Analysis and simulations of droplet generation regimes in a coaxial microfluidic device
Listen
Radio episode about this paper
Transcript
Introduction to the show: ident: Physics. The day in general physics: instruments, methods and results across the field.
Kai: We're looking at the paper: "Analysis and simulations of droplet generation regimes in a coaxial microfluidic device".
Mira: Here's the gist of the paper. The generation of microdroplets via segmentation in microfluidic devices is of interest in many applications, from biochemical to pharmaceutical <ref:2601.09483#pg6>.
Kai: This work evaluates the suitability of the open-source software Basilisk to accurately predict microdroplet generation by segmentation <ref:2601.09483#pg6> and proposes new scaling laws for the prediction of droplet diameters in the dripping and jetting regimes <ref:2601.09483#pg6>.
Mira: Modeling Fluid Dynamics The dynamics of droplet generation in a microfluidic device are described where two non-miscible fluids enter from separate channels, come in contact at a junction, and flow out from the same channel <ref:2601.09483#pg8> The physical parameters characterizing the problem are the viscosities and densities of the two fluids, µi and ρi, and surface tension σ <ref:2601.09483#pg8> Based on these physical, geometrical and flow parameters, four dimensionless groups can be formed <ref:2601.09483#pg8> These groups include the viscosity and density ratios, respectively µd/µc and ρd/ρc <ref:2601.09483#pg8>, as well as the Reynolds and Capillary number <ref:2601.09483#pg8>.
Kai: Numerical Implementation of Basilisk Basilisk is a library of solvers that can be used for a variety of fluid dynamics applications, from industrial to geophysical flows <ref:2601.09483#pg9> All the different parts share a common discretization based on the finite volume method on Cartesian grid <ref:2601.09483#pg9>.
Title and authors: Kai: So we’re talking about "Analysis and simulations of droplet generation regimes in a coaxial microfluidic device." The authors are Alessio Innocenti, Andrea Poggi, Simone Camarri, and Maria Vittoria Salvetti from the University of Pisa. It sounds like they’ve put a lot of effort into simulating these systems using Basilisk.
Mira: They aren't just running one simulation; they’re systematically analyzing how the production mechanisms change depending on the physical and flow parameters in a coaxial setup. That suggests they are trying to map out the entire landscape of droplet formation, from one condition to another.
Lev: What I find interesting is that they are testing Basilisk against analytical solutions from Bretherton theory and literature experimental results. For me, it means they aren't just making up numbers; they’re grounding their numerical predictions in established theoretical frameworks and what we already know from the lab.
Kai: Exactly. They’re using this comparison to see how well Basilisk predicts droplet generation by segmentation, which is a specific technique they are interested in because it gives better control over size and uniformity than older batch methods do.
Mira: And when you look at the regimes they're studying, they’re focusing on how the balance between surface tension, viscous stresses from the outer fluid, and inertial forces from the inner fluid dictates whether you get dripping or jetting.
Lev: The paper also mentions that they are performing several campaigns of numerical simulations for a coaxial device to analyze these different regimes. That suggests they aren't just looking at one perfect setup; they’re testing how the system behaves under varying conditions.
Kai: So, in short, this paper is about using a specific computer tool to explore the physics of droplet making in coaxial microchannels and checking if the results match what theory and experiments suggest. It sets up a good foundation for understanding these systems.
Mira: And it’s important because they are looking at how things like the viscosity ratio play into this, which is something that often gets overlooked when you just look at simple flow setups. That’s a big piece of the puzzle here for understanding droplet behavior.
Lev: It gives us a way to see what these simulations would actually need to run on real hardware later, by testing the software’s accuracy against known physics first.
The paper's summary: Kai: To summarize, this paper dives into the dynamics of droplet generation in a coaxial microfluidic device where two non-miscible fluids meet and flow out. They are using segmentation to generate these droplets and they’re evaluating Basilisk software for its accuracy in predicting this whole process.
Mira: The summary highlights that the geometrical configurations can be quite varied, ranging from cross-flow and T-junction setups to co-flow and flow focusing setups, which shows the breadth of what they are studying.
Lev: It’s clear they are looking at how surface tension provides a stabilizing effect while viscous stresses and inertial forces destabilize the system on different sides of the junction. That’s a classic fluid dynamics problem in microchannels that requires careful balancing.
Kai: And they quantify this balance using dimensionless numbers, specifically the Capillary number and the Weber number, to characterize these different regimes of formation and breakup.
Mira: The summary points out that at very low capillary numbers, surface tension dominates, leading to the jet tip growing until it blocks the junction before breaking off. But at larger capillary numbers, you get a steady dripping regime or a convective instability leading to a jetting regime.
Lev: So they are essentially defining the transition points between these two distinct modes—dripping and jetting—based on how those dimensionless numbers behave as you change the flow conditions.
Kai: What this means for us is that we have a clearer picture of what drives the breakup instability, which helps us understand how to design channels to get a specific droplet size or generation rate.
Mira: It also sets up the next step: they propose new scaling laws for predicting droplet diameters in both the dripping and jetting regimes, moving beyond just describing the mechanisms.
Lev: That’s a big leap because those scaling laws are what you actually use to predict things in a design phase, not just theoretical descriptions of how it happens.
The paper's improvements: Kai: The authors suggest some specific improvements to their work, and one of the most important is taking additional physical effects into account, specifically the viscosity ratio. They feel that incorporating this parameter will refine the existing scaling laws significantly.
Mira: I agree with that. When you’re modeling flows where you have two fluids with different viscosities, ignoring that ratio can lead to big errors in predicting things like droplet diameters and pinch-off positions.
Lev: From an engineering standpoint, if they incorporate the viscosity ratio properly into the scaling laws, it makes the model much more useful for real hardware design because it accounts for a real physical difference between the fluids.
Kai: They derive a specific scaling law for the dripping regime: ddrop over D equals zero point two three three times Ca to the negative zero point three three times c to the positive zero point two times d multiplied by mu sub d over mu sub c raised to the negative zero point two two and Re to the negative zero point zero zero zero.
