Automated tuning of high-order waveforms for picoliter resolution jetting of rheologically challenging materials

Abstract Drop on demand piezoelectric jetting has become an important tool for direct patterning as well as adaptive material deposition in micro- and nano-fabrication applications. Key performance and reliability metrics for jetting include lowest dispense volume, mean volume accuracy, volume variance, and drop position accuracy. Using physics-based models that accurately describe drop formation is challenging as these models require several jet parameters that can vary due to manufacturing tolerances from one jet to the next, and are unavailable non-destructively. Such manufacturing variations are particularly important when pushing the limits of volume resolution while achieving other performance and reliability targets. Current practice involves ad-hoc manual recalibration of the actuation waveform to obtain reliable jetting of a variety of materials, which prevents the use of waveforms defined by large numbers of parameters. In order to achieve the above mentioned metrics, a promising approach is to explore high order waveforms; however, such waveforms are unsuited for ad-hoc manual techniques. Here, we propose an automatic waveform tuning method that: (i) circumvents the need for complex forward models by estimating drop volume and velocity measurements via computer vision; (ii) uses genetic algorithm based stochastic optimization to tune the waveform; (iii) using a real-time sensing based on image processing that has been benchmarked against an International Electrotechnical Commission (IEC) standard for imaging-based measurement of droplet volume and (iv) enables exploration of high order waveforms (defined by up to 10 parameters) to maximize the performance/reliability of a given combination of jetting system and material. The experimental results presented in this paper show that with such methods, fluids considered rheologically impossible to jet in the literature can be jetted reliably with impressive performance metrics such as drop volumes of 1.1 pL achieved for ethyl acetate. The drop radii can also be reduced to as low as 25.6% of the jet aperture radius while achieving drop velocity sufficient to maintain position accuracy for ethyl acetate.