Fabrication of an anode functional layer for an electrolyte-supported solid oxide fuel cell using electrohydrodynamic jet printing^*

8.0 g nickel oxide (NiO) powder and 6.4 g scandia-stabilized zirconia (ScSZ) were thoroughly mixed with ethanol using ball milling for 24 h. After milling, the mixture was filtrated and dried in an oven at 80 °C for 24 h. Then, a mixture of 6 wt% α-terpineol and 94 wt% ethyl cellulose was added into the milled NiO and ScSZ powder and stirred consecutively to form an anode paste. The paste was finely grinded using a triple roll mill for 30 min The viscous anode paste whose viscosity measured by M-VROC viscometer (US) with 530 cp was used as the ink for the next printing steps.

Typically, the Scandia stabilized zirconia electrolyte plate with the thickness of 0.25 mm was mounted onto the stage of the E-jet printer (PSjet-300-V, Printing Solution, Korea). A syringe containing about 1.0 ml ink was mounted onto the printer. A back pressure was used to deliver ink from the syringe to a nozzle connected to the syringe. The height from surface of the electrolyte plate to the tip of the nozzle was 300 μm. The printing process was carried out at various applied voltage, printing speed, and grid interval. The optimal parameters at which the printed lines were not defective were applied to the next steps. The first printed layer was then dried at 100 °C for 2 h. Subsequently, the second, the third and the fourth layers were perpendicularly printed on the previously printed dried layer to form a multi-layer structured anode functional layer (figure 1). Finally, the dried layers were then sintered at 1300 °C for 3 h.

Zoom In Zoom Out Reset image size

On the other hand, the screen printing method with 80-mesh pore size, 125-micro thickness was also used to make a comparison with E-jet printing method.

To investigate the effect of anode on microstructure and electrochemical performance, a cell with a separately fired anode was also prepared. As shown in figure 2, the first printed layer with 56:44%wt NiO: ScSZ was then dried at 100 °C for 2 h. Subsequently, the second were printed on the previously printed dried layer and at 100 °C for 2 h and then sintered 1300 °C for 3 h. The process was carried out similarly with cathode side with the percentage 50:50 wt% LSM:ScSZ. Finally, a pure lanthanum strontium manganate (LSM) current collector ink was screen printed and fired at 1200 °C for 1 h.

Zoom In Zoom Out Reset image size

The anode functional layers derived from E-jet printing and screen printing process were characterized by using an optical microscope (MX51, Olympus), 3D optical profilometry microscope (Sensofar, Spain), and a field-emission scanning electron microscope (FE-SEM, Hitachi SU8010, Japan).

To get the electrochemical performance, fuel cell evaluation system (Auto-SOFC, TOYO, Japan) was applied. After reduction process of NiO at the temperature 1000 °C for 1 h, I–V measurement was conducted by means of current interruption method in hydrogen fuel with flow rate of 80 ml × min⁻¹ in the anode, and air at a flow rate of 150 ml × min⁻¹ in the cathode side.

The particle size of the powder for the printed ink was characterized by FESEM. As shown in figure 3, the size of the ink particles was approximately from 50 to 150 nm.

Zoom In Zoom Out Reset image size

The novel anode functional layers for electrolyte-supported SOFC were fabricated using E-jet printing technique. The printing parameters such as applied voltage, printing speed, and grid interval were altered to investigate their effect on the structure of the printed layers. Applied voltage is one of the important parameters since it determines the strength of the electric field needed for ink to be jetted. Various applied voltages were investigated for the printing of the anode functional layers. In this case, the pump rate was kept constant at 0.001 ml × min⁻¹. The printing speed and the grid interval were set at 10 mm × s⁻¹ and 0.15 mm, respectively. The applied voltage was varied to investigate its effect on the jetting behavior of the ink. In the range of voltage from 1400 V to 2000 V, the ink deposited in a steady E-jet mode in which the Taylor cone presented continuously and forming a continuous stream of ink from the nozzle to the substrate. At the voltage below 1400 V, the ink jetted in a non-continuous mode because the strength of electric field was not large enough to maintain a stable Taylor cone. At the voltage above 2000 V, complex jetting behavior occurred with tilted jets and multiple jets. Figure 4 shows the optical microscope images of the anode functional layers printed at different applied voltages. In the images, the substrate is bright, the printed lines are the grey lines with the line thickness of around 60 μm.

Zoom In Zoom Out Reset image size

As seen in the figure 4, the line width varies negligibly when increasing the applied voltage from 1400 V to 1850 V. This indicates that the applied voltage in the range for the steady cone-jet mode had an insignificant effect on the deposition speed of ink. In this case, the pump speed was kept the same for different applied voltages that delivered a similar amount of ink to the head of the nozzle. In addition, the printing speed was the same that subsequently led to a similar deposition speed of ink at different applied voltages.

