Airflow assisted printhead for high-resolution electrohydrodynamic jet printing onto non-conductive and tilted surfaces

Advancements in technology and industry developments have created a significant demand for high-resolution fabrication capabilities. Some of these demands, such as small-scale productions of highly customized and high-resolution features, are currently underserved due to cost and material restrictions of traditional high-resolution fabrication techniques (e.g., optical lithography and inkjet printing) Electrohydrodynamic jet (e-jet) printing has emerged as a cost effective technology to meet this need.

E-jet is generated when the ink in a nozzle is charged above a critical voltage level due to an applied voltage potential between a conductive nozzle and grounded substrate; the resultant electrostatic force draws ink out of the nozzle to form a Taylor cone, releasing <20 μm diameter ink droplets.¹ E-jet is compatible with a large variety of ink materials such as protein suspensions,² silver nano-particle suspensions,³ and viscous materials.⁴ It is also cost effective for small scale production and provides great adaptability to design changes.

Despite all the merits, e-jet's application has been limited due to print quality sensitivity to standoff height changes between the nozzle and the substrate. Non-flat substrates (>5 μm variation) result in scattered printing, poor printing accuracy and resolution, or the inability to release material. We term the impact of varying the substrate condition during e-jet printing as substrate effects.

Prior arts and strategies to overcome substrate effects were reviewed in Ref. 5; currently, there are no general open loop e-jet printing techniques that have successfully demonstrated high-resolution (<20 μm) printing on a contoured surface.^5–10 

Printing resolution of an e-jet nozzle with an integrated extractor ring, presented in Refs. 6 and 9, cannot be increased after the design has been scaled-down due to several challenges. The key challenge is that a scaled-down printhead does not provide sufficient distance between the nozzle and extractor ring for the released droplets to accelerate. As such, charged ink droplets will not gain enough momentum to overcome the attraction force from the extractor ring. Instead of passing through the grounded extractor ring to the substrate below,^6,9 positively charged droplets will be electrically pulled towards the grounded extractor ring.

To help the released droplets overcome the attraction force from the extractor, we incorporate airflow into an e-jet system to redirect the droplets onto the substrate. The first functional prototype that implemented this concept was published previously in Ref. 10. Extending from the original two nozzle design, we have designed, characterized, and fabricated a three nozzle airflow assisted e-jet printhead. The extension from two to three nozzles was implemented to achieve independent control of the electrohydro and airflow dynamics. Independent control is a key for achieving high-resolution printing with minimized scattering defects. Figure 1 shows a schematic diagram of the modified three nozzle airflow assisted e-jet printhead. The extractor nozzle is a grounded gold sputtered nozzle for creating the electrohydrodynamics. The printing nozzle is a gold sputtered nozzle charged with positive voltage. They are oriented to maximize printing consistency and to minimize droplet scattering. The air jetting nozzle is a non-sputtered glass nozzle used to supply airflow. Electrohydrodynamics generates a Taylor cone between the printing and extractor nozzles, releasing droplets towards the extractor nozzle tip. Downward airflow released from the air jetting nozzle redirects the droplets towards the substrate. Controlling parameters of the printhead are listed in Table I. Applied voltage V can be a constant DC voltage or a pulsed voltage. In this paper, we showcase pulse printing. Patterns printed using constant DC voltage can be found in the supplementary material.¹¹ 

To redirect the charged droplets away from their original trajectory along the electrostatic field, the air jetting nozzle is slightly offset from the direction of the electrostatic field. With this orientation, the airflow distorts rather than stabilizes the outer profile of the Taylor Cone, resulting in higher initiating voltages as compared to co-flow of CO₂ around the Taylor cone in an electrospray process.¹² In airflow assisted e-jet, droplet generation is driven by electrohydrodynamics, while the droplet trajectory is controlled by airflow. This differs from aerodynamically assisted jet printing¹³ and wire arch spray;¹⁴ two processes that use airflow to generate and eject droplets.

The key design parameters are determined using a combination of basic physics and empirical analysis.

Step 1: Determine d_N as a function of the desired resolution of the printed pattern. Utilizing a dripping ejection mode to enable controlled droplet ejection, the printed droplet diameter is approximately 0.5 d_N. We aim to print feature sizes between 1 and 20 μm; thus, we chose $dN=2 μm$⁠.

