Optimization and Characterization of Drop-on-demand Inkjet Printing ...

Optimization and Characterization of Drop-on-demand Inkjet Printing Process for Platinum Organometallic Inks

Gerard Cummins1, Robert Kay1, Jonathan Terry2, Marc P.Y. Desmulliez1 and Anthony J. Walton2 1 Institute of Integrated Systems, Microsystems Engineering Centre (MISEC), School of Engineering and Physical Sciences, Heriot Watt University, Edinburgh, United Kingdom, EH14 4AS 2 Institute of Integrated Systems, School of Engineering, University of Edinburgh, Edinburgh, United Kingdom, EH9 3JF Email: [email protected] Abstract Inkjet printing has been extensively used over the past 30 years in the graphic arts and packaging industries. This technology involves dispensing accurately positioned droplets of ink onto a substrate, which then solidifies through the evaporation of the constituent solvent, the cross-linking of a polymer or through crystallization. The mask-less, flexible, rapid and low cost nature of inkjet printing, combined with the development of a range of functional inks, has led to the adoption of this technology in system manufacturing. The SMART Microsystems research project underway at the Institute for Integrated Systems is investigating the use of this technique in the rapid customization of CMOS foundry wafers for More-than-Moore applications. This paper presents results obtained during the development and optimization of a drop-on-demand inkjet printing process for initial batches of platinum organometallic inks. Drop-on-Demand (DOD) inkjet printing works by inducing a transient pressure pulse in the ink reservoir through electrical excitation of either a thermal or piezoelectric element. The correct implementation of this excitation signal is necessary to produce a pressure pulse capable of reproducibly and reliably generating a series of droplets. The effects of system parameters on the formation of these droplets are investigated. Methods used to characterize droplet ejection are also described. Introduction Inkjet printing has been widely used in the graphics industry for many years. However, the advantages of cheap equipment, direct writing, low material wastage and increasing availability of functional inks has made this deposition method increasingly attractive for use in manufacturing applications[1]. The deposition of functional materials by inkjet printing for such applications requires not only that the formulation of inks meet the viscosity, surface tension and density requirements similar to standard inkjet inks but also have the desired functional properties such as electrical conductivity[2]. These requirements are usually met by using nanoparticle inks, which consist of metallic nanoparticles suspended in an organic solvent. The metallic nanoparticles can lead to blocking of the print-head nozzles if suitable agents are not added to the ink to prevent precipitation and agglomeration. These stabilizing agents can prevent the formation of connections between adjacent metallic nanoparticles by the creation of an intermediate insulating layer. The removal of this insulating layer and formation of a continuous electrical

path through the necking and eventual fusing of nanoparticles is achieved through thermal annealing. An organometallic compound is one that contains at least one metal-carbon bond. Organometallic materials are already used in the semiconductor industry to repair photolithography masks and integrated circuits using focused ion beam deposition[3][4]. Other methods that have been reported for depositing organometallic material include direct laser writing and e-beam writing. The non-particulate nature of organometallic ink minimizes the blocking of nozzles and can produce a void free film. Inkjet printing relies on the generation of sequences of droplets. This can be achieved either by continuous (CIJ) or drop-on-demand (DOD) inkjet. CIJ is not commonly used in printing for manufacturing as many functional inks can be affected by exposure to the environment. The DOD inkjet Printing process consists of five stages, which are drop ejection, drop flight, drop impact, drop spreading and drop evaporation[5]. The primary focus of this paper is drop ejection, which involves the generation and breaking away of drops at the print head. Droplet ejection occurs by inducing a transient pressure pulse in the ink reservoir by electrical excitation of either a thermal or piezoelectric element. The correct implementation of this excitation signal can produce a pressure pulse capable of forcing a small volume of ink out of the nozzle to reproducibly generate a series of droplets[1][2]. The effects of these system parameters on the ejection of droplets of a novel organometallic ink are investigated in this paper. Piezoelectric Drop on Demand The voltage waveform applied to the piezoelectric element is shown in Fig. 1 and consists of four stages. Each stage can be described by a slew rate, duration and a magnitude. The voltage is reduced during the first stage to relax the piezoelectric element. This increases the volume of the channel to create a lower pressure, which draws fluid from the reservoir into the channel.

