out with a plasma pencil using a low-pressure argon plasma instrument from Diener Electronic. Each plasma sintering experiment consisted of the following steps: (1) The chamber was evacuated to a pressure lower than 0.1 mbar; (2) argon gas was introduced into the chamber at a controlled flow rate (0.2 sccm) that was maintained for the duration of the experiment; and (3) after the flow rate and pressure stabilized, a plasma was initiated. Four power values (100, 250, 400 and 500 W) and five treatment times (5, 10,15, 20 and 30 min) were used for the experiments.

Surface tension and contact angle of the formulated silver particle-free inks on glass substrates were measured with a drop-shaped analyzer (Krüss DSA100, Germany), using the pendant drop method and sessile drop method, respectively. The viscosity measurements were performed on a Haake rotational viscometer at 20 °C under 500 rpm. The thermal decomposition behaviors of the silver particle-free inks were investigated with a differential scanning calorimeter (Netzsch DSC 204 F1) and a thermogravimetric analyzer (Netzsch TG 209 F1). Measurements were carried out in the temperature range 30–300 °C with a heating rate of 10 °C min⁻¹ under an argon atmosphere and a nitrogen atmosphere, respectively. The gas released from the thermogravimetric analyzer during the heating process was monitored with a coupled mass spectrometer (Netzsch QMS 403C). X-ray diffraction (XRD) analysis was carried out over a range of 20° to 90° on a Bruker D8 Advance X-ray diffractometer with a Cu Kα 1 radiation (λ = 0.15406 nm). The surface and cross-section morphologies of the obtained silver films were observed via a scanning electron microscope (HITACHI S-4100). Resistivity values were calculated via ρ = R_s × t, where R_s and t are sheet resistance and the average thickness of the obtained silver films. R_s was measured on a four-point probe instrument (Jandel Engineering), and t was measured from the cross-section image.

In this work, silver oxalate was chosen as the silver precursor of the ink because it has a higher silver content (71 wt%) than most of the organic silver precursors such as silver acetate and can be thermally reduced to metallic silver at low thermal decomposition temperatures (less than 200 °C) without organic residues [17, 34]. Methanol, a volatile and reducible alcohol, was used as the ink solvent to adjust the rheological properties as it has lower surface tension and viscosity than other alcohols such as ethanol and 2-propanol. AMP, an alkanolamine, was selected as the co-solvent of the ink due to its high viscosity (102 mPa·s, 30 °C, Fisher Scientific) and medium boiling point (165 °C).

The influences of the amount of AMP on the ink fluid and electrical properties were first investigated. Inks with different AMP content that are varied from 0.2 ml to 0.6 ml were formulated, as shown in Fig. 1a. When the content of AMP is below 0.2 ml, a transparent ink cannot be formulated. This means that the amount of AMP is not enough to completely react with the silver oxalate. And above this amount, a solution ink can be produced, mainly via the following complexing process (Fig. 1b). The lone pair of electrons on the nitrogen atom of AMP can coordinate with silver cations to produce a soluble silver-amine complex, thereby resulting in clear and transparent ink. FT-IR analysis was employed to confirm this process, as shown in Fig. 1c. The peaks associated with NH₂ groups of AMP undergo a dramatic red-shift after the reaction, implying that a coordination process occurred on the bond of NH₂. This phenomenon is similar to that of other silver-amine particle-free inks such as silver citrate-amine ink [35] and silver carbonate-amine ink [36].

The above results indicate that the AMP can not only act as the solvent in the ink but also as the complexing agent for silver oxalate. This is different from the common silver particle-free ink formula where a complexing agent is always necessary. In other words, our ink system is very simple, only consisting of a silver precursor and solvents.