Step 2: Determine d_E between $30 μm≤dE<dE,bulk$⁠. Experimental testing identified 30 μm as a lower limit for effective extraction. Extractor nozzle tips with $dE<30 μm$ do not dominate the electrohydrodynamics, losing robustness in the electrohydrodynamics across varying standoff heights (see Fig. S1 in supplementary material¹¹). Importantly, the use of a larger extractor nozzle increases the bulk volume of the printhead, hindering the printhead from printing at low standoff heights H. The selection of 30 μm provides a reasonable trade-off between size and electrohydrodynamic stability.

Step 3: Determine $dA$ as the smallest diameter that delivers sufficient airflow for redirecting the ink droplets onto the substrate. We experimentally tested nozzle diameters of ${2,5,10,30}$ μm. From these experiments, we determined that when $dA={2,5}μm$ the airflow was too small to redirect the droplets, while a $dA=30 μm$ was very sensitive to changes in air pressure. A nozzle diameter of 10 μm was able to redirect the droplets with minimal scattering and material displacement. See Fig. S2 in supplementary material for further details.¹¹ 

Step 4: Determine tilting angle α for the printing nozzle. Figure 2 depicts the electrostatic field of an airflow assisted e-jet printhead from COMSOL 4.4 (model details in supplementary material).¹¹ The electrostatic field generated by the printing and extractor nozzles tilted at angles of α and β creates a strong horizon force component upon the ink, preventing droplets from landing vertically, and leading to droplet scattering (droplet trajectory 2). To mitigate this effect, angle α should be minimized, while avoiding physical contact between the air jetting nozzle (labeled airflow in Fig. 2) and the printing nozzle. Smaller angles combined with airflow will generate droplet trajectory 1. For our design, $α=30°$ is the smallest angle that avoids physical contact between the nozzles.

Step 5: Determine the tilting angle $β$ for the extractor nozzle. Similar to α, the angle should be minimized, while avoiding physical contact between the extractor and air jetting nozzles. A printing nozzle with a large angle α scatters the droplets by repelling the ink drops sideways (Fig. 2); correspondingly, an extractor nozzle with a large angle $β$ scatters the droplets by attracting the droplets with a horizontal potential field. For our design, we select $β=30°$⁠.

Step 6: Determine the tip distance between the printing and extractor nozzles $S$⁠. Experimental testing indicated that a distance of S < 10 d_N requires a strong air jet to direct the droplets away from the extractor nozzle, leading to droplet scattering. Conversely, a distance of S > 10 d_N μm leads to significant electrohydrodynamic interference from the substrate. As such, we determined that the ideal value for our design was S = 10 d_N. The tip of the extractor nozzle should be placed slightly below the printing nozzle so that the meniscus can form properly at the printing nozzle tip.

Step 7: Determine the standoff height h between the air jetting nozzle and the printing nozzle based on fluid dynamics. According to fluid models in literature,¹⁵ subsonic airflow released from a nozzle generates a potential core region at the nozzle opening, in which the airflow velocity is equivalent to the velocity at the nozzle exit. Our experimental results suggest that disruption of the potential core leads to droplet scattering; therefore, we place the printing nozzle tip at point B Fig. 3. According to the literature, potential core length can be as long as $6⋅dA$⁠, hence point B should be located outside this region h > 8d_A.

Step 8: Determine air pressure (P_A) supply for the air jetting nozzle. The optimum level of P_A is the lowest air pressure sufficient to redirect the ink droplets away from the extractor nozzle and towards the substrate. High P_A values lead to scattering or material displacement on the substrate. To determine P_A, start with 0 psi and increase until droplets are redirected onto the substrate. P_A values can vary between 3 and 15 psi depending on the nozzle sizes and ink materials used.

To demonstrate the ability of the airflow assisted printhead to decrease substrate sensitivity, we compare the printing performance using Norland Optical Adhesive (NOA) 81 as our ink for three cases:

1. E-jet printing (traditional): Positively charged conductive printing nozzle with $dN=2 μm$⁠; conductive grounded substrate (gold sputtered glass slide).

2. E-jet printing (traditional): Positively charged conductive printing nozzles with $dN=2 μm$⁠; non-conductive substrate (glass slide) with no grounded plate underneath.

3. Airflow assisted e-jet printing: Positively charged conductive printing nozzle with $dN=2 μm$⁠; non-conductive substrate (glass slide); grounded extractor nozzle with $dE=30 μm$⁠, air jetting nozzle with $dA=10 μm$⁠; S = 20 μm, h = 100 μm; α = β = 30° from vertical; and experimentally determined P_A.