Fig. 1 Amplitude of ejection waveform as a function of time The voltage is held constant while the fluid is being drawn in. The voltage is increased during the second stage of the 2011 13th Electronics Packaging Technology Conference

978-1-4577-1982-0/11/$26.00 ©2011 IEEE 256

waveform, which causes the PZT bimorph to compress. This results in a sudden reduction in the volume of the channel thereby generating a transient pressure wave. This pressure wave pushes fluid out of the channel nozzle to form a droplet. The waveform voltage is reduced in the remaining stages. As a result the channel slightly expands causing the droplet to neck and eventually separate from the nozzle. Finally the piezoelectric element is returned to its initial slightly deflected stage prior to the next waveform. Characterization of Droplet Ejection The performance of the droplet ejection process can be characterized by responses such as the droplet velocity, volume, consistency, shape and directionality. Droplets jetted from the nozzle should have a velocity ranging between 1 and 30 ms-1 depending on the fluid properties of the ink [2][6]. A lower velocity can result in increased dot positioning errors or insufficient kinetic energy to break free from the print head, while too high a velocity can result in splashing of the droplet upon impact with the substrate. The desired size of the droplets depends upon the application. Larger droplet sizes result in quicker printing for large area applications, whereas small droplet sizes are ideal for applications where high resolutions and small dimensions are needed. Droplet size is commonly measured using either gravimetric[7] or optical methods[8]. The droplet velocity and volume should stay within a certain band of values to minimize irregularities in the printed features. Print quality can be negatively affected if the droplet includes satellite droplets or long tails. Ideally the shape of the droplet should be spherical. Finally, to ensure that droplet positioning is accurate and reproducible it is required that the path of the droplets should be perpendicular to the substrate. Material Characterization Behaviour of the ink during droplet ejection is dependent on material properties such as density, viscosity and surface tension. It is necessary to measure these properties to better understand how suitable these fluids are for inkjet printing. This is often achieved through the use of dimensionless numbers such as the Ohnesorge number as the fluid mechanics and physics behind DOD inkjet printing are complex. The Ohnesorge (Oh) number is the ratio between the viscous forces and inertial and surface tension forces. To calculate the Ohnesorge number the viscosity (η), density (ρ) and surface tension (γ) of the ink were measured. The characteristic diameter (L) was taken as being 21.5 μm, which is the diameter of a print-head nozzle.

Oh =

η ργ L

(1)

Table I Fluid properties of ink Z Ink η (mPa.s) γ (mNm-1) ρ (kgm-3) Oh Pt 12.9 31.02 1300 0.283 3.53 Fromm[9], Derby[2] and others have shown that the mechanisms of drop formation can be characterized by the reciprocal of the Ohnesorge number (Z). Stable drop formation can be expected from fluids when 10 > Z > 1.The viscosity of the ink was measured at room temperature using an AR 2000 rheometer with shear rates between 200 and 20 s1 . The ink exhibited Newtonian properties over this range. The

surface tension was measured using the pendant drop method with a FTA 200 with FTA32 image processing software. The density was calculated from mass and volume measurements. The measurements obtained are shown in Table I. The Ohnesorge number of this ink is 0.283, giving a reciprocal of 3.53. This is within the range of values described by Fromm as being suitable for use in inkjet printing. Design of Experiment A fractional factorial design was used to characterize and optimize droplet ejection for this ink. Fourteen factors were identified that could affect the droplet area (DA), velocity (DV), shape (DS) or stability (DStab) of the ejection process. These factors included the ink temperature (CT), jetting frequency (JF), and slew rate (SR), duration (VD) and amplitude (VA) of each of the four regions of the ejection waveform. A 14 factor, 2 level fractional factorial design was used, which resulted in 32 experimental runs. The order in which the experiments were run was randomized to reduce the influence of any external factors. The experimental design matrix is shown in Table II Experimental Method The ink was degassed before printing by placing it in a Bransonic B3 ultrasonic cleaner for two hours. Afterwards the ink is injected into the fluid module of a DMC-11610 10pl piezoelectric drop-on-demand print-head cartridge. Droplet ejection and flight are recorded using the integrated stroboscopic drop-watching camera of the Fujifilm Dimatix DMP 2800 printer for each of the 32 experimental runs. The print-head nozzles are purged for 0.5 μs between each experimental run to ensure that the nozzles are wetted, fresh ink is used in all runs and to minimize any “first-drop” problems[10] [11]. Sequential frames are taken from the recorded movies. These images are analyzed using the MATLAB image processing toolbox. First the RGB images are cropped to only show droplets in flight and not droplet formation at the nozzles. These images are converted to gray scale and the background subtracted to compensate for non-uniform illumination. Next the complement of the image is generated and the contrast adjusted before conversion to a binary image. Morphological erosion and dilation is used to remove any remaining noise from the image. Additional morphological analysis is used to detect the boundaries of each drop in the image. The size, shape and stability of each drop are calculated. The number of pixels within the drop boundary are counted and multiplied by the pixel resolution of the image to determine the cross-sectional area. The roundness of the drop shape is calculated using the figure of merit Rshape, which is defined as the ratio of the square of the effective perimeter of the droplet to the cross-sectional area. A perfect circle would be signified by a value of 1. 4π (Area) RShape = (2) (Perimeter)2 One requirement for a process to be considered stable is that the size of the droplets be consistent in size. The standard deviation of the droplet area over time is used as an indicator of the stability of droplet ejection induced by the waveform. 2011 13th Electronics Packaging Technology Conference

