Nanofiber diameter characterization is a tedious work. Generally researchers take a snapshot of nanofiber collection, then use image analysis software (such as ImageJ) to manually measure fiber diameters. Quicker and more reliable techniques should be developed for online quality control.

Here are some other methods:

“Measuring Fiber Diameter Distribution in Nonwovens” Pourdeyhimi,  Textile Research Journal April 1999 vol. 69 no. 4 233-236 (Link)

“Distance transform algorithm for measuring nanofiber diameter” M Ziabari, Korean Journal of Chemical Engineering, Volume 25, Number 4, 905-918 (Link)

“Artificial Vision System for the Automatic Measurement of Interfiber Pore Characteristics and Fiber Diameter Distribution in Nanofiber Assemblies” E Tomba, Ind. Eng. Chem. Res., 2010, 49 (6), pp 2957–2968 (Link)

Update 17.09.2011 -pdate for artificial vision system is published. You can read the preprint here.


Here is one new electrospinning optimization article:

Electrospinning fundamentals: optimizing solution and apparatus parameters


And new review paper:

Technological advances in electrospinning of nanofibers


Qiang Huang is assistant professor at USC IE department and editor of special issue of  IIE Transactions “Quality, Sensing and Prognostics Issues in Nanomanufacturing”. He has two NSF grants related to my blog theme.

First one is “Collaborative Research: Nanostructure Growth Process Modeling and Optimal Experimental Strategies for Repeatable Fabrication of Nanostructures for Application in Photovoltaics”. Total amount awarded is 300.000 $. Here is important sections from abstract:

The research objective of this award is to establish statistics-transformed nanostructure growth process models and efficient experimental strategies for improving process repeatability in the fabrication of nanostructures for the application in photovoltaic cells. To achieve repeatable fabrication of photovoltaic cells with respect to yield (productivity) and uniformity (quality), it is essential to identify and optimize the growth conditions rooted on predictive process models. […] The methodology will be validated through controlled growth of nanowires and fabrication of photovoltaic cells.

Successful completion of the proposed research will lead to new tools and methods for improving process repeatability and yield in nanomanufacturing, particularly in the large scale fabrication of photovaic cells. […] (Emphasis by me)

Second one is called “In Situ Nanomanufacturing Process Control Through Multiscale Nanostructure Growth Modeling” Total amount awarded is 350.000 $. Here is important sections from abstract:

The objective of the proposed research is to generate knowledge of in situ nanomanufacturing process control through multiscale nanostructure growth modeling and growth of metal-oxide nanowires with excellent optical properties. Standard statistical quality control (SQC) faces new challenges of scale effects which is unique to quality control of nanofabrication processes. Particularly, key process variables, varying with location and time, are measured at macro/micro scales. The quality characteristics of nanostructures would better be characterized as space-time random field measured in nanoscale. Relating macroscale process variables to nanoscale critical quality characteristics and defects requires multiscale model integration for in situ process control. The research therefore aims to model nanofabrication process, more specifically, nanostructure growth for in situ quality control in nanomanufacturing. Novel metal-oxide nanowires will be synthesized and characterized for wide applications in nanoscale electronic and optoelectronic devices. […] (Emphasis by me)

I will be following Qiang Huang’s papers.



Today I read paper by Xinwei Deng from University of Wisconsin-Madison titled “Applications of statistical
quantification techniques in nanomechanics and nanoelectronics”. You can download the paper from here.

In the paper Deng talks about the infeasibility of standard characterization methods in nanoscale. For example ordinary tensile strength test cannot be done at the nanoscale, because you material cannot be clamped by the holder without sliding.

Deng argues that statistical methods can be better alternative, and demonstrates his idea in two areas: nanomechanics and nanoelectronics.

Other two related papers are:

Statistical approach to quantifying the elastic deformation of nanomaterials

Quantifying the elastic deformation behavior of bridged nanobelts

Electrospinning process is affected by various parameters. (See my previous post) Most of the papers on parameter effect analysis are focusing on only one parameter. Since process is complex this approach has limited relevance for our goal. Some of the researchers tried to use Design of Experiments (DOE) approach in order to better understand interactions between parameters.

All articles are interested in fiber diameter. Characterization method is random selection of arbitrary number of fibers, measuring their diameter and taking average. Why is this a problem?

So here is the list of papers using DOE:

1) A design of experiments (DoE) approach to material properties optimization of electrospun nanofibres. (SR Coles et al.)

Polymer type: PVOH, PLA

Parameters studied: Conductivity (Salt-No Salt), Concentration (High-Medium-Low), Potential (High-Medium-Low), Collection distance (High-Medium-Low)

Measurements: mean fiber diameter

Conclusion: Large number of unsuccessful runs meant it was impossible to carry out a statistical analysis of this work. Interaction plots were generated for PVOH.

2) Regeneration of Bombyx mori silk by electrospinning. (S. Sukigara et al.)

Polymer type: B. Mori silk fibroin fibres

Parameters studied: Electric field (2-3-4 (kV/cm)), Concentration (12-15-16), Distance (5-7 cm)

Measurements: mean fiber diameter

Conclusion: Contour plots relating fiber diameter to electric field and solution concentration were generated for spinning distances of 5 and 7 cm.

3) Investigation on Process Parameters of Electrospinning System through Orthogonal Experimental Design. (W. Cui et al.)

