BiQadx
MicrofluidicsQ3 2025 · 12 min read

Capillary Action Modeling using Finite Element Analysis

Finite Element Analysis using COMSOL Multiphysics 6.2 was used to model two-phase capillary flow through 50 µm diagnostic channels on the Stratos cartridge platform. Simulated routing time predictions achieved 94.8% concordance with physical bench measurements, enabling virtual design iteration before CNC prototyping.

BQ
BiQadx Core Engineering
Q3 2025
12 min read
94.8%
Model Concordance
vs. physical bench data
50 µm
Channel Width
Stratos cartridge geometry
11.2s
Fill Time Achieved
Target ≤12s per chamber
◆ Engineering Process Flow
1
MODEL
2
PROTOTYPE
3
FILL TEST
4
ITERATE
5
RELEASE
◆ Key Findings
  • Corner radius is the dominant fill-time variable: 90° corners increase fill time 31% and cause air entrapment in 27% of runs
  • Optimised geometry (15µm radius, 0.25mm vent) achieves 11.2s fill time with zero air inclusion events across n=30
  • COMSOL FSI coupling reduced model error from +15.8% to +3.7% for flexible substrate sections
01

Physics of Two-Phase Capillary Flow at Microscale

Below 200 µm channel width, capillary pressure dominates over gravitational forces (Bond number Bo < 0.01). The Washburn equation governs fill rate: L = √(Rγcosθ·t/2η) where R is the hydraulic radius, γ is surface tension (0.072 N/m for plasma/air), θ is the contact angle (38° for plasma on COC), and η is dynamic viscosity (1.2 mPa·s). At 50 µm width, predicted fill velocity is 2.8 mm/s with a contact angle hysteresis correction of +4.2% due to surface roughness Ra 0.3 µm from CNC milling.

02

COMSOL Model Setup & Mesh Strategy

The Level Set method (LSM) was selected for interface tracking over Volume-of-Fluid (VOF) due to superior mass conservation at the capillary scale. A physics-controlled swept mesh with 3 µm maximum element size at channel walls was applied — 1.2M elements total. Surface tension was implemented as a continuum surface force (CSF). Blood plasma was modelled as a Newtonian fluid (validated for Hct < 45% in channels < 1mm). Solver: PARDISO direct solver, adaptive time-stepping 0.001–0.05s. Compute time: 4.3 hours per geometry variant on 32-core workstation.

03

Parametric Optimisation Findings

A 72-variant parametric sweep (channel width 30–100 µm, corner radius 0–25 µm, vent hole diameter 0.1–0.4 mm) revealed that corner radius is the dominant variable: sharp 90° corners create air entrapment vortices that increase fill time by 31%. Rounded corners (r = 15 µm) eliminated entrapment in 96% of simulated cases. Vent hole sizing follows a critical threshold: below 0.15 mm, back-pressure build-up stalls flow; above 0.30 mm, evaporation rate exceeds acceptable limits (>5 µL/h at 37°C).

04

Experimental Validation Protocol

Physical validation used high-speed microscopy (Photron FASTCAM Nova, 1000 fps) to track the fluid front through fluorescent-dyed plasma. 30 cartridge replicates per geometry variant were tested at 25°C. Fill time, meniscus shape, and air inclusion events were compared to COMSOL predictions. The 94.8% concordance was achieved for the final optimised geometry; early variants showed only 71% concordance, prompting the addition of fluid-structure interaction (FSI) coupling for deformable PDMS sections.

Channel Geometry Variants — COMSOL vs. Bench Fill Time (seconds)
Geometry VariantCOMSOL PredictedBench MeasuredΔ ErrorAir Inclusion Events
50µm / 90° corners / 0.10mm vent18.4s21.3s+15.8%8/30
50µm / 15µm radius / 0.10mm vent13.1s13.9s+6.1%2/30
50µm / 15µm radius / 0.20mm vent11.7s12.2s+4.3%0/30
50µm / 15µm radius / 0.25mm vent (FINAL)10.8s11.2s+3.7%0/30
n=30 per variant. Plasma at 25°C. Hct 38–42%. Concordance calculated as (1 − mean|Δ|/mean bench) × 100.BiQadx Engineering Data

Research Context Only: This document is published as an engineering log for transparency. All content describes R&D-phase investigations. No clinical diagnostic claims are made. This is not a regulatory filing or clinical performance specification.

Engineering LibraryINS-005 / BiQadx © 2026
BiQadx content is R&D / prototype / pilot-stage. No clinical claims. For planning and technical understanding only. Not medical advice.