Goal
To create a modular, low-cost embedded microfluidics platform capable of automating enzyme evolution workflows, including droplet sorting, incubation, and optical analysis—all using off-the-shelf microcontrollers and custom-fabricated chips.
Hardware Architecture
- Microcontroller: ESP32-S3 w/ native USB and camera support
- Control Interface: Stepper drivers (A4988), 5V/12V mosfet banks, GPIO breakout
- Pumps: Custom syringe + peristaltic modules w/ motor feedback
- Imaging: OV2640 + 850nm LED for fluorescence; USB/UART for live logging
- Display/UI: OLED 0.96″ + rotary encoder for standalone debug
- Power: 12V main rail with 5V/3.3V regulation; optional Li-ion + solar input
Notes
- ESP32’s dual-core lets us separate UI from real-time control loops
- Isolation diodes needed between MOSFETs and shared power rail
- Image acquisition latency is ~200ms for 320×240; acceptable for endpoint assays
Microfluidic Chip Fabrication
- Material: PMMA (acrylic), 1.5 mm base, 1 mm channel layer
- Method: CO2 laser rastering for channels, solvent-bonded with ethanol + UV
- Channel Specs: 0.5 mm width, 1 mm depth (routed), Y-junctions + incubation loops
- Hydrophobic Treatment: Vapor-phase HMDS post-bonding
Lessons Learned
- Melt-over from laser cutting introduces edge deformation → DCM polishing helps
- HMDS vapor passivation lasts ~1 week before noticeable wetting increase
- Clamping during UV bonding critical to avoid microchannel deformation
Software Stack
- Firmware: PlatformIO (ESP-IDF), FreeRTOS split tasks
- Command Protocol: Serial JSON over USB/UART
- UI: Simple knob navigation; debug output to OLED and USB
- Control Loops: PID flow regulation based on motor encoder feedback
- Logging: External SD card, or Wi-Fi upload in batch mode
Current Functional Modules
Next Steps
- Fabricate V3 of chip with integrated droplet trap zone
- Finish tuning PID for consistent 1-10 uL injection pulses
- Implement high-speed capture for droplet count per flow event
- Integrate magnet-based sorting and validate on fluorescent droplet batches