Polymer type: Biodegradable poly(DL-lactide) (PDLLA)

Parameters studied: electrical voltage (15-20-25 kV), solution concentration (10-20-30 %), polymer molecular weight (50-100-165 kDa), solvent system(acetone, two other mixtures of acetone and chloroform), flow velocity (1.8-5.4-9.0 mL/h), and the needle size of syringe (0.45-0.60-0.80 mm)

Measurements: mean fiber diameter and beads percent

Conclusion: Orthogonal analysis and validation tests are performed. Quantitative equations of regression analysis could be used to prepare electrospun fibers with predetermined diameters and surface morphologies.

4) Effects of electrospinning parameters on polyacrylonitrile nanofiber diameter: An investigation by response surface methodology. (OS Yördem et al.)

Polymer type: Polyacrylonitrile (PAN)

Parameters studied: molecular weight, solution concentration, applied voltage, collector distance.

Measurements: mean fiber diameter

Conclusion: investigation the interactive effects of the parameters on the resultant fiber and to establish a prediction scheme for the domain/window of the parameters where targeted PAN fiber diameter can be achieved.

5) A new approach for optimization of electrospun nanofiber formation process. (M. Ziabari et al.)

Polymer type: Polyvinyl alcohol (PVA)

Parameters studied: solution concentration (%8-10-12), spinning distance (20-10-15 cm), applied voltage (25-20-15 kV), and volume flow rate (0.2-0.3-0.4 mL/h).

Measurements: mean fiber diameter

Conclusion: Full factorial design, 15 tests for evaluating the model, response surface plots were generated.

6) Process Optimization and Empirical Modeling for Electrospun Poly(D,L-lactide) Fibers using Response Surface Methodology (SY Gu et al.)

Polymer type: PDLA

Parameters studied: concentration (3-5-7 %), applied voltage (10-14-18 kV)

Measurements: mean fiber diameter

Conclusion: interactions between parameters is established.

7) Fabrication of electrospun poly(methyl methacrylate) nanofibrous membranes by statistical approach for application in enzyme immobilization (JP Chen et al.)

Polymer type: poly(methylmethacrylate) (PMMA)

Parameters studied: solution concentration (15-20-25-30-35 %), flow rate (0.3-0.6-0.9-1.2-1.5 ml/h), applied voltage (16-18-20-22-24 kV), temperature (10-15-20-25-30 °C), and distance (6-9-12-15-18 cm).

Meausurements: mean fiber diameter

Conclusion: establishing interactions between parameters, validating the model. Calculation of parameters for production of nanofibers with minimum diameter – 36 nm. By providing enormous surface area for C. rugosa lipase immobilization, predesigned PMMA NFM with the minimum fiber diameter could provide an enzyme loading 5.2 times the maximum values reported previously.

8 ) Prediction of water retention capacity of hydrolysed electrospun polyacrylonitrile fibers using statistical model and artificial neural network. (Dev et al.)

Polymer type: polyacrylonitrile (PAN)

Parameters studied: alkali concentration(3-6-9 5%), temperature (50-60-70°C), time (30, 45, 60).

Meausurements: mean fiber diameter

Conclusions: establishment interactions between parameters and verification of the model.

I am tracking the citations to these articles, so as soon as a new article is published I will share it here.

My favorite article is number 7 by JP Chen, because he used the knowledge of interactions to optimize fiber diameter. Other articles just established the interactions.

Nanowire characterization is very important for optimizing the manufacturing process. You cannot improve, if you cannot properly measure something. Right now this process is not standardized, researchers select the “most suitable” portion of the nanowire collection, they measure the desired metric for arbitrary number of nanowires etc. That’s why nanomanufacturing process improvement results can be misleading.

I found an article about this topic: “A Statistics-Guided Approach to Precise Characterization of Nanowire Morphology” Article describes statistical methods for three metrics: length, density, diameter.

Nanowire length

They used three-dimensional geometric model to get a distribution of nanowire length (mean, variance). They measured projected lengths of nanowires for three different tilting angles.

Nanowire density

Nanowire density is usually defined to be the number of nanowires located per unit area. Method they propose: divide the substrate to n1 and also to n2 parts. Randomly pick n1 cells from n12 cells such that one region is selected from each row. Then from each chosen large cell, randomly pick one small cell inside the large cell to perform measurement. After this there is one more selection of cells such that there is one cell selected from each column.

Nanowire diameter

Using same logic as in length, they measure diameter projection for six different angles. For hexagonal nanowires, randomly measured tilt angles are between zero and 180/6 = 30

I will look for papers which cited this paper.

Today I read the paper “Optimizing and Improving the Growth Quality of ZnO Nanowire Arrays Guided by Statistical Design of Experiments“.

High aspect ratio of ZnO nanowires increases antireflectivity of photovoltaic devices. Xu et al. managed to increase aspet ratio from 10 to 23.

Firstly they made 34-2 factorial design. Tested parameters:

  • temperature of furnace
  • growth time
  • zinc concentration
  • capping agent

Some of the runs resulted in poor aspect ratios. Aspect ratio was calculated by measuring lengths and widths of around 20 nanowires using Photoshop. In the second stage they used this information to narrow down focus.

In the third stage they aimed at reducing noise effects (decreasing noise means getting consistent results in each trial with same parameter settings). After that, they had 3 more stages.

I liked the stage by stage refinement of parameters. Will be helpful for my research. What I did not like was aspect ratio calculation process: 20 nanowires is not enough, how they selected these 20 nanowires is not clear.