Mira: That formula shows how droplet diameter relates directly to all those parameters, including that viscosity ratio term they mentioned earlier, which is a nice refinement of what was previously there.
Lev: And for the jetting regime, they give a scaling law as well: ddrop over D equals two point four eight multiplied by the ratio of flow rates divided by the sum of flow rates over two times that ratio to the power zero and mu sub d over mu sub c raised to zero and zero point one two four.
Kai: So they’ve essentially taken empirical relations and refined them by adding these new physical dependencies, which is exactly what you need when you want accurate predictions for a device.
Mira: The improvement here is moving from just describing the regimes to actually providing predictive equations for the droplet diameters in those regimes, which is a much more practical thing to have.
Conclusion: Kai: So, wrapping up on this paper, what we’ve seen is that they’ve used Basilisk to test droplet generation by segmentation in coaxial microfluidics and found it works well against theory and experiments. They are also proposing new scaling laws for the dripping and jetting regimes that include things like the viscosity ratio.
Mira: The big implication is that we now have better tools for predicting droplet sizes in these complex flows, not just a general description of how they form, but specific equations for different flow types.
Lev: From my perspective on running this on real hardware, it means the simulations are getting more physically grounded when we talk about scaling them up or trying to translate them into something you can build.
Kai: It’s definitely a step forward in using these simulation tools to guide the design of microfluidic devices for things like drug delivery, where precise control over droplet size is so important.
Mira: The authors acknowledge that their study focuses on the effects of varying the Capillary number and Weber number, and they point out that fully three-dimensional simulations are still a bit ahead in development right now.
Lev: So they admit the limitation is that while they’ve mapped out many regimes, getting those truly full three-dimensional simulations done is still a bit of a work in progress for the software.
Kai: Exactly. It shows where we are now, and what’s next—that continued improvement of solvers will let them push into those fully three-dimensional scenarios soon.
Mira: So, to finish up on "Analysis and simulations of droplet generation regimes in a coaxial microfluidic device," we've seen a solid validation of the tools and some new predictive scaling laws that incorporate viscosity ratios for both dripping and jetting regimes.
Lev: It gives us a clearer path for how these numerical models can be applied to actual device design, by accounting for those real physical differences between the fluids.
Alessio Innocenti, Andrea Poggi, Simone Camarri, Maria Vittoria Salvetti
University of Pisa
physics.flu-dyn
Submitted: 2026-01-14
Updated: 2026-11-04
Comments: 35 pages. Accepted version
DOI: 10.1016/j.compfluid.2026.107334
License: http://creativecommons.org/licenses/by-nc-nd/4.0/
Importance score: 80/100
The gist: The generation of microdroplets via segmentation in microfluidic devices is of interest in many applications, from biochemical to pharmaceutical <ref:2601.09483#pg6>.
Key concepts
- Basilisk
- Basilisk is an open-source library of solvers used to model various fluid dynamics problems. In this study, it was tested to see if it accurately predicts how microdroplets form when fluids interact in a microfluidic device.
- Dripping and Jetting Regimes
- These are two distinct flow regimes during droplet generation. Dripping involves slow, discrete release of droplets from the junction, while jetting involves a continuous stream of fluid exiting the device. The paper maps these regimes based on physical parameters like Capillary number.
- Capillary Number (Ca)
- The Capillary number is a dimensionless group that relates viscous forces to surface tension forces in fluid flow. It is used to categorize the flow behavior, helping researchers determine whether a system operates in the dripping or jetting regime for droplet formation.
Terminology
Summary
The generation of microdroplets via segmentation in microfluidic devices is of interest in many applications, from biochemical to pharmaceutical <ref:2601. p. 0 Analysis and simulations of droplet generation regimes
in a coaxial microfluidic device
Alessio Innocenti∗
, Andrea Poggi, Simone Camarri, Maria Vittoria Salvetti
University of Pisa, Dipartimento di Ingegneria Civile e Industriale, Largo Lucio
Lazzarino, 56122, Pisa, Italy
Abstract
The generation of microdroplets via segmentation in microfluidic devices is
of interest in many applications, from biochemical to pharmaceutical. This
technique permits indeed much higher control on the droplet size, uniformity
and generation rate than in standard batch generation processes. This
technique permits indeed much higher control on the droplet size, uniformity
and generation rate than in standard batch generation processes. In this work
we have evaluated the suitability of the open-source software Basilisk to accurately predict microdroplet generation by segmentation. We have then performed several campaigns of
numerical simulations for a coaxial device, analyzing the different regimes of
droplet generation, and evaluating how the physical and flow parameters affect the production mechanisms and the diameters of the generated droplets. Finally we have proposed new scaling laws for the prediction of droplet diameters in the dripping and jetting regimes, refining existing ones by taking
∗Corresponding author
Email address: alessio.Where the paper says it
Where the paper says it
Where the paper says it
Modeling Fluid Dynamics
The dynamics of droplet generation in a microfluidic device are described where two non-miscible fluids enter from separate channels, come in contact at a junction, and flow out from the same channel <ref:2601.09483#pg8> The physical parameters characterizing the problem are the viscosities and densities of the two fluids, µi and ρi, and surface tension σ <ref:2601. p. 5 Modeling and governing equations
We describe the dynamics of droplet generation in a microfluidic device,
where two non-miscible fluids enter from separate channels, come in contact
at a junction, and flow out from the same channel. The physical parameters
characterizing the problem are the viscosities and densities of the two fluids,
µi and ρi
, and surface tension σ. ), the additional parameters characterizing the flow are the inlet velocities
of the two fluids, Uc and Ud, and a characteristic dimension of the channels,
typically the diameter, D, in coaxial configurations or the channel height,
H, in square section channels. Based on these physical, geometrical and flow
parameters, four dimensionless groups can be formed: firstly the viscosity
and density ratios, respectively µd/µc and ρd/ρc; secondly the Reynolds and
Capillary number, that for these kind of flows can be defined based on the
continuous phase properties as follows
Rec =
ρcUcD
µc
, Cac =
µcUc
σ
(1)
