Quality Control InManufacturing of ElectrospunNanofiber CompositesBy Dmitry M. Luzhansky, Principal Engineer, Donaldson Company, Inc.

ORIGINAL PAPER/PEER-REVIEWED

38 INJ Winter 2003

Key words: Nanofibers, electrospinning, composites, quali-ty control, filtration efficiency

IntroductionThe extremely small fiber size and low mass of nanofiber

composites makes quality control measurement very difficult.Most commercially available contact measurement equip-ment cannot work with the nanofiber composite withoutaltering its characteristics. Generally, non-contacting opticalmethods are not effective because the wavelength of the lightused is greater than the diameter of the nanofibers. Non-con-tact Gamma or Beta gauges lack the necessary resolution.Research methods such as Scanning Electron Microscopy(SEM) are not practical in a production setting. Large-scaleproduction of nanofiber composites requires unique qualitycontrol (QC) methods.

Figures 1a, 1b, and 1c show examples of nanofiber compos-ites manufactured by Donaldson.

As Donaldson Company’s nanofiber production hasincreased to over 10,000 m2 per day during the last 20 years,QC methods have been developed to address these needs.Four evaluation tools that have been developed for commer-cial production of nanofibers will be described. These evalu-ation tools focus on: uniformity of nanofiber distribution,fiber size distribution, durability of fiber layer in the compos-ite, and environmental resistance.

Uniformity of nanofiber distributionThe layer of nanofibers in the composite has low solidity

(typically less then 25%), low thickness, is extremely lightweight, and has small fiber size. Existing off-line methods of

measuring filtration efficiency are very accurate, but have sev-eral limitations. Such methods use small sample areas andrequire multiple parallel measurements. High concentrationsof aerosols used in such methods can destroy the sample,making such methods impractical for use in on-line measure-ments. The biggest problem, however, is the intrinsic delaybetween production and test, which makes these methodsunsuitable for process control. Conventional on-line mea-surement methods using different parts of electro-magneticspectra (β-radiation, optical methods) have not workedbecause they do not see the nanofiber, which is smaller thanthe wavelength.

To control the process and product quality, an on-line webfiltration efficiency test method and on-line efficiency monitorwere invented and patented. The test system uses two con-ventional laser particle counters with the addition of a spe-cialized sampling head. The sampling head incorporates anair curtain effect to allow efficiency measurements at speedsup to several hundred feet per minute. This test method canbe applied to nanofiber web development, manufacturingprocess development, and production process control.Details of the on-line efficiency monitor and operating proce-dure are described elsewhere [1,2].

The schematic of the on-line efficiency monitor is presentedin Figure 2. The on-line efficiency monitor includes a sourceof clean air (250 cfm), a source of aerosol, upstream anddownstream particle counters, sampling head, and a PC thatcollects and analyzes data and controls the process.

Whereas off-line methods are very accurate but not usablein production settings, sacrificing a small amount of accuracymakes it possible to use on-line methods to test the efficiency

of each foot of media in a production environment. As shownin Figure 2, the on-line method uses non-contact measurementat low concentrations of aerosol. The apparatus is a subsys-tem of the production line. Analysis of measurements is usedimmediately to control the production process.

The nanofiber layer accounts for most of the filtration effi-ciency in the nanofiber composite. Results of measurementsdescribe quality by assessing the uniformity of the nanofiberlayer in the machine and cross machine directions. Results ofin-control process and out-of-control process testing are pre-sented in Figures 3a and 3b.

Automated Fiber SizingFull process control of the

nanofibers production requiresmeasurement and control of fibersize, fiber size distribution, andquantity of fibers. On-line effi-ciency monitor directly measuresfiltration efficiency, which is afunction of all three parameters.Therefore, there is a need in thetool that can directly measuremean fiber size and fiber size dis-tribution. Routine sampling andmeasurement of fiber size andfiber size distribution are doneusing SEM. Measuring nanofiberdiameters usually consists of

manually comparing the diameter of fibers in a photomicro-graph to a known scale. The process is very time consumingand operator consistency and fatigue can reduce the accuracy.Automating the fiber sizing is a natural solution to the prob-lem. Authomated fiber sizing directly measures fiber diame-ter and fiber diameter distribution, which along with on-lineefficiency monitor allows full process control.

We analyzed several commercially available fiber sizingsoftware packages and found they didn’t meet our require-ments. The automated fiber sizing method we developed isbased on a proprietary algorithm.

In the automated fiber sizing process, the SEM image is firstcropped to a desired size and unwanted details are eliminat-ed. The calibration bar from the SEM image is used to set thenumber of pixels per micron. Next, the program converts theimage to black and white using a gray scale function. Theblack (non-fiber) areas are sorted according to size. Startingwith the largest black area, a straight line is defined on the

39 INJ Winter 2003

Figure 1 EXAMPLES OF NANOCOMPOSITES.A), B) – NANOFIBERS ON THE SUR-

FACE OF FILTER PAPER; C) – NANOFIBERS ON THE SURFACE

OF SPUNBOND NON-WOVEN

Figure 2SCHEMATIC OF ON-LINE EFFICIENCY MONITOR

A B

C

border pixels and a diameter is drawn from one pixel acrossthe white area at 900. The diameter stops when it encountersanother black pixel. This is repeated around the black shape

about every five pixels, and then around each black area inorder of size (Figure 4).