The fluid properties such as surface tension, viscosity, and contact angle of each ink are given in Table 1; the corresponding images are given in Fig. S1 in the supporting information. As the content of AMP in the ink increases, the viscosity and surface tension of the ink increase gradually. SEM and four-point probe analyses were employed to investigate the morphology and resistivity differences of the film from each ink to determine the optimal content; the results are given in Fig. S2, where the AMP content has a significant influence on the resistivity of the film. As the AMP content increased from 0.3 to 0.6 ml, the resistivity of the film almost quadrupled. This can be easily understood. At high AMP content, the methanol content is relatively small, the ink viscosity is relatively large, the mass transfer is affected, and more organic residues are left, thereby making the conductivity poor. The corresponding SEM images indicate that the film obtained by the ink with 0.3 ml AMP was most compact and uniform. Therefore, this recipe was chosen for the following experiments, where the surface tension and viscosity of the ink were 25.86 mN/m and 6.64 mPa·s, respectively. These values are similar to those reported in our recent work about the printing research [37], which are within the range of fluid parameters required by a piezoelectric printer [4], indicating that the ink is printable. However, due to the lack of high-boiling-point solvents (T_b beyond 150 °C) in this ink formula, early drying and crystallization may appear when printing.

The thermal decomposition behavior of the optimal silver ink was investigated by DSC-TG-MS; the results are given in Fig. 2. The DSC curve shows three main endothermic peaks. The first two, around 75 °C and 130 °C, are attributed to the evaporation of the MA and the partial thermal reduction of Ag⁺. This is in line with the results of MS (mass spectrometry) where MA and CO₂ gases were detected, respectively. The exothermic peak around 150 °C is from the decomposition of AMP [38]. The TG curve displays two main mass loss steps. The first step starts from 30 °C to 130 °C, corresponding to 40.2 wt% mass loss. The second step is between 130 and 180 °C, with a mass loss of 45.9 wt%. Thus, the final mass residue was 13.9 wt%, close to the theoretical silver content (13.2 wt%) of the ink. From the TG-DSC curve, it can be concluded that the decomposition temperature of the silver-amine complex in ink was about 130 °C, indicating the possibility of low-temperature sintering of conductive patterns. In combination with the decomposition temperature of silver oxalate (166–185 °C) [34], it can be concluded that AMP can lower the decomposition temperature of silver oxalate. This mild reducing ability may be related to the properties of the alkanolamine.

For silver particle-free ink, a post-treatment such as drying, curing, or sintering is required for the printed films or patterns to acquire specific electronic properties. Here, thermal sintering and plasma sintering were utilized for our ink system. An ink droplet of 75 μl was spin-coated on the glass substrates in order to achieve homogeneous thin films to assure reliable conductivity measurements. Silver films, obtained at various sintering parameters, were investigated by SEM and XRD to see the differences in the microstructure.

Figure 3a shows the surface morphology of silver films sintered at 120 °C, 130 °C, 140 °C, 150 °C, 160 °C, and 170 °C for 60 min. All the films have a microstructure composed of interconnected domains of silver nanoparticles and few voids. This microstructural feature is a result of the fast evaporation of the solvent and thermal reduction of silver-AMP complex ions, where a lot of bubbles such as MA or CO₂ are produced and need to go through the films, thereby leaving a structure with pores. The film at 120 °C shows a compact microstructure consisting of small particles, mainly due to the insufficient decomposition and evaporation at this temperature. With further increasing the temperature, the size of the voids increased but the particle size became larger and most of them have a good connection with each other. Further increasing the sintering temperature to 170 °C, the microstructure development of silver films evolved to a stage dominated seemingly by the sintering of nanoparticles, as evidenced by such microstructural features as enlarged domain size and coalescence of pores. Here, the porosity and the average particle size of the films sintered at 130 °C, 150 °C and 170 °C were obtained using an image process technique [39] and Nanomeasre software. As the temperature increases, the porosity increases from 8.9% to 21.5%, and then 30.0%, while the average particle size increase from 99.2 to 133.8 nm, and then 206.1 nm. The average pore size for 170 °C is about 150 nm. These results indicate that the film is porous. According to reference [40] and [41], the silver porous nanostructures was an excellent substrate for chemical sensing using the SERS method. Therefore, our Ag porous nanostructures might be suitable for the SERS-related applications such as monitoring pesticide residues in agricultural products. Figure 3b gives the cross-section images of each film. The average thickness is about 300 ± 30 nm after sintering at a temperature above 150 °C for 60 min. The images also show that the film is made up of spherical interconnected silver nanoparticles.