An alternative, quite common in microfluidics literature, is to replace
one of these two dimensionless number with the Weber number, defined as
follows:
W ec =
ρcU
2
c D
σ
= Rec · Cac (2)
The two fluids are governed by the Navier-Stokes incompressible equations with variable density and viscosity, which in their dimensionless form
6 ), the additional parameters characterizing the flow are the inlet velocities
of the two fluids, Uc and Ud, and a characteristic dimension of the channels,
typically the diameter, D, in coaxial configurations or the channel height,
H, in square section channels. Based on these physical, geometrical and flow
parameters, four dimensionless groups can be formed: firstly the viscosity
and density ratios, respectively µd/µc and ρd/ρc; secondly the Reynolds and
Capillary number, that for these kind of flows can be defined based on the
continuous phase properties as follows
Rec =
ρcUcD
µc
, Cac =
µcUc
σ
(1)
An alternative, quite common in microfluidics literature, is to replace
one of these two dimensionless number with the Weber number, defined as
follows:
W ec =
ρcU
2
c D
σ
= Rec · Cac (2)
The two fluids are governed by the Navier-Stokes incompressible equations with variable density and viscosity, which in their dimensionless form
6 ), the additional parameters characterizing the flow are the inlet velocities
of the two fluids, Uc and Ud, and a characteristic dimension of the channels,
typically the diameter, D, in coaxial configurations or the channel height,
H, in square section channels. Based on these physical, geometrical and flow
parameters, four dimensionless groups can be formed: firstly the viscosity
and density ratios, respectively µd/µc and ρd/ρc; secondly the Reynolds and
Capillary number, that for these kind of flows can be defined based on the
continuous phase properties as follows
Rec =
ρcUcD
µc
, Cac =
µcUc
σ
(1)
An alternative, quite common in microfluidics literature, is to replace
one of these two dimensionless number with the Weber number, defined as
follows:
W ec =
ρcU
2
c D
σ
= Rec · Cac (2)
The two fluids are governed by the Navier-Stokes incompressible equations with variable density and viscosity, which in their dimensionless form
6Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Numerical Implementation of Basilisk
Basilisk is a library of solvers that can be used for a variety of fluid dynamics applications, from industrial to geophysical flows <ref:2601.09483#pg9> All the different parts share a common discretization based on the finite volume method on Cartesian grid <ref:2601. p. 7 build a specific solver assembly with the desired features. All the different parts share a common discretization based on the finite volume method
on Cartesian grid. All simulations performed with Basilisk have been
done with adaptive mesh refinement (AMR) on quadtree grids. The interface is advected with a geometric piecewise-linear interface construction (PLIC) VOF method ([21, 22]), generalized for the quad/octree
spatial discretization, which can ensure a sharp interface while preserving
mass conservation. All simulations performed with Basilisk have been
done with adaptive mesh refinement (AMR) on quadtree grids. The interface is advected with a geometric piecewise-linear interface construction (PLIC) VOF method ([21, 22]), generalized for the quad/octree
spatial discretization, which can ensure a sharp interface while preserving
mass conservation. sion successively, with a fractional step method that guarantees exact mass
conservation. The surface tension term in Equation (4) is modeled with the continuum
surface force (CSF) approach [24].Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Validation and Simulation Results
The work validates Basilisk by analyzing a 2-dimensional Taylor bubble configuration, finding excellent agreement with theoretical and empirical literature laws for droplet terminal velocity <ref:2601.09483#pg8> For the coaxial configuration, simulations correctly reproduce the generation mechanisms, and the diameter and pinch-off position are in agreement with experimental images <ref:2601. p. 0 We have validated
the numerical tool with analytical solutions for the dynamics of droplets in
confined flows, namely with Bretherton theory, and by comparison with literature experimental results. We have then performed several campaigns of
numerical simulations for a coaxial device, analyzing the different regimes of
droplet generation, and evaluating how the physical and flow parameters affect the production mechanisms and the diameters of the generated droplets. Even if not always all these
aspect are considered rigorously, the literature on numerical approaches to
droplet generation in microdevices is rapidly increasing, and different methods and softwares are now commonly used in device design. [7]
have simulated coflowing devices with a front-tracking method [8] to track the
interface between the two immiscible phases, exploiting mainly the dripping
and jetting regimes at different values of the Capillary and Weber numbers.Where the paper says it
Where the paper says it
Proposed Scaling Laws
New scaling laws are proposed to predict droplet diameters in the dripping and jetting regimes, refining existing ones by taking into account additional physical effects like the viscosity ratio <ref:2601.09483#pg8> The derived scaling law for the dripping regime is ddrop/D = 0. p. 4 New scaling laws
are finally proposed to predict droplet diameters in the dripping and jetting
regimes, improving existing ones by including additional effects such as the
viscosity ratio. The work is structured as follows: in Section 2 we briefly describe the
mathematical formulation of the problem; in Section 3 we report the details
of the numerical implementation of Basilisk; in Section 4 we validate and
assess the accuracy of Basilisk for a Taylor bubble configuration; in Section
5 we analyze in details a coaxial configuration for droplet generation with
validation against literature experiments, regime mapping and derivation of
scaling laws for the droplet diameter. ), the additional parameters characterizing the flow are the inlet velocities
of the two fluids, Uc and Ud, and a characteristic dimension of the channels,
typically the diameter, D, in coaxial configurations or the channel height,
H, in square section channels. Based on these physical, geometrical and flow
parameters, four dimensionless groups can be formed: firstly the viscosity
and density ratios, respectively µd/µc and ρd/ρc; secondly the Reynolds and
Capillary number, that for these kind of flows can be defined based on the
continuous phase properties as follows
Rec =
ρcUcD
µc
, Cac =
µcUc
σ
(1)
An alternative, quite common in microfluidics literature, is to replace
one of these two dimensionless number with the Weber number, defined as
follows:
W ec =
ρcU
2
c D
σ
= Rec · Cac (2)
The two fluids are governed by the Navier-Stokes incompressible equations with variable density and viscosity, which in their dimensionless form
6 innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. New scaling laws
are finally proposed to predict droplet diameters in the dripping and jetting
regimes, improving existing ones by including additional effects such as the
