How accurate is this droplet size estimate?

The Garstecki squeezing-regime correlation (d/w $\approx$ 1 + Qd/Qc) is typically accurate to ±30% for flow-focusing geometries in the squeezing regime. It is a starting point for lab optimization, not a replacement for empirical measurement. Always verify with a test run before production.

What is the Capillary number and why does it matter?

The Capillary number (Ca = μv/σ) measures the ratio of viscous drag to surface tension. At low Ca (< 0.01) the nozzle plugs up and droplets form by squeezing — size is geometry-dominated and monodisperse. At intermediate Ca (0.01–0.3), dripping produces slightly smaller drops at higher frequency. At high Ca (> 0.3), jetting forms polydisperse droplets and is usually undesirable.

Why must the target diameter be larger than the nozzle width?

In the squeezing regime, the dispersed phase fills the entire nozzle cross-section before pinching off. The droplet is always at least as wide as the nozzle. Requesting a droplet smaller than the nozzle means you are not in the squeezing regime — use a smaller nozzle or switch to a dripping-regime chip.

What happens if I use less than 0.5% surfactant?

Below ~0.5% surfactant (PFPE-PEG for HFE-7500, Span 80 for mineral oil), the oil/water interface is not fully covered, and droplets coalesce on contact. The tool flags this because the size estimate is meaningless if droplets merge downstream.

Can I use this for T-junction generators?

The Garstecki correlation was originally developed for T-junctions and later validated for flow-focusing geometries. The physics is similar — both are squeezing-regime — so the estimate is applicable, but flow-focusing typically gives better monodispersity.

Does my data leave the browser?

No. All math runs in your browser. No values are sent to any server.

ToolsConductScience tool

MicrofluidicsFree in-browser calculator

Microfluidic Droplet Generator Parameter Estimator.

Estimate starting Qc/Qd flow rates, droplet volume, generation frequency, and Capillary number regime for flow-focusing droplet generators. Free. Client-side.

Run calculator Request a Quote

PrivateData stays in your browser

LiveNo sign-up required

Validated2026-04-07

CitableMethods and citation included

Calculator

Results update in place

Try it out

Load example microfluidic droplet parameter estimator data to see the full workflow

⚠️ Estimates only. Squeezing-regime droplet diameter is set by a Garstecki-style correlation (d/w ≈ 1 + Qd/Qc) that is good to ~30%. Always validate with a test run before production. Capillary number is the regime indicator — actual droplet size also depends on chip geometry and surfactant chemistry.

Nozzle Geometry

Nozzle width (μm) — flow-focusing orifice

Channel height (μm)

Phases

Continuous (oil) phase

Dispersed (aqueous) phase

Surfactant in continuous phase (% w/w)

Target & Flow

Target droplet diameter (μm)

Total flow constraint (μL/min)

Suggested Starting Point

Continuous phase Qc

100.0 μL/min

Dispersed phase Qd

100.0 μL/min

Flow ratio Qd/Qc

1.00

Droplet volume

523.6 pL

Frequency

3.18 kHz

Capillary number

0.1867

Regime: dripping

Dripping regime (Ca 0.01–0.3): droplet size set by competition between viscous drag and interfacial tension. Most flexible regime; tune flow ratio to fine-tune size.

When to use

Choosing starting Qc/Qd for a new droplet generation chip
Estimating droplet volume and frequency for a reagent budget
Checking whether your flow conditions are in the squeezing, dripping, or jetting regime
Sizing a chip nozzle to hit a target droplet diameter
Teaching the Garstecki scaling law interactively

Do not use for

For precise droplet sizing (validate empirically — this is a ±30% starting estimate)
For emulsions made by shaking, vortexing, or membrane extrusion
For gas-in-liquid (bubble) generators (surface tension and density are different)
For double emulsions (requires a two-step model not covered here)

Start in the squeezing regime

Ca < 0.01 is the default for any new chip. Drops are monodisperse, size depends only on geometry and flow ratio, and the system is forgiving of pump pulsation. Move to dripping only if you need smaller drops.

