Skip to content

Machine learning algorithms to predict electrospun nanfobiers morphologies

Notifications You must be signed in to change notification settings

JeanSchoeller/SpinLearn

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

27 Commits
 
 
 
 

Repository files navigation

SpinLearn

Introduction:

In this project, we try to predict the diamater, morphologies and electrospinnability of electrospun nanofibers using supervised and unsupervised techniques.

Electrospinning

Electrospinning is a powerful method to produce non-woven fiber meshes/membranes from polymer solutions or polymer melts and has been described in detail by several research groups. As the term electrospinning indicates, fibers are generated by applying a high voltage in the range of kV to overcome the surface tension of a polymer solution. The electrical force causes the pendant polymer droplet at the tip of the spinneret (blunt needle tip) to become charged and the accumulation of charges on the surface of the droplet deforms it into a cone shape referred to as Taylor's cone. Here is a schematic of the electrospinning setup:

Electrospinning setup

When the applied voltage produces a sufficiently strong electric field able to counterbalance the surface tension of the polymer solution, a charged jet is ejected from the needle tip. While the polymer jet travels to the grounded/oppositely charged collector, the solvent evaporates and fibers are randomly deposited on the collector. In this way, a non-woven mesh of fibers with diameters ranging from nano- to micrometers can be generated. These fiber membranes are highly porous materials and their surface to volume ratio is among the highest in material science. Due to this high porosity, e-spun fiber membranes have shown great potential for diverse applications in tissue engineering, drug delivery, solar cells, membranes for environmental bioengineering, chemical sensors and more. By varying process parameters (such as polymer molecular weight, chemical composition of the polymer, surface tension, viscosity, electrical conductivity, the force of the electrical field, the distance between the tip of the spinneret and the collector, the temperature and relative humidity, the type of collector) the morphology of the e-spun constructs can be tailored for specific structure and function. The fiber morphology is highly dependent on the equilibrium between the surface tension and the electrical field. If an equilibrium is not reached, the appa-rition of beads within the construct is inevitable and a broad distribution of the fiber diameter can be observed. The eruption of the jet from the Taylor cone is described as chaotic and pre-sents bending instability. Thus, using a typical electrospinning setup, only nonwoven meshes can be produced. Nevertheless, more ordered constructs can be obtained by using dif-ferent collector morphology. Such collectors are planar collectors, rotating drums or electrode collectors.

Aims and Scope

The many parameters on which electrospinning relies make it difficult to optimize the process. Therefore, we aim to train machine learning algorithms (both supervised and unsupervised learning) to predict the morphologies of electrospun nanofibers. The parameters involved in the electrospinning process with which we will build the training dataset are depicted in the table below:

About

Machine learning algorithms to predict electrospun nanfobiers morphologies

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published