viscosity ratio. The work is structured as follows: in Section 2 we briefly describe the
mathematical formulation of the problem; in Section 3 we report the details
of the numerical implementation of Basilisk; in Section 4 we validate and
assess the accuracy of Basilisk for a Taylor bubble configuration; in Section
5 we analyze in details a coaxial configuration for droplet generation with
validation against literature experiments, regime mapping and derivation of
scaling laws for the droplet diameter. ), the additional parameters characterizing the flow are the inlet velocities
of the two fluids, Uc and Ud, and a characteristic dimension of the channels,
typically the diameter, D, in coaxial configurations or the channel height,
H, in square section channels. Based on these physical, geometrical and flow
parameters, four dimensionless groups can be formed: firstly the viscosity
and density ratios, respectively µd/µc and ρd/ρc; secondly the Reynolds and
Capillary number, that for these kind of flows can be defined based on the
continuous phase properties as follows
Rec =
ρcUcD
µc
, Cac =
µcUc
σ
(1)
An alternative, quite common in microfluidics literature, is to replace
one of these two dimensionless number with the Weber number, defined as
follows:
W ec =
ρcU
2
c D
σ
= Rec · Cac (2)
The two fluids are governed by the Navier-Stokes incompressible equations with variable density and viscosity, which in their dimensionless form
6 innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Where the paper says it
Regime Mapping and Conclusion
The systematic parametric study identified different regimes of droplet generation in terms of Capillary number of the continuous phase and Weber number of the dispersed phase <ref:2601.09483#pg8>. The results support and complement literature results, providing valuable information for preliminary design of microfluidic coaxial devices <ref:2601. p. 0 We have validated
the numerical tool with analytical solutions for the dynamics of droplets in
confined flows, namely with Bretherton theory, and by comparison with literature experimental results. We have then performed several campaigns of
numerical simulations for a coaxial device, analyzing the different regimes of
droplet generation, and evaluating how the physical and flow parameters affect the production mechanisms and the diameters of the generated droplets. In this work we want to evaluate the suitability of the opensource software Basilisk (https://basilisk. fr) to tackle those kind of applications. into account additional physical effects, like the viscosity ratio. Introduction
The processes of droplet generation at the milli to micro scale has attracted considerable attention due to the numerous applications in which
micro-droplets are involved, such as: (i) lab-on-a-chip devices (miniaturized
reproductions of laboratory processes at the micro/milli scale) for sample
preparation, mixing, separation and detection; (ii) drug delivery through
microparticles, nanoparticles, coated microparticles and microencapsulation,
which has the great advantage of delivering small doses of drug only in the
desired site, reducing accumulation and undesired systemic effects, together
with a better control on the release rate, by modifying the microenvironment
[1]; (iii) encapsulation and manipulations of single cells in aqueous droplets
for cell culture, which allows isolating each individual cell in its microenvironement [2].Where the paper says it
Where the paper says it
Where the paper says it
Acknowledgments
This work is supported by PNRR M4C2 - HPC, Big data and Quantum Computing (Simulazioni, calcolo e analisi dei dati e altre prestazioni - CN1) - CUP I53C22000690001 SPOKE 6 Multiscale modeling & Engineering applications <ref:2601.09483#pg8>. We also acknowledge the CINECA award under the ISCRA initiative for the availability of high performance computing resources and support <ref:2601.09483#pg8>.
References
[1] Ava M Vargason, Aaron C Anselmo, and Samir Mitragotri. The evolution of commercial drug delivery technologies. p. 1 into account additional physical effects, like the viscosity ratio. Introduction
The processes of droplet generation at the milli to micro scale has attracted considerable attention due to the numerous applications in which
micro-droplets are involved, such as: (i) lab-on-a-chip devices (miniaturized
reproductions of laboratory processes at the micro/milli scale) for sample
preparation, mixing, separation and detection; (ii) drug delivery through
microparticles, nanoparticles, coated microparticles and microencapsulation,
which has the great advantage of delivering small doses of drug only in the
desired site, reducing accumulation and undesired systemic effects, together
with a better control on the release rate, by modifying the microenvironment
[1]; (iii) encapsulation and manipulations of single cells in aqueous droplets
for cell culture, which allows isolating each individual cell in its microenvironement [2]. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
Where the paper says it
[2] Sarah Köster, Francesco E Angile, Honey Duan, Jeremy J Agresti, Anton Wintner, Christian Schmitz, Amy C Rowat, Christoph A Merten, Dario Pisignano, Andrew D Griffiths, et al. Drop-based microfluidic devices for encapsulation of single cells. p. 1 Introduction
The processes of droplet generation at the milli to micro scale has attracted considerable attention due to the numerous applications in which
micro-droplets are involved, such as: (i) lab-on-a-chip devices (miniaturized
reproductions of laboratory processes at the micro/milli scale) for sample
preparation, mixing, separation and detection; (ii) drug delivery through
microparticles, nanoparticles, coated microparticles and microencapsulation,
which has the great advantage of delivering small doses of drug only in the
desired site, reducing accumulation and undesired systemic effects, together
with a better control on the release rate, by modifying the microenvironment
[1]; (iii) encapsulation and manipulations of single cells in aqueous droplets
for cell culture, which allows isolating each individual cell in its microenvironement [2]. Droplet generation via segmentation in microfluidic devices
has the great advantage with respect to batch methods of ensuring excellent
control on size distribution, allowing the generation of almost monodisperse
droplets, and on generation rates, at the cost of lower processed volumes. into account additional physical effects, like the viscosity ratio. Introduction
The processes of droplet generation at the milli to micro scale has attracted considerable attention due to the numerous applications in which
micro-droplets are involved, such as: (i) lab-on-a-chip devices (miniaturized
reproductions of laboratory processes at the micro/milli scale) for sample
preparation, mixing, separation and detection; (ii) drug delivery through
microparticles, nanoparticles, coated microparticles and microencapsulation,
which has the great advantage of delivering small doses of drug only in the
desired site, reducing accumulation and undesired systemic effects, together