For each case, we determined the initiating voltages V_o at varying standoff heights: $H={30,60,100,...,1200}μm$⁠. The experiment was repeated 3 times for each case, with data averaged across the 3 trials and plotted in Fig. 4. All printing experiments were conducted within temperature range of 19–22 °C, and humidity level of 20%–60%. Since NOA 81 is a single component liquid polymer, humidity changes do not significantly disrupt the printing dynamics.

The curves for cases 1 and 2 follow the trend predicted from Eqs. (1) and (2);¹⁶ the initiating voltage required to overcome capillary action increases as a function of increased standoff height:

In Eq. (1), $Q$ is the ink releasing rate, $dN$ is the nozzle diameter, $μ$ is the viscosity of the liquid, $L$ is the length of the nozzle, $ΔP$ is the pressure difference between the ink chamber and the ambient air, $εo$ is permittivity of free space, $E$ is the electrostatic field intensity, and $γ$ is the surface tension of the air-ink interface.

Equation (1) shows that ink will be released (Q > 0) if the ink-air pressure difference and the electrostatic field surpass the capillary action (⁠$4γ/dN$⁠). Since the viscosity of NOA 81 is relatively low, e-jet printing can be initiated without applied pressure (ΔP = 0); thus, when E = E_o = $8γ/εodN$⁠, printing will be initiated.

Equation (2) relates initiating electrostatic field intensity $Eo$ to nozzle diameter, standoff height, and applied initiating voltage $Vo$⁠. Since $Eo$ and $dN$ are constants, Eq. (2) predicts that $Vo$ increases as $H$ increases, matching the curves for cases 1 and 2 in Fig. 4. The curve for case 3 demonstrates the effect of the airflow dynamics. Standoff heights greater than 400 μm show a small, yet steady increment in initiating printing voltage; an unexpected result if one was to simply apply the electrohydrodynamic properties defined in Eqs. (1) and (2). The printing voltage increase across a change of 1170 μm in standoff height for the airflow assisted printhead is 21%, which is significantly lower than the 101% increase for the conventional e-jet setups on both conductive and non-conductive substrates. This indicates the improved robustness of the airflow assisted printhead to variations in standoff height.

Figure 5 presents printed patterns of airflow assisted e-jet printing of NOA 81 on non-conductive surfaces tilted at 25°, 45°, and 65°. The printing conditions for the dot matrices are listed in the table within Fig. 5 and the printing environmental factors are the same as previously described. Note that the same peak voltage (415 V) was used for all three tilted surfaces. All patterns were printed with a stationary printhead. The substrate jogging rate was 50 μm/s, a slow rate chosen for demonstration purposes only. The top row of each matrix was printed with H = 400 μm; note that 415 V would not initialize traditional e-jet printing with standoff heights of H ≥ 300 μm. Due to the tilted angle, the standoff height changes at each row for a total standoff height variation from 400 to 1300 μm for the 65° tilted surface.

For comparison, dot matrices of NOA 81were printed using traditional e-jet printing onto non-conductive glass surfaces tilted at 25°, 45°, and 65° (Fig. 6). Starting the top row at a standoff height of 400 μm, note that the initiating voltages required to print across the surfaces are much higher than 415 V. Comparing Figs. 5 and 6, one can clearly identify the improvement in printing consistency due to the airflow assisted printhead. Traditional e-jet fails to achieve consistent printing results, ceasing to print at all after the first 5 rows due to insufficient applied voltage for the 45° and 65° surfaces. Additionally, more scattering is observed with the traditional e-jet printed results.

Finally, to validate droplet size and placement control, airflow assisted e-jet printing was printed using the pulse printing technique from Ref. 17. Figure 7 shows average droplet diameter on a 65° tilted non-conductive surface reduced from 18.9 ± 0.4 μm to 10.9 ± 0.4 μm by shortening the pulse width from 100 ms to 20 ms. Ink droplet size was not affected by the standoff height variation of 214 μm.

In conclusion, airflow assisted e-jet printing can achieve consistent high-resolution (<20 μm) printing across large changes in standoff height. This open-loop consistency surpasses conventional e-jet printing using either conductive or non-conductive surfaces. The introduction of airflow into the e-jet process provides the necessary mechanism for droplet redirection onto the substrate, while enabling controlled droplet release.

The research for this paper was financially supported by National Science Foundation CAREER Award No. 1351469.