257

The velocity is determined manually by measuring the presence of instabilities in several of the ejection recipes in the displacement of a droplet over a number of image frames. matrix and variation in behavior between nozzles. Droplet velocity from one nozzle was measured due to the Table II Experimental matrix used to optimise the droplet ejection process E X P

CT (OC)

JF (kHz)

S R 1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50 30 50

5 5 40 40 5 5 40 40 5 5 40 40 5 5 40 40 5 5 40 40 5 5 40 40 5 5 40 40 5 5 40 40

0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5

V A 1 (V) 0 0 0 0 0 0 0 0 2 2 2 2 2 2 2 2 0 0 0 0 0 0 0 0 2 2 2 2 2 2 2 2

V D 1 (μs) 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 3 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10

S R 2 0.5 1.5 1.5 0.5 1.5 0.5 0.5 1.5 0.5 1.5 1.5 0.5 1.5 0.5 0.5 1.5 0.5 1.5 1.5 0.5 1.5 0.5 0.5 1.5 0.5 1.5 1.5 0.5 1.5 0.5 0.5 1.5

V D 2 (μs) 2 3.5 3.5 2 2 3.5 3.5 2 3.5 2 2 3.5 3.5 2 2 3.5 2 3.5 3.5 2 2 3.5 3.5 2 3.5 2 2 3.5 3.5 2 2 3.5

V A 2 (V) 20 40 20 40 40 20 40 20 40 20 40 20 20 40 20 40 20 40 20 40 40 20 40 20 40 20 40 20 20 40 20 40

S R 3 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 1.5 1.5 0.5 0.5 0.5 0.5 1.5 1.5 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 1.5 1.5 0.5 0.5 0.5 0.5 1.5 1.5

V D 3 (μs) 2.5 7.5 7.5 2.5 2.5 7.5 7.5 2.5 2.5 7.5 7.5 2.5 2.5 7.5 7.5 2.5 7.5 2.5 2.5 7.5 7.5 2.5 2.5 7.5 7.5 2.5 2.5 7.5 7.5 2.5 2.5 7.5

V A 3 (V) 10 18 10 18 18 10 18 10 10 18 10 18 18 10 18 10 18 10 18 10 10 18 10 18 18 10 18 10 10 18 10 18

S R 4 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 0.5 0.5 1.5 1.5 1.5 1.5 0.5 0.5 1.5 1.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.5 1.5 1.5 0.5 1.5 0.5 0.5

V D 4 (μs) 0.5 4 0.5 4 0.5 4 0.5 4 4 0.5 4 0.5 4 0.5 4 0.5 4 0.5 4 0.5 4 0.5 0.5 4 4 0.5 0.5 4 0.5 0.5 4 4

V A 4 (V) 4 4 8 8 4 4 8 8 8 8 4 4 8 8 4 4 8 8 4 4 8 8 4 4 4 4 8 8 4 4 8 8

D A (μm2)

D S

D Stab

D V (ms-1)

0.00 18.93 105.23 10.46 112.91 89.21 16.82 87.46 1.99 264.55 6.22 122.28 30.32 406.42 0.00 47.87 1.03 28.80 220.07 371.60 20.06 79.66 127.10 179.89 26.88 33.44 3.66 40.42 7.41 52.23 4.11 85.72