with a better control on the release rate, by modifying the microenvironment
[1]; (iii) encapsulation and manipulations of single cells in aqueous droplets
for cell culture, which allows isolating each individual cell in its microenvironement [2].Where the paper says it
Where the paper says it
[3] Shelley Lynn Anna. Droplets and bubbles in microfluidic devices. p. 1 Introduction
The processes of droplet generation at the milli to micro scale has attracted considerable attention due to the numerous applications in which
micro-droplets are involved, such as: (i) lab-on-a-chip devices (miniaturized
reproductions of laboratory processes at the micro/milli scale) for sample
preparation, mixing, separation and detection; (ii) drug delivery through
microparticles, nanoparticles, coated microparticles and microencapsulation,
which has the great advantage of delivering small doses of drug only in the
desired site, reducing accumulation and undesired systemic effects, together
with a better control on the release rate, by modifying the microenvironment
[1]; (iii) encapsulation and manipulations of single cells in aqueous droplets
for cell culture, which allows isolating each individual cell in its microenvironement [2]. Droplet generation via segmentation in microfluidic devices
has the great advantage with respect to batch methods of ensuring excellent
control on size distribution, allowing the generation of almost monodisperse
droplets, and on generation rates, at the cost of lower processed volumes. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[4] Taotao Fu, Yining Wu, Youguang Ma, and Huai Z Li. Droplet formation and breakup dynamics in microfluidic flow-focusing devices: From dripping to jetting. p. 1 Droplet generation via segmentation in microfluidic devices
has the great advantage with respect to batch methods of ensuring excellent
control on size distribution, allowing the generation of almost monodisperse
droplets, and on generation rates, at the cost of lower processed volumes. The
generation of microdroplets via segmentation is generally achieved at a junction of separate channels fed with immiscible fluids, generally oil and an aqueous solution, through different possible regimes, namely squeezing, dripping
and jetting [3]. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[5] Piotr Garstecki, Howard A Stone, and George M Whitesides. Mechanism for flow-rate controlled breakup in confined geometries: A route to monodisperse emulsions. p. 1 Introduction
The processes of droplet generation at the milli to micro scale has attracted considerable attention due to the numerous applications in which
micro-droplets are involved, such as: (i) lab-on-a-chip devices (miniaturized
reproductions of laboratory processes at the micro/milli scale) for sample
preparation, mixing, separation and detection; (ii) drug delivery through
microparticles, nanoparticles, coated microparticles and microencapsulation,
which has the great advantage of delivering small doses of drug only in the
desired site, reducing accumulation and undesired systemic effects, together
with a better control on the release rate, by modifying the microenvironment
[1]; (iii) encapsulation and manipulations of single cells in aqueous droplets
for cell culture, which allows isolating each individual cell in its microenvironement [2]. Droplet generation via segmentation in microfluidic devices
has the great advantage with respect to batch methods of ensuring excellent
control on size distribution, allowing the generation of almost monodisperse
droplets, and on generation rates, at the cost of lower processed volumes. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[6] Thomas Cubaud and Thomas G Mason. Capillary threads and viscous droplets in square microchannels. p. 9 More
details on geometries and boundary conditions are given in the following
sections. 2-dimensional Taylor bubble
Before studying the droplet generation in a coaxial device, we assess the
accuracy of Basilisk in reproducing the dynamics of a Taylor bubble, since
the literature on Basilisk applications to problems with bubbles/droplets
in microchannels is still very limited to date. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[7] Jinsong Hua, Baili Zhang, and Jing Lou. Numerical simulation of microdroplet formation in coflowing immiscible liquids. p. 3 Even if not always all these
aspect are considered rigorously, the literature on numerical approaches to
droplet generation in microdevices is rapidly increasing, and different methods and softwares are now commonly used in device design. [7]
have simulated coflowing devices with a front-tracking method [8] to track the
interface between the two immiscible phases, exploiting mainly the dripping
and jetting regimes at different values of the Capillary and Weber numbers. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[8] Salih Ozen Unverdi and Grétar Tryggvason. A front-tracking method for viscous, incompressible, multi-fluid flows. p. 3 Afkhami and coworkers
have explored the use of Volume-of-Fluid (VOF) methods for the interface
tracking, together with the geometrical height function method for the calculation of curvature and the estimation of surface tension, in a T-junction [11]
and in a co-flowing configuration in a Hele-Shaw cell [12]. A different VOF
implementation has been adopted in another work [13] to study the effect
of the injection angle in a 3-dimensional cross-flow geometry, while the most
widely used numerical tool for such kind of flows probably remains OpenFOAM with its interFoam solver, which is still based on the VOF method,
but with a diffuse interface approach [14, 15, 16]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[9] M De Menech, Piotr Garstecki, F Jousse, and Howard A Stone. Transition from squeezing to dripping in a microfluidic T-shaped junction. p. 3 De Menech et al. [9] have analyzed the transition from squeezing to dripping
in a T-junction with the phase-field method, which adopts a diffuse interface
approach, justified by the low capillary numbers (dominant effect of surface
tension) in microfluidics, and relating the breakup dynamics to the upstream
pressure variations. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[10] Long Wu, Michihisa Tsutahara, Lae Sung Kim, and ManYeong Ha. Three-dimensional lattice Boltzmann simulations of droplet formation in a cross-junction microchannel. p. 3 [9] have analyzed the transition from squeezing to dripping
in a T-junction with the phase-field method, which adopts a diffuse interface
approach, justified by the low capillary numbers (dominant effect of surface
tension) in microfluidics, and relating the breakup dynamics to the upstream
pressure variations. [10] have performed three dimensional simulations with the lattice Boltzmann method for a cross-junction configuration,
with a surface tension calculation based on the continuum surface force (CSF)
method, with a good comparison with experiments. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[11] Shahriar Afkhami, Alex M Leshansky, and Yuriko Renardy. Numerical investigation of elongated drops in a microfluidic T-junction. p. 1 Droplet generation via segmentation in microfluidic devices
