Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency
The present work describes the phenomenological approach to automatically determine the frequency range for positive and negative dielectrophoresis (DEP)—an electrokinetic force that can be used for massively parallel micro- and nano-assembly. An experimental setup consists of the microfabricated ch...
Main Authors: | , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
MDPI AG
2022-02-01
|
Series: | Micromachines |
Subjects: | |
Online Access: | https://www.mdpi.com/2072-666X/13/3/399 |
_version_ | 1797444846724579328 |
---|---|
author | Matthew Michaels Shih-Yuan Yu Tuo Zhou Fangzhou Du Mohammad Abdullah Al Faruque Lawrence Kulinsky |
author_facet | Matthew Michaels Shih-Yuan Yu Tuo Zhou Fangzhou Du Mohammad Abdullah Al Faruque Lawrence Kulinsky |
author_sort | Matthew Michaels |
collection | DOAJ |
description | The present work describes the phenomenological approach to automatically determine the frequency range for positive and negative dielectrophoresis (DEP)—an electrokinetic force that can be used for massively parallel micro- and nano-assembly. An experimental setup consists of the microfabricated chip with gold microelectrode array connected to a function generator capable of digitally controlling an AC signal of 1 V (peak-to-peak) and of various frequencies in the range between 10 kHz and 1 MHz. The suspension of latex microbeads (3-μm diameter) is either attracted or repelled from the microelectrodes under the influence of DEP force as a function of the applied frequency. The video of the bead movement is captured via a digital camera attached to the microscope. The OpenCV software package is used to digitally analyze the images and identify the beads. Positions of the identified beads are compared for successive frames via Artificial Intelligence (AI) algorithm that determines the cloud behavior of the microbeads and algorithmically determines if the beads experience attraction or repulsion from the electrodes. Based on the determined behavior of the beads, algorithm will either increase or decrease the applied frequency and implement the digital command of the function generator that is controlled by the computer. Thus, the operation of the study platform is fully automated. The AI-guided platform has determined that positive DEP (pDEP) is active below 500 kHz frequency, negative DEP (nDEP) is evidenced above 1 MHz frequency and the crossover frequency is between 500 kHz and 1 MHz. These results are in line with previously published experimentally determined frequency-dependent DEP behavior of the latex microbeads. The phenomenological approach assisted by live AI-guided feedback loop described in the present study will assist the active manipulation of the system towards the desired phenomenological outcome such as, for example, collection of the particles at the electrodes, even if, due to the complexity and plurality of the interactive forces, model-based predictions are not available. |
first_indexed | 2024-03-09T13:17:22Z |
format | Article |
id | doaj.art-5471acbe83ae4470af15613bd89daffe |
institution | Directory Open Access Journal |
issn | 2072-666X |
language | English |
last_indexed | 2024-03-09T13:17:22Z |
publishDate | 2022-02-01 |
publisher | MDPI AG |
record_format | Article |
series | Micromachines |
spelling | doaj.art-5471acbe83ae4470af15613bd89daffe2023-11-30T21:33:37ZengMDPI AGMicromachines2072-666X2022-02-0113339910.3390/mi13030399Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied FrequencyMatthew Michaels0Shih-Yuan Yu1Tuo Zhou2Fangzhou Du3Mohammad Abdullah Al Faruque4Lawrence Kulinsky5Department of Mechanical and Aerospace Engineering, University of California Irvine, 5200 Engineering Hall, Irvine, CA 92627-2700, USADepartment of Electrical Engineering and Computer Science, University of California Irvine, 2200 Engineering Hall, Irvine, CA 92627-2700, USADepartment of Mechanical and Aerospace Engineering, University of California Irvine, 5200 Engineering Hall, Irvine, CA 92627-2700, USADepartment of Electrical Engineering and Computer Science, University of California Irvine, 2200 Engineering Hall, Irvine, CA 92627-2700, USADepartment of Mechanical and Aerospace Engineering, University of California Irvine, 5200 Engineering Hall, Irvine, CA 92627-2700, USADepartment