Programmable Droplet Microfluidics Based on Machine Learning and Acoustic Manipulation

Abstract

Typical microfluidic devices are application-specific and have to be carefully designed to implement the necessary functionalities for the targeted application. Programmable microfluidic chips try to overcome this by offering reconfigurable functionalities, allowing the same chip to be used in multiple different applications. In this work, we demonstrate a programmable microfluidic chip for the two-dimensional manipulation of droplets, based on ultrasonic bulk acoustic waves and a closed-loop machine-learning-based control algorithm. The algorithm has no prior knowledge of the acoustic fields but learns to control the droplets on the fly. The manipulation is based on switching the frequency of a single ultrasonic transducer. Using this method, we demonstrate 2D transportation and merging of water droplets in oil and oil droplets in water, and we performed the chemistry that underlies the basis of a colorimetric glucose assay. We show that we can manipulate drops with volumes ranging from ∼200 pL up to ∼30 nL with our setup. We also demonstrate that our method is robust, by changing the system parameters and showing that the machine learning algorithm can still complete the manipulation tasks. In short, our method uses ultrasonics to flexibly manipulate droplets, enabling programmable droplet microfluidic devices.

Acoustic Manipulation of Particles in Microfluidic Chips with an Adaptive Controller that Models Acoustic Fields

“…Compared to the ε ‐greedy control method we have reported previously, [ 1,9 ] the new controller shows far better path following precision and the manipulations are faster (e.g., ≈10 min for AMA and ≈108 min for ϵ‐greedy). The poor precision and long manipulation times were the bottlenecks of applying the previous controller in practical applications of droplet microfluidics and particle sorting.…”

“…The original frequencies were 100 frequencies linearly spaced between 65 and 700 kHz and the new frequencies were 100 frequencies linearly spaced between 71.5 and 770 kHz, but as an example, the controller still uses the model developed for 65 for the 71.5 kHz. The 10% frequency shift is enough to make all the acoustic field models meaningless [ 9 ] and mimics sudden changes in the chamber, such as a bubble entering the chamber or a large change in temperature. 2) By physically changing the chip to a completely new one, this time with three outlets instead of one.…”

“…The particles were pumped into the chamber with a syringe pump (Aladdin, World Precision Instruments). For the manipulation experiments with droplets, we used the flow‐focusing device from our previous work [ 9 ] with 2.5 vol% sorbitan monooleate (Span 80, Sigma‐Aldrich) in hexadecane oil (Fisher Chemicals) as the continuous phase and blue‐dyed deionized water as the dispersed phase (droplets). The flow rates were = 0.5 μL min −1 and = 50 μL min −1 .…”

“…In our previous works, we reported reinforcement learning methods for manipulating particles [ 1 ] and droplets [ 9 ] inside closed‐channel microfluidic chips with bulk acoustic waves. The methods were derived from the family of multiarmed bandit algorithms; specifically, the UCB‐1 and ε‐greedy algorithms.…”

“…

Acoustic microfluidics-acoustofluidics-is a rapidly growing field, where an acoustic source creates ultrasonic waves that are used to contactlessly manipulate particles, [1][2][3][4] cells, [5][6][7][8] or droplets [4,9,10] within microfluidic chips. [11] Many different applications have been demonstrated, including particle sorting, [1,12] blood fractionation, [6,13] cell sorting, [6,14,15] cell patterning, [8,[16][17][18][19][20] and droplet microreactors, [9,17,21] among others.The contactless nature of the acoustic force simplifies the fabrication of the chip. In the case of bulk acoustic waves, the chip itself can be a passive glass device, with the ultrasonic waves produced by a piezoelectric transducer attached to the chip.

…”