has the great advantage with respect to batch methods of ensuring excellent
control on size distribution, allowing the generation of almost monodisperse
droplets, and on generation rates, at the cost of lower processed volumes. The
generation of microdroplets via segmentation is generally achieved at a junction of separate channels fed with immiscible fluids, generally oil and an aqueous solution, through different possible regimes, namely squeezing, dripping
and jetting [3]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[12] Shahriar Afkhami and Yuriko Renardy. A volume-of-fluid formulation for the study of co-flowing fluids governed by the Hele-Shaw equations. p. 3 Afkhami and coworkers
have explored the use of Volume-of-Fluid (VOF) methods for the interface
tracking, together with the geometrical height function method for the calculation of curvature and the estimation of surface tension, in a T-junction [11]
and in a co-flowing configuration in a Hele-Shaw cell [12]. A different VOF
implementation has been adopted in another work [13] to study the effect
of the injection angle in a 3-dimensional cross-flow geometry, while the most
widely used numerical tool for such kind of flows probably remains OpenFOAM with its interFoam solver, which is still based on the VOF method,
but with a diffuse interface approach [14, 15, 16]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[13] Wei Yu, Xiangdong Liu, Yuanjin Zhao, and Yongping Chen. Droplet generation hydrodynamics in the microfluidic cross-junction with different junction angles. p. 1 Droplet generation via segmentation in microfluidic devices
has the great advantage with respect to batch methods of ensuring excellent
control on size distribution, allowing the generation of almost monodisperse
droplets, and on generation rates, at the cost of lower processed volumes. The
generation of microdroplets via segmentation is generally achieved at a junction of separate channels fed with immiscible fluids, generally oil and an aqueous solution, through different possible regimes, namely squeezing, dripping
and jetting [3]. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[14] Duong A Hoang, Volkert van Steijn, Luis M Portela, Michiel T Kreutzer, and Chris R Kleijn. Modeling of low-capillary number segmented flows in microchannels using OpenFOAM. p. 2 However, even if the fluid dynamics of such devices is
simple at first sight, given the very low Reynolds numbers involved, which
implies that the flows are inherently laminar and can generally be described
by the Stokes equations, interfaces can complicate the numerical solution
significantly, leading to possible lack of accuracy if not properly handled, or
in the worst case to numerical instabilities, due to the amplification of capillary waves. This latter complication, with an explicit treatment of surface
tension, can only be overcome at the expense of using very small time steps,
3 The
generation of microdroplets via segmentation is generally achieved at a junction of separate channels fed with immiscible fluids, generally oil and an aqueous solution, through different possible regimes, namely squeezing, dripping
and jetting [3]. The analysis of different design configurations, geometries
and response to variations of physical parameters is crucial for the optimization of the devices and to maintain the control, first of all on the production
regime, secondly on droplet volume, production rates and monodispersity. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
Where the paper says it
[15] Duong A Hoang, Volkert van Steijn, Luis M Portela, Michiel T Kreutzer, and Chris R Kleijn. Benchmark numerical simulations of segmented two-phase flows in microchannels using the Volume of Fluid method. p. 8 Accurate computation of the curvature is
carried out using the height-function method [19]. The device geometry is defined in Basilisk through the embedded boundary method (EBM), originally proposed in [25, 26], and widely used in
Basilisk literature for moving boundaries [27], solidification modelling [28],
and two-phase flows with complex geometries [29]. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[16] Maryam Fatehifar, Alistair Revell, and Masoud Jabbari. Non-Newtonian droplet generation in a cross-junction microfluidic channel. p. 5 Modeling and governing equations
We describe the dynamics of droplet generation in a microfluidic device,
where two non-miscible fluids enter from separate channels, come in contact
at a junction, and flow out from the same channel. The physical parameters
characterizing the problem are the viscosities and densities of the two fluids,
µi and ρi
, and surface tension σ. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[17] Sara Tomasi Masoni, Alessandro Mariotti, Chiara Galletti, Roberto Mauri, Maria Vittoria Salvetti, and Elisabetta Brunazzi. Formation of sodium-alginate droplets in an X-microdevice: Characterization of the pinching efficiency. p. 9 of the equilibrium contact angle, which depends on the specific properties of
the fluid and of the solid surface, was not carried out, as it would require
a more detailed characterization of the experimental conditions, which is
beyond the scope of the present work, as the comparison is performed with
an experiment from the literature. For the velocity and pressure fields we set an inlet boundary condition
with constant velocity at the entrance of the channel, a no-slip condition at
the walls, and an outflow condition with fixed pressure at the outlet of the
channel. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[18] David Gabriel Harispe and Pablo A Kler. Accurate numerical prototypes of microfluidic droplet generators with open source tools. p. 0 In this work
we have evaluated the suitability of the open-source software Basilisk to accurately predict microdroplet generation by segmentation. We have validated
the numerical tool with analytical solutions for the dynamics of droplets in
confined flows, namely with Bretherton theory, and by comparison with literature experimental results. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[19] Stéphane Popinet. An accurate adaptive solver for surfacetension-driven interfacial flows. p. 8 Accurate computation of the curvature is
carried out using the height-function method [19]. The device geometry is defined in Basilisk through the embedded boundary method (EBM), originally proposed in [25, 26], and widely used in
Basilisk literature for moving boundaries [27], solidification modelling [28],
and two-phase flows with complex geometries [29]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[20] Stéphane Popinet. Gerris: a tree-based adaptive solver for the incompressible Euler equations in complex geometries. p. 5 ), the additional parameters characterizing the flow are the inlet velocities
of the two fluids, Uc and Ud, and a characteristic dimension of the channels,
typically the diameter, D, in coaxial configurations or the channel height,
H, in square section channels. Based on these physical, geometrical and flow
parameters, four dimensionless groups can be formed: firstly the viscosity
and density ratios, respectively µd/µc and ρd/ρc; secondly the Reynolds and