Figure 5 illustrates the optical microscope images of the anode functional layers printed at different printing speeds. In this case, the applied voltage and the pump rate were 1600 V and 0.004 ml × min⁻¹, respectively. As seen in the figure, the printed lines are 2–3 times bigger than those in figure 4. This is due to the higher pump rate used in this case that leads to more ink delivered to the head of the nozzle. The images show that when increasing the printing speed, the printed lines became thinner. Table 1 shows the size of the lines printed at different printing speeds. As seen in the table, when the printing speed increased from 6 to 30 mm × s⁻¹, the average line thickness decreased from 196.6 to 132.1 μm. The layer thickness of the printed lines was measured by breaking the samples and observing the cross sections using an optical microscope. The average layer thickness decreased from 5.0 to 2.8 μm when the printing speed increased from 6 to 30 mm × s⁻¹. In the E-jet printing, at a constant pump rate and applied voltage, the stream of ink jetted from the nozzle is quite stable with an almost constant ink delivery speed. As the printing speed increased, the substrate moved faster during printing. Thus the printed lines received less amount of ink due to the shorter time necessary to get the same printing distance compared to that at lower printing speed.

Zoom In Zoom Out Reset image size

After investigating the E-jet printing parameters, the optimal parameters (table 2) were applied to fabricate the multi-layer structured anode functional layers.

Figure 6(a) shows the optical microscope images of the anode functional layers printed at an applied voltage of 1800 V and a pump rate of 0.001 ml × min⁻¹ with grid interval 0.04 mm. It is obvious to see that the printed lines were almost continuous and quite with the same width, about 0.13 mm. Figure 6(b) shows the optical microscope image of multilayer-structured anode functional layers, the next printed layer being fabricated on the previously printed dried layer (four layers). The functional layers almost covered the surface area of the electrolyte plate completely due to the grid structure of anode functional layers.

Zoom In Zoom Out Reset image size

The FESEM images of the anode functional layer at the surface (figure 7) and the cross section (figure 8) with 1 layer, 4 layers printed by E-jet printing and the layer printed by screen printing. Figures 7(a) and (b) show the FESEM images of single layer anode strips on the electrolyte plate. The thinness of anode caused the discontinuous bands and left some unfilled areas when observed at high magnitude. Figures 7(c) and (d) show the FESEM images of the layers with grid structure and many pores created from the perpendicular printed lines. The grid structure formed an enough thin and foamy membrane to cover a consecutive surface area. Such a structure is expected to have various advantages compared to the continuous layers fabricated using the conventional methods. As seen in the images, the second printed layer was fabricated on the first layer with the printed lines perpendicular to those of the first layer.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Some printed lines of the second layer were fused together which might be due to the rough surface of the substrate after fabricating the first printed layer that made ink at the higher positions flowing to the lower positions and spreading out there. For example, with many pores of the grid structure, the anode functional layer can facilitate the diffusion of fuel gas in and gaseous products out of the layer during the operation of SOFCs. In addition, this grid-structured anode functional layer may have higher active surface area, and greater number of reaction sites for better electrochemical reactions. Thus, the performance of SOFCs could be improved.

Moreover, because the operating temperature of SOFCs is very high, a continuous electrode layer fabricated using the conventional methods commonly faces the problem of cracking and delamination from the electrolyte layer due to thermal expansion mismatch between the electrode material and the electrolyte. This problem makes the electrode not in good contact with the electrolyte layer, thus decreases the performance of the SOFC. A grid-structured electrode layer may be more flexible than a continuous layer, thus can mitigate the cracking and delamination of electrode. With this technique of printing of multiple electrode layers, the thickness of the anode functional layers can be easily adjusted to meet different requirements in the fabrication of SOFCs.

Figure 8 shows the thickness of layers derived from E-jet printing and screen printing. The images indicate that the thickness of E-jet printed layers was roughly 7–10 times less than that of the screen printed one, with 15 μm of thickness (figure 8(f)). Nonetheless, the thickness of E-jet printed samples did not proportionate to the number of layers printed. In the figures 8(b) and (d), the thickness of 4 layers just only doubled the one of single layer. It could be inferred that the viscosity of the anode ink caused the ink flowing down to the slits between ledges of the first printed strips at the second time printing which resulted in no increase in the height of the second layers. Thus, the thickness of 2 layers was as much as the one of single layer. At the third and fourth time, the ink overlaid on the first layer or filled the indented spaces left over the second printing. As a result, the height in total 4 layers folded 2 times the first one. Only about 2 μm of thickness, the 4-time printed one was 7 times thinner than the screen printing but still thick enough to obtain the anode functional layer with the same activities area.

The 3D optical images of the thickness of one anode function layers printed at voltage applied 1800 V and the screen printing layer are shown in figures 9 and 10, with the thickness of anode layers approximately 3 and 18 μm, respectively. There was a significant difference about the thickness measured from FE-SEM and 3D images. The difference might be due the different positions and principle of each measurement.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The cell performance tests were executed at operation temperature at 800 °C and the results of I-V and I-P curves are presented in figure 11 with fuelled by H₂ and air at the flow rates of 80 and 150 ml × min⁻¹, respectively. Obviously, the OCV values of both cells are approximately 1.1 V, which indicates that no gas leakage occurred through the ScSZ electrolyte layer of the cell and no crack existed on ScSZ film during the performance test. The maximum power densities are 0.33 and 0.30 W cm⁻² by using screen printing and E-jet printing, respectively. The difference between the performance of the E-jet printed cell and screening-printing cell is relatively negligible. However, E-jet printing technology benefits from the thickness which is 7–10 times thinner than the screen-printing one, resulting in low cost procedure.

Zoom In Zoom Out Reset image size