The ±30% rule

The Garstecki correlation is accurate to ±30%. Do NOT set a spec based on this tool alone — run a 1 mL test to measure actual diameter with a microscope before committing to a production run.

Surfactant concentration is non-negotiable

Below 0.5%, drops coalesce. Below 1%, the interface may not be fully saturated, giving time-dependent coalescence. Use at least 2% for any production application.

Channel height is part of the geometry

The Garstecki model assumes the droplet fills the channel height. If your chip is much deeper than the nozzle is wide, the simple scaling breaks down — the droplet becomes a disc, not a sphere.

Method

Garstecki squeezing-regime scaling: d/w = 1 + α(Qd/Qc), $\alpha$ = 1. From target diameter and nozzle width: flowRatio = (d/w − 1). Qc = totalFlow / (1 + flowRatio). Qd = Qc $\times$ flowRatio. Droplet volume = (4/3)π(d/2)³ μ $\text{m}^{3}$ $\times$ 1e-3 → pL. Frequency = (Qd μL/min $\times$ 1e6 / 60) / volumePL → Hz. Capillary number Ca = μc $\times$ v / $\sigma$ , where v = Qc_m3s / channelArea. Regime: squeezing (Ca < 0.01), dripping (0.01–0.3), jetting (> 0.3). Surface tension and viscosity from published phase datasheets (3M HFE-7500, Sigma M5904 mineral oil) with recommended surfactant.

Validated

Last validated 2026-04-07. Calculations are designed for planning and documentation support; verify procurement decisions against manufacturer specifications or institutional SOPs.

How to cite

How to Cite

ConductScience Microfluidic Droplet Generator Parameter Estimator (v1.13.0). ConductScience, Inc. 2026. Available at: https://conductscience.com/tools/microfluidic-droplet-parameter-estimator

The Garstecki Squeezing-Regime Model

In 2006, Garstecki et al. showed that droplet size in a confined microfluidic channel follows a beautifully simple scaling:

d/w

\approx

1 +

\alpha

\times

(Qd/Qc)

where d is the droplet diameter, w is the nozzle width, Qd is the dispersed-phase flow rate, Qc is the continuous-phase flow rate, and $\alpha$ is a geometry-dependent prefactor of order 1.

The key insight: in the squeezing regime, surface tension dominates over viscous stress, so the droplet length depends only on the flow-rate ratio, not on the absolute flow rates. This makes the squeezing regime extremely predictable and the default for biological applications.

Squeezing, Dripping, Jetting — The Three Regimes

Squeezing (Ca < 0.01): The dispersed phase completely fills the nozzle cross-section. Droplets form when the continuous phase pinches the thread against the channel wall. Size is set by geometry + flow ratio. Monodisperse, low throughput.

Dripping (Ca 0.01–0.3): Viscous drag partially thins the dispersed thread before pinch-off. Droplets are smaller than squeezing at the same Qd/Qc. Most flexible regime for tuning size vs throughput.

Jetting (Ca > 0.3): The dispersed phase forms a thin jet that breaks up downstream by Rayleigh–Plateau instability. High throughput but polydisperse — avoid for applications requiring uniform drops.

Surfactant: The Unsung Hero of Droplet Stability

Surface-active molecules (PFPE-PEG for fluorinated oils, Span 80 for hydrocarbon oils) adsorb to the oil/water interface and lower $\sigma$ , which:

1. Prevents coalescence — drops can touch without merging 2. Speeds up drop formation (lower $\sigma$ → higher Ca at same velocity) 3. Can reduce drop size (weaker surface tension → earlier pinch-off)

At least 0.5% surfactant is required for stable drops. Below that, any downstream contact — in a delay line, in a storage array, at a detector — causes coalescence and destroys the emulsion.

Frequently asked

325

Free tools

1,200+

Institutions

100%

Client-side

Uploads required

Next steps

Equipment