Figure 4a shows the surface morphology of silver films sintered with low-pressure plasma, which is very different from the structure of the films with thermal sintering. The surface is composed of small-sized and multilayered silver nanoparticles and has many filamentous cracks. This structure is a result of the combined effects of plasma effect and rapid solvent evaporation. As the evaporation continues, the film will wrinkle and generate tensile stress. When the tensile stress exceeds the adhesion to the substrate, cracks will occur [42, 43]. However, the morphology of the crack is determined by many factors in the evaporation process, such as the characteristics of the substrate (wettability, roughness, etc.), and the evaporation rate, temperature, film thickness, etc. [44, 45]. These factors increase the difficulty of controlling film morphologies. Segmented sintering with different intensities of argon pressure may help to obtain a uniform film morphology, which can slow down the evaporation rate of the solvents and the decomposition of the silver-amine complex.

At 100 W of power, some small particles were generated on the surface of the glass substrates. These particles have an average size of 50 nm and most of them seem to be surrounded or wrapped by organic molecules. We suspect these small particles are the result of the thermal effect of the plasma [46], which will cause the evaporation of the ink solvents and the reduction of silver ions, thereby leading to the generation of silver particles. As power increases, the number of silver particles becomes more and more and the organic molecules decrease. Further increasing the power of plasma to 500 W, the microstructure development of silver films evolved to a multilayer with few organic matters, where some of the interconnected silver particles are covered on the surface of others. This is attributed to another effect of plasma, the etching effect. The high energy and very active plasma particles break the chemical bonds in residual organic molecules [46] and remove them from the surface of the silver particles so that the particles have good contact. The produced silver nanoparticles merge and connect with adjacent particles, thereby resulting in a conductive network.

Figure 4b gives the morphologies and cross-section images of silver films sintered with plasma at 500 W (the maximum power of our instrument) for different lengths of time. The surface morphology gradually changes as time increases, evolving to a multilayer composed of interconnected silver particles. The average thickness is about 400 ± 30 nm after sintering at 500 W for 15–30 min. The images show that the film is made up of small silver nanoparticles stacked inline form. Here, it should be noted that the sintering of silver particle-free ink is very different from silver nanoparticle ink as the latter can be dried before sintering, that is why the treated films are not as good as reported in the literature in terms of microstructures.

X-ray diffraction was used to probe the crystalline structure of the silver films obtained with thermal sintering and plasma sintering, respectively; the results are given in Fig. S3. All the films show sharp distinct peaks corresponding to the characteristic values for metallic silver, revealing well-crystallized silver films. With the increase of sintering temperature or power, the intensity of diffraction peaks of silver films increases, meaning a better crystallinity.

Figure 5 shows the resistivity of the silver films against sintering temperature or plasma power and time. For thermal sintering, the resistivity decreases close to a plateau value of 24.3 μΩ·cm with increasing sintering temperature to above 160 °C (Fig. 5a). At 120 °C, a resistivity below 10^–5 Ω·cm has already been obtained. This value is higher than that of bulk silver and is associated with the reduction degree of silver-AMP complex ions and the residual degree of AMP that needs a temperature higher than that of 160 °C to remove. However, it can meet the electrical requirements of most electronic products. The changes in the resistivity with time are shown in Fig. 5b. After 5 min of sintering, the resistivity of the silver film is about 40 μΩ·cm, mainly due to the uncompleted reduction reaction. A significant decrease in resistivity was achieved by further increasing the sintering time to 30 min. After that, the resistivity decreases moderately and reaches 25 μΩ·cm at 60 min. For plasma sintering, the resistivity decreases significantly with the increase of plasma power (Fig. 5c) and time (Fig. 5d), reaching a value of 29 μΩ·cm at 500 W for 30 min. Further increasing the power could achieve a lower resistivity. However, this is limited by our available plasma instrument where the maximum power is 500 W.

Here, it is worth noting that the resistivity of the films sintered for 30 min by the two methods is almost the same. This means that the formation of a low resistivity film with plasma sintering of a silver particle-free ink is also possible, although the films from plasma sintering are not as uniform as the thermal sintering. Compared with thermal sintering, in terms of the control and adjustment of film microstructure, plasma sintering is more difficult due to the influences of many factors as previously mentioned. Here, we tried to slow down the drying rate of the ink by increasing the argon flow to 2 sccm in the chamber and the result shows insufficient sintering because a high resistivity was obtained. Only under very low pressure can good electrical performance be achieved.