Characterisation of droplet-compartmentalised chemical oscillator networks using computer vision
Characterisation of droplet-compartmentalised chemical oscillator networks using computer vision
The challenge of overcoming limitations in computing technology drives the development of unconventional computing materials and architectures. Microfluidic and droplet-in-oil technology is a field of study that allows a large number of communicating chemical droplets of different compositions to be efficiently produced as a new form of engineering and computing substrate. To develop this technology, this project utilises the Belousov-Zhabotinsky (BZ) chemical oscillator of different compositions as aqueous droplets in an asolectin in hexadecane system. Previous studies were limited to structured homogeneous solutions and emulsions or to arrays of micrometre-sized droplets of identical composition. In comparison, this project uses millimetre-sized droplets enabling wave patterns to be observed while using droplets of different compositions coupled in different topologies. These droplet array topologies are defined using laser-cut acrylic (PMMA) with complex designs. This allows the effect of changing individual BZ components to be observed over many repeats and in different topologies. Wave characteristics of different BZ droplet compositions are statistically analysed over many repeats with computer vision techniques. Changing H2SO4, NaBrO3, malonic acid and ferroin concentrations changes the oscillation lifetime, frequency and total wave count as well as the amplitude and area of each peak. When droplets are coupled in arrays of more than one droplet, wave sources are formed at droplet interfaces but without wave propagation through the interface and no change to wave characteristics. If the coupled droplet becomes exhausted, the wave source at that interface disappears and other wave sources become dominant. To study wave propagation through droplet interfaces, a BZ composition using both malonic acid and 1,4-cyclohexanedione is developed. These droplets have a shorter induction phase and waves that can propagate through droplet interfaces of at least 20 droplets over 55 mm. Arrays of up to 399 droplets are formed and stable for at least 4 hours. Microfluidic devices are important for future applications as they allow forming large networks of coupled droplets in different topologies to be scaled up and precisely controlled. To test the compatibility of BZ with polydimethylsiloxane (PDMS) microfluidic devices, BZ droplets were mixed in situ and observed in a millifluidic chip for at least 3 hours. These developments reveal opportunities to study and advance future applications of BZ or other chemical droplets in large arrays of coupled droplets for future applications in unconventional computing.
University of Southampton
Chang, Kai Ming
4a4a15be-862b-4cd2-a229-bc4562f6b47e
February 2018
Chang, Kai Ming
4a4a15be-862b-4cd2-a229-bc4562f6b47e
De Planque, Maurits
a1d33d13-f516-44fb-8d2c-c51d18bc21ba
Zauner, Klaus-Peter
c8b22dbd-10e6-43d8-813b-0766f985cc97
Chang, Kai Ming
(2018)
Characterisation of droplet-compartmentalised chemical oscillator networks using computer vision.
University of Southampton, Doctoral Thesis, 262pp.
Record type:
Thesis
(Doctoral)
Abstract
The challenge of overcoming limitations in computing technology drives the development of unconventional computing materials and architectures. Microfluidic and droplet-in-oil technology is a field of study that allows a large number of communicating chemical droplets of different compositions to be efficiently produced as a new form of engineering and computing substrate. To develop this technology, this project utilises the Belousov-Zhabotinsky (BZ) chemical oscillator of different compositions as aqueous droplets in an asolectin in hexadecane system. Previous studies were limited to structured homogeneous solutions and emulsions or to arrays of micrometre-sized droplets of identical composition. In comparison, this project uses millimetre-sized droplets enabling wave patterns to be observed while using droplets of different compositions coupled in different topologies. These droplet array topologies are defined using laser-cut acrylic (PMMA) with complex designs. This allows the effect of changing individual BZ components to be observed over many repeats and in different topologies. Wave characteristics of different BZ droplet compositions are statistically analysed over many repeats with computer vision techniques. Changing H2SO4, NaBrO3, malonic acid and ferroin concentrations changes the oscillation lifetime, frequency and total wave count as well as the amplitude and area of each peak. When droplets are coupled in arrays of more than one droplet, wave sources are formed at droplet interfaces but without wave propagation through the interface and no change to wave characteristics. If the coupled droplet becomes exhausted, the wave source at that interface disappears and other wave sources become dominant. To study wave propagation through droplet interfaces, a BZ composition using both malonic acid and 1,4-cyclohexanedione is developed. These droplets have a shorter induction phase and waves that can propagate through droplet interfaces of at least 20 droplets over 55 mm. Arrays of up to 399 droplets are formed and stable for at least 4 hours. Microfluidic devices are important for future applications as they allow forming large networks of coupled droplets in different topologies to be scaled up and precisely controlled. To test the compatibility of BZ with polydimethylsiloxane (PDMS) microfluidic devices, BZ droplets were mixed in situ and observed in a millifluidic chip for at least 3 hours. These developments reveal opportunities to study and advance future applications of BZ or other chemical droplets in large arrays of coupled droplets for future applications in unconventional computing.
Text
Final Thesis
- Version of Record
More information
Published date: February 2018
Identifiers
Local EPrints ID: 423559
URI: http://eprints.soton.ac.uk/id/eprint/423559
PURE UUID: 4166ec12-3d98-43c0-9edb-c9a778a4bf75
Catalogue record
Date deposited: 26 Sep 2018 16:30
Last modified: 15 Mar 2024 21:49
Export record
Contributors
Author:
Kai Ming Chang
Thesis advisor:
Maurits De Planque
Thesis advisor:
Klaus-Peter Zauner
Download statistics
Downloads from ePrints over the past year. Other digital versions may also be available to download e.g. from the publisher's website.
View more statistics