0.00 0.17 0.59 0.35 0.55 0.80 0.52 0.79 0.36 0.83 0.75 0.67 0.30 0.72 0.60 0.52 0.36 0.37 0.60 0.58 0.41 0.42 0.55 0.86 0.80 0.69 0.65 0.58 0.5 .52 0.5 0.0

0.00 48.60 121.25 90.40 119.28 62.24 42.29 69.99 29.12 306.16 35.02 140.81 66.62 560.06 0.00 296.3 8.82 158.58 259.15 1316.5 87.62 211.32 138.05 129.12 57.39 246.17 27.38 87.08 33.31 254.68 17.58 73.42

0.25 0.00 0.40 2.33 0.44 6.00 0.33 0.30 0.89 0.63 0.40 0.30 0.29 0.38 0.00 6.00 0.11 0.10 0.20 33.0 038 0.43 0.33 0.20 0.25 0.40 0.43 25.0 0.13 0.33 0.00 0.01

Fig. 2 Main Responses for Droplet Area

2011 13th Electronics Packaging Technology Conference

258

Fig. 3 Main Responses for Droplet Shape

Fig. 4 Main Responses for Droplet Velocity

Fig. 5 Main Responses for Standard Deviation of Droplet Area The main effects for droplet area are shown in Fig. 2. The Results slew rate of the third stage of the waveform has a greater These measurements are inserted into the experimental effect than those of the other stages on the droplet area. This matrix to find the sensitivity of the responses to each of the 14 may be explained by taking into account that the primary factors. Increasing temperature is shown to increase all responses due to the associated changes of the fluid properties purpose of this stage of the waveform is to induce necking and eventually separation of droplets from the nozzle. A higher of the ink. slew rate may generate a transient pressure pulse of sufficient magnitude required to cause larger droplets to break away. 2011 13th Electronics Packaging Technology Conference

259

The effect of amplitude on droplet area can also be seen from this graph. An amplitude of 0V for stage 1 results in a larger droplet area as the piezoelectric element is relaxed, which increases the volume of the reservoir. A larger volume of ink in the reservoir means that a larger volume of ink is present for ejection. Likewise a larger amplitude for stage 2 results in an increase in droplet area due to the larger force produced. The main effects for droplet shape are shown in Fig. 3. This graph shows that changing slew rates of stage 1 and 3 have a negligible effect on the roundness of the droplet. The droplet roundness is sensitive to changes in the slew rates of the other stages. Increasing the duration of stages 1, 2 and 3 are also shown to result in an increased roundness of any droplet. These stages are involved with droplet formation and breaking away. The effect of these factors on the droplet velocity is shown in Fig. 4. The velocity is demonstrated to be most sensitive to the amplitude of the third stage compared to those of the other stages. The breaking away of the droplet from the nozzle is induced by this stage of the waveform and imparts momentum to the droplet at this point. An increase in amplitude of second stage results in an increase in droplet velocity as expected. Low slew-rates for the first and second stage increase the droplet velocity. The effect of these factors on the standard deviation of the droplet area is shown in Fig. 5. It can be seen from this graph that the stability of the ejection process is sensitive to changes in all factors, which is known from previous experience. The relatively low viscosity of the ink makes it more sensitive to transients induced by switching between waveform amplitudes or increased jetting frequency due to the poor damping of these materials. Analysis of these results enabled the development of a stable jetting waveform. The parameters for each of the four stages of the waveform are shown in Table III. The ink is heated to 50OC and jetted at a frequency of 5kHz. The area and shape factor of the resulting droplets were measured to be 304.2 μm2 and 1.06 respectively. The velocity of the droplets was set at 7 ms-1 for all nozzles by adjusting the applied voltage. An example of the droplet ejection achieved with this waveform is shown in Fig. 6

Fig. 6 Example of droplet ejection achieved with jetting waveform Drops produced with the optimized waveform were approximately 43 μm in diameter on a silicon oxide substrate at 50OC. The morphology of any printed features is affected by factors such as the contact angle between the ink and substrate, the print-head velocity, substrate temperature and

drop pitch. Stable, straight lines were printed with this ink using a drop pitch of 37 μm. Three passes of the print-head were used to print Greek cross structures for the measurement of the sheet resistance of annealed films [12][13]. Multiple passes of the print-head were necessary as a single pass produced a film with an average height of 9 nm and the low metal content of the ink (2.31% of total mass). An example of the Greek cross structures printed is shown in Fig. 7. The average height of the resulting film was approximately 90 nm. Next the Greek cross structures were annealed for 2 minutes at 350OC. The sheet resistance was measured to be 3.98E+05 Ω/ using a HP 4156B semiconductor parameter analyzer.