Capillary number, that for these kind of flows can be defined based on the
continuous phase properties as follows
Rec =
ρcUcD
µc
, Cac =
µcUc
σ
(1)
An alternative, quite common in microfluidics literature, is to replace
one of these two dimensionless number with the Weber number, defined as
follows:
W ec =
ρcU
2
c D
σ
= Rec · Cac (2)
The two fluids are governed by the Navier-Stokes incompressible equations with variable density and viscosity, which in their dimensionless form
6 it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[21] Cyril W Hirt and Billy D Nichols. Volume of fluid (VOF) method for the dynamics of free boundaries. p. 3 [10] have performed three dimensional simulations with the lattice Boltzmann method for a cross-junction configuration,
with a surface tension calculation based on the continuum surface force (CSF)
method, with a good comparison with experiments. Afkhami and coworkers
have explored the use of Volume-of-Fluid (VOF) methods for the interface
tracking, together with the geometrical height function method for the calculation of curvature and the estimation of surface tension, in a T-junction [11]
and in a co-flowing configuration in a Hele-Shaw cell [12]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[22] Grétar Tryggvason, Ruben Scardovelli, and Stéphane Zaleski. Direct Numerical Simulations of Gas–Liquid Multiphase Flows. p. 0 We have validated
the numerical tool with analytical solutions for the dynamics of droplets in
confined flows, namely with Bretherton theory, and by comparison with literature experimental results. We have then performed several campaigns of
numerical simulations for a coaxial device, analyzing the different regimes of
droplet generation, and evaluating how the physical and flow parameters affect the production mechanisms and the diameters of the generated droplets. Analysis and simulations of droplet generation regimes
in a coaxial microfluidic device
Alessio Innocenti∗
, Andrea Poggi, Simone Camarri, Maria Vittoria Salvetti
University of Pisa, Dipartimento di Ingegneria Civile e Industriale, Largo Lucio
Lazzarino, 56122, Pisa, Italy
Abstract
The generation of microdroplets via segmentation in microfluidic devices is
of interest in many applications, from biochemical to pharmaceutical. This
technique permits indeed much higher control on the droplet size, uniformity
and generation rate than in standard batch generation processes.Where the paper says it
Where the paper says it
[23] Eugenio Aulisa, Sandro Manservisi, Ruben Scardovelli, and Stephane Zaleski. Interface reconstruction with least-squares fit and split advection in three-dimensional Cartesian geometry. p. 8 The method embeds a
sharp representation of a solid surface on a regular Cartesian grid. The
reconstruction of the solid interface is done using a piecewise linear reconstruction, using a signed distance function sampled on the vertices of the
Cartesian grid for the geometrical properties. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[24] Jeremiah U Brackbill, Douglas B Kothe, and Charles Zemach. A continuum method for modeling surface tension. p. 3 [9] have analyzed the transition from squeezing to dripping
in a T-junction with the phase-field method, which adopts a diffuse interface
approach, justified by the low capillary numbers (dominant effect of surface
tension) in microfluidics, and relating the breakup dynamics to the upstream
pressure variations. [10] have performed three dimensional simulations with the lattice Boltzmann method for a cross-junction configuration,
with a surface tension calculation based on the continuum surface force (CSF)
method, with a good comparison with experiments. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[25] Hans Johansen and Phillip Colella. A Cartesian grid embedded boundary method for Poisson’s equation on irregular domains. p. 8 The device geometry is defined in Basilisk through the embedded boundary method (EBM), originally proposed in [25, 26], and widely used in
Basilisk literature for moving boundaries [27], solidification modelling [28],
and two-phase flows with complex geometries [29]. The method embeds a
sharp representation of a solid surface on a regular Cartesian grid. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[26] Peter Schwartz, Michael Barad, Phillip Colella, and Terry Ligocki. A Cartesian grid embedded boundary method for the heat equation and Poissons equation in three dimensions. p. 8 The device geometry is defined in Basilisk through the embedded boundary method (EBM), originally proposed in [25, 26], and widely used in
Basilisk literature for moving boundaries [27], solidification modelling [28],
and two-phase flows with complex geometries [29]. The method embeds a
sharp representation of a solid surface on a regular Cartesian grid. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[27] Arthur R Ghigo, Stéphane Popinet, and Anthony Wachs. A conservative finite volume cut-cell method on an adaptive Cartesian tree grid for moving rigid bodies in incompressible flows. p. 8 The
reconstruction of the solid interface is done using a piecewise linear reconstruction, using a signed distance function sampled on the vertices of the
Cartesian grid for the geometrical properties. The discrete operators in the
cut cell are then properly defined with the finite volume method with 2nd
order accuracy [29] for single phase flow / 1st order accuracy when coupled
with the VOF method.Where the paper says it
[28] Alexandre Limare, Stéphane Popinet, Christophe Josserand, Zhonghan Xue, and A Ghigo. A hybrid level-set/embedded boundary method ap33 implied to solidification-melt problems. p. 8 Accurate computation of the curvature is
carried out using the height-function method [19]. The device geometry is defined in Basilisk through the embedded boundary method (EBM), originally proposed in [25, 26], and widely used in
Basilisk literature for moving boundaries [27], solidification modelling [28],
and two-phase flows with complex geometries [29]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[29] Mathilde Tavares, Christophe Josserand, Alexandre Limare, Jose Ma Lopez-Herrera, and Stéphane Popinet. A coupled VOF/embedded boundary method to model two-phase flows on arbitrary solid surfaces. p. 8 The discrete operators in the
cut cell are then properly defined with the finite volume method with 2nd
order accuracy [29] for single phase flow / 1st order accuracy when coupled
with the VOF method. Concerning the boundary condition of the fluid volume fraction at the cut
cell, a particular treatment must be adopted when in the same cell there are
the two liquid phases and the solid, indeed a condition on the contact angle
need to be set [30, 31]. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[30] Shahriar Afkhami, Stephane Zaleski, and Markus Bussmann. A meshdependent model for applying dynamic contact angles to VOF simulations. p. 8 The discrete operators in the
cut cell are then properly defined with the finite volume method with 2nd
order accuracy [29] for single phase flow / 1st order accuracy when coupled
with the VOF method. Concerning the boundary condition of the fluid volume fraction at the cut