All fiber diameter lengths are recorded and a running his-togram is generated (Figure 5).

The process that used to take hours of painstaking mea-surements can be completed in seconds.

Accelerated Environmental Resistance TestMany nanofiber composites used in filtration are exposed to

hot, humid, and aggressive environments and are often in theapplication for several years. Therefore, long term fiber sta-bility in to such environments becomes a vital element of theirperformance. To “accelerate” environmental testing we mea-sure the filtration efficiency of the nanofiber filter materialusing ASTM test method F1215-89 [3], soak it in hot water for5 minutes, dry, and measure the efficiency again. Likewise,we can measure the resistance of nanofiber composites toaggressive chemical environments including solvent vaporsand oils.

DL Bending Tester [4] Knowledge of strain-stress characteristics is impor-tant in understanding the performance of ananofiber composite under dynamic stress. Due tothe small size of fibers and extremely low weightof the layer, traditional methods do not give usefulresults because the substrates dominate themechanical properties of the nanofiber composite.

The DL bending tester was developed by theauthor to study strain properties and failure mech-anisms of nanofibers applied to the surface ofanother media. Later, this apparatus becameknown internally as the DL bending tester, incor-porating the author’s initials. First a sample issecured in the tester and positioned in the opticalmicroscope. A motor then bends and extends thesample around a cylinder with known diameter.The strain that the sample is close to the plainstrain conditions because the thickness of the sam-ple small enough comparing to diameter of the

40 INJ Winter 2003

Figure 3aIN-CONTROL PROCESS

Figure 5RESULTS OF AUTOMATED FIBER SIZE ANALYSIS

Figure 4NANOFIBER SEM PHOTO WITH FIBER

DIAMETER MEASUREMENT RESULTS

Figure 3bOUT-OF-CONTROL PROCESS

cylinder. The area that is being observed is approximately 6x4mm (4 mm in the direction of stretch) and it can be varied.

The basic schematic of the tester is shown in Figures 6a and6b and the general view of the tester is shown in Figure 7.

Angle can be recalculated into the strain measure:

α - angle of stretch; R – cylinder diameter; L0 – initial length of the specimen.

The angle is measured using the scale on the apparatus. Acamera mounted on the microscope sends a dynamic image toa monitor that is used to observe the sample throughout thetest. The first sign of relative movement between componentsof the composite structure is an indicator of critical strain. Anoperator records what kind of movement occurred and anglesof the first destruction in the nanofiber layer, full destructionof nanofiber layer, and full destruction of the composite.Comparison of the angles for different composites gives themeasure of stability of the structure.

In Figures 8a - 8c are presented frames from such movie. Careful analysis of video helps to understand probable

causes of failures in the composite.

ConclusionsLarge-scale production of nanofiber composites requires

quality and process control. Critical parameters include theuniformity of nanofiber distribution, fiber size distribution,durability of the fiber layer in the composite, and environ-mental resistance. Fundamentally new approaches are

required due to the low solidity, low thickness, extremely lowweight and small fiber size. The methods and instrumentsdescribed herein have shown their relevance and reliability.

AcknowledgementsThe author expresses his gratitude to the Corporate

Technology and Nanofiber Media Business Units ofDonaldson Company and Natick Soldier Center, Soldier,Biological and Chemical Command, for supporting thisresearch.

References1. M. A. Gogins and J. W. Schaefer, “Variability of Filter

Media Efficiency,” Advances in Filtration and SeparationTechnology, vol. 8, American Filtration and SeparationsSociety, Chicago, Illinois, 1994.

41 INJ Winter 2003

Figure 6aSPECIMEN BEFORE DEFORMATION1 - SAMPLE, 2 - JAWS, 3 – CYLINDER

Figure 6bSPECIMEN AFTER DEFORMATION

Figure 7DL BENDING TESTER

2. Gogins M. U.S. Patent 5,203,201, assigned to DonaldsonCompany, Dec. 20, 1991.

3. “Standard Test Method for Determining the InitialEfficiency of a Flatsheet Filter Medium in an Airflow UsingLatex Spheres,” ASTM F1215-89, ASTM International,Philadelphia, 1989.

4. Effort sponsored by the Natick Soldier Center, Soldier,Biological and Chemical Command, USA, under cooperativeagreement number DAAD16-01-3-0001. The U.S. Governmentis authorized to reproduce and distribute reprints forGovernmental purposes notwithstanding any copyright nota-tion thereon. The views and conclusions contained herein arethose of the authors and should not be interpreted as neces-sarily representing the official policies or endorsements, eitherexpressed or implied, of the Natick Soldier Center or the U.S.Government. — INJ

42 INJ Winter 2003

Figure 8SEQUENTIAL VIDEO FRAMES OF A TYPICAL DL BENDING TEST

A) INITIAL FRAME; B) FIRST MOVEMENT AND BREAKAGE IN THE NANOFIBER LAYER (IN OVAL);

C) TOTAL DESTRUCTION OF THE COMPOSITE

A B C