Slew Rate Duration (μs) Amplitude (V)

Stage 1 1 2 0

Stage 2 1.25 2 20

Stage 3 1 2 8

Stage 4 1.5 0.5 4

Table III Parameters of the stable jetting waveform

Fig. 7 Printed Greek cross structure before annealing Conclusions An experimental matrix was designed for the characterization of an excitation signal used in the inkjet printing of a novel platinum organometallic ink. The results of these experiments were used to produce a waveform that has been shown to be able to reliably and reproducibly produce round droplets with a cross sectional area of 304.2 μm2. The inkjet printing process was used to print Greek cross structures, which provided initial sheet resistance measurements of this material. Acknowledgements The authors wish to thank Michelle Harvie and Jim Thomson at Ceimig Ltd. (Dundee, UK). The authors would also like to acknowledge the financial support of the Innovative electronic Manufacturing Research Centre (IeMRC) through the flagship project Smart Microsystems.

References [1] K. Hon, L. Li, and I. Hutchings, “Direct writing technology—Advances and developments,” CIRP Annals Manufacturing Technology, vol. 57, no. 2, pp. 601-620, 2008.


260

[2] B. Derby, “Inkjet Printing of Functional and Structural Materials: Fluid Property Requirements, Feature Stability, and Resolution,” Annual Review of Materials Research, vol. 40, no. 1, pp. 395-414, 2010. [3] T. Tao, J. Ro, J. Melngailis, Z. Xue, and H. D. Kaesz, “Focused ion beam induced deposition of platinum,” Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures, vol. 8, no. 6, pp. 1826-1829, Nov. 1990. [4] T. Tao, W. Wilkinson, and J. Melngailis, “Focused ion beam induced deposition of platinum for repair processes,” Journal of Vacuum Science & Technology B: Microelectronics and Nanometer Structures, vol. 9, no. 1, pp. 162-164, Jan. 1991. [5] M. Singh, H. M. Haverinen, P. Dhagat, and G. E. Jabbour, “Inkjet Printing—Process and Its Applications,” Adv. Mater., vol. 22, no. 6, pp. 673-685, 2010. [6] A. L. Yarin, “Drop Impact Dynamics: Splashing, Receding, Bouncing,” Annual Review of Fluid Mechanics, vol. 38, no. 1, pp. 159-192, 2006. [7] R. M. Verkouteren and J. R. Verkouteren, “Inkjet Metrology: High-Accuracy Mass Measurements of Microdroplets Produced by a Drop-on-Demand Dispenser,” Anal. Chem., vol. 81, no. 20, pp. 8577-8584, Sep. 2009. [8] H. Dong, W. W. Carr, and J. F. Morris, “Visualization of drop-on-demand inkjet: Drop formation and deposition,” Review of Scientific Instruments, vol. 77, no. 8, 2006. [9] J. E. Fromm, “Numerical Calculation of the Fluid Dynamics of Drop-on-Demand Jets,” IBM Journal of Research and Development, vol. 28, no. 3, p. 322, 1984. [10] A. Famili, S. A. Palkar, and J. W. J. Baldy, “First drop dissimilarity in drop-on-demand inkjet devices,” Physics of Fluids, vol. 23, no. 1, 2011. [11] H. R. Kang, “Water-based ink-jet ink III: Performance studies,” Journal of Imaging Science and Technology, vol. 35, no. 3, pp. 195-201, 1991. [12] S. Enderling et al., “Sheet Resistance Measurement of Non-Standard Cleanroom Materials Using Suspended Greek Cross Test Structures,” IEEE Transactions on Semiconductor Manufacturing, vol. 19, no. 1, pp. 2-9, Feb. 2006. [13] D. K. Schroder, Semiconductor material and device characterization. John Wiley & Sons, 2006.


261