cell, a particular treatment must be adopted when in the same cell there are
the two liquid phases and the solid, indeed a condition on the contact angle
need to be set [30, 31]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[31] Dominique Legendre and Marco Maglio. Comparison between numerical models for the simulation of moving contact lines. p. 9 of the equilibrium contact angle, which depends on the specific properties of
the fluid and of the solid surface, was not carried out, as it would require
a more detailed characterization of the experimental conditions, which is
beyond the scope of the present work, as the comparison is performed with
an experiment from the literature. For the velocity and pressure fields we set an inlet boundary condition
with constant velocity at the entrance of the channel, a no-slip condition at
the walls, and an outflow condition with fixed pressure at the outlet of the
channel. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[32] Jose-Maria Fullana, Yue Ling, Stéphane Popinet, and Christophe Josserand. Droplet in micro-channels: A numerical approach using an adaptive two phase flow solver. p. 3 A different VOF
implementation has been adopted in another work [13] to study the effect
of the injection angle in a 3-dimensional cross-flow geometry, while the most
widely used numerical tool for such kind of flows probably remains OpenFOAM with its interFoam solver, which is still based on the VOF method,
but with a diffuse interface approach [14, 15, 16]. It is worth mentioning
that, in parallel, research is also being conducted using commercial codes,
4Where the paper says it
[33] Francis Patton Bretherton. The motion of long bubbles in tubes. p. 9 More
details on geometries and boundary conditions are given in the following
sections. 2-dimensional Taylor bubble
Before studying the droplet generation in a coaxial device, we assess the
accuracy of Basilisk in reproducing the dynamics of a Taylor bubble, since
the literature on Basilisk applications to problems with bubbles/droplets
in microchannels is still very limited to date. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[34] Pascale Aussillous and David Quéré. Quick deposition of a fluid on the wall of a tube. p. 8 For more details on the numerical implementation of
the contact angle condition in Basilisk, see [29]. In all the present simulations,
when the contact between the interface and the wall is present, we have used
an equilibrium contact angle of αw = 30◦
(measured between the wall and the
interface in the continuous phase), which is within the typical range of values
adopted in numerical simulations [17, 16, 15]. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601. 09483v1 [physics.Where the paper says it
Where the paper says it
[35] Wenjie Lan, Shaowei Li, and Guangsheng Luo. Numerical and experimental investigation of dripping and jetting flow in a coaxial microchannel. p. 0 We have then performed several campaigns of
numerical simulations for a coaxial device, analyzing the different regimes of
droplet generation, and evaluating how the physical and flow parameters affect the production mechanisms and the diameters of the generated droplets. Finally we have proposed new scaling laws for the prediction of droplet diameters in the dripping and jetting regimes, refining existing ones by taking
∗Corresponding author
Email address: alessio. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
[36] Andrew S Utada, Alberto Fernandez-Nieves, Howard A Stone, and David A Weitz. Dripping to jetting transitions in coflowing liquid streams. p. 3 Even if not always all these
aspect are considered rigorously, the literature on numerical approaches to
droplet generation in microdevices is rapidly increasing, and different methods and softwares are now commonly used in device design. [7]
have simulated coflowing devices with a front-tracking method [8] to track the
interface between the two immiscible phases, exploiting mainly the dripping
and jetting regimes at different values of the Capillary and Weber numbers. innocenti@unipi. it (Alessio Innocenti)
Preprint submitted to Computers and Fluids January 15, 2026
arXiv:2601.Where the paper says it
Where the paper says it
Green — the paper says it, and the page is printed. Red — no passage carries it, so treat that sentence as unverified. The same check runs on your own text: check a summary against its source, $5.
Improvements for AI systems
-
Accurate prediction of droplet generation regimes in coaxial microfluidic devices can be achieved by evaluating open-source software like Basilisk against analytical solutions from Bretherton theory and experimental results.
-
The improved AI system can analyze physical and flow parameters to predict the different regimes of droplet generation, specifically squeezing, dripping, and jetting, by mapping them onto Capillary number and Weber number spaces.
-
The system can propose new scaling laws for predicting droplet diameters in the dripping and jetting regimes by incorporating additional physical effects such as the viscosity ratio into existing empirical relations.
-
The AI system can accurately determine the equivalent diameter of generated droplets using Basilisk tag functions to calculate area, providing a quantitative measure of droplet size.
-
The improved system can perform regime mapping based on Capillary number and Weber number to identify transition zones between dripping and jetting regimes, which offers valuable insights for device design.
Abstract
The generation of microdroplets via segmentation in microfluidic devices is of interest in many applications, from biochemical to pharmaceutical. This technique permits indeed much higher control on the droplet size, uniformity and generation rate than in standard batch generation processes. In this work we have evaluated the suitability of the open-source software Basilisk to accurately predict microdroplet generation by segmentation. We have validated the numerical tool with analytical solutions for the dynamics of droplets in confined flows, namely with Bretherton theory, and by comparison with literature experimental results. We have then performed several campaigns of numerical simulations for a coaxial device, analyzing the different regimes of droplet generation, and evaluating how the physical and flow parameters affect the production mechanisms and the diameters of the generated droplets. Finally we have proposed new scaling laws for the prediction of droplet diameters in the dripping and jetting regimes, refining existing ones by taking into account additional physical effects, like the viscosity ratio.
Related papers
- Emergent Transfer of a Physics Foundation Model from Simulation to Laboratory Turbulence
- Early onset of secondary shear instability in Kelvin-Helmholtz braids at high Reynolds number
- Rotational covariance, continuous spectrum and non-modal growth of three-dimensional planar weak viscous shock layers
- Deep Learning-based Algebraic Reynolds Stress Closures for RANS Simulations of Turbulent Flows
- Neptuna: A Comprehensive Machine Learning Framework for Benchmarking Complex Multiphase Flows
- Size and shape of condensing steam jets and plumes