of Mechanical and Aerospace Engineering, University of California Irvine, 5200 Engineering Hall, Irvine, CA 92627-2700, USAThe present work describes the phenomenological approach to automatically determine the frequency range for positive and negative dielectrophoresis (DEP)—an electrokinetic force that can be used for massively parallel micro- and nano-assembly. An experimental setup consists of the microfabricated chip with gold microelectrode array connected to a function generator capable of digitally controlling an AC signal of 1 V (peak-to-peak) and of various frequencies in the range between 10 kHz and 1 MHz. The suspension of latex microbeads (3-μm diameter) is either attracted or repelled from the microelectrodes under the influence of DEP force as a function of the applied frequency. The video of the bead movement is captured via a digital camera attached to the microscope. The OpenCV software package is used to digitally analyze the images and identify the beads. Positions of the identified beads are compared for successive frames via Artificial Intelligence (AI) algorithm that determines the cloud behavior of the microbeads and algorithmically determines if the beads experience attraction or repulsion from the electrodes. Based on the determined behavior of the beads, algorithm will either increase or decrease the applied frequency and implement the digital command of the function generator that is controlled by the computer. Thus, the operation of the study platform is fully automated. The AI-guided platform has determined that positive DEP (pDEP) is active below 500 kHz frequency, negative DEP (nDEP) is evidenced above 1 MHz frequency and the crossover frequency is between 500 kHz and 1 MHz. These results are in line with previously published experimentally determined frequency-dependent DEP behavior of the latex microbeads. The phenomenological approach assisted by live AI-guided feedback loop described in the present study will assist the active manipulation of the system towards the desired phenomenological outcome such as, for example, collection of the particles at the electrodes, even if, due to the complexity and plurality of the interactive forces, model-based predictions are not available.https://www.mdpi.com/2072-666X/13/3/399guided micro-/nano-assemblyartificial intelligencedielectrophoresiselectrokinetic assembly |
spellingShingle | Matthew Michaels Shih-Yuan Yu Tuo Zhou Fangzhou Du Mohammad Abdullah Al Faruque Lawrence Kulinsky Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency Micromachines guided micro-/nano-assembly artificial intelligence dielectrophoresis electrokinetic assembly |
title | Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency |
title_full | Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency |
title_fullStr | Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency |
title_full_unstemmed | Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency |
title_short | Artificial Intelligence Algorithms Enable Automated Characterization of the Positive and Negative Dielectrophoretic Ranges of Applied Frequency |
title_sort | artificial intelligence algorithms enable automated characterization of the positive and negative dielectrophoretic ranges of applied frequency |
topic | guided micro-/nano-assembly artificial intelligence dielectrophoresis electrokinetic assembly |
url | https://www.mdpi.com/2072-666X/13/3/399 |
work_keys_str_mv | AT matthewmichaels artificialintelligencealgorithmsenableautomatedcharacterizationofthepositiveandnegativedielectrophoreticrangesofappliedfrequency AT shihyuanyu artificialintelligencealgorithmsenableautomatedcharacterizationofthepositiveandnegativedielectrophoreticrangesofappliedfrequency AT tuozhou artificialintelligencealgorithmsenableautomatedcharacterizationofthepositiveandnegativedielectrophoreticrangesofappliedfrequency AT fangzhoudu artificialintelligencealgorithmsenableautomatedcharacterizationofthepositiveandnegativedielectrophoreticrangesofappliedfrequency AT mohammadabdullahalfaruque artificialintelligencealgorithmsenableautomatedcharacterizationofthepositiveandnegativedielectrophoreticrangesofappliedfrequency AT lawrencekulinsky artificialintelligencealgorithmsenableautomatedcharacterizationofthepositiveandnegativedielectrophoreticrangesofappliedfrequency |