By Fred Tepper and Leonid Kaledin

Summary: A form of bioactive nano-alumina fibers attracts and retains virus and other macromolecules by electrostatic forces. These fibers have been incorporated into fibrous (“depth”) filters and were found to have up to a log 7-8 retention for virus even at flow rates several orders of magnitude greater than could be obtained using microporous membranes.

Asmall—about 2 nanometers (nm), or 0.002 microns (µm), in diameter—fiber with a high surface area of roughly 300-600 square meters per gram (m2/g) has been developed using boehmite (AlOOH) as the main component—and has exhibited some unusual filtration capabilities of the fibers.

The large number of hydroxyl (OH-) groups available on the nanofibers generates a positive charge in water solution such that it will attract and retain negatively charged particles including bacteria, virus, organic and inorganic colloids and negatively charged macromolecules. Data are presented on the adsorption capability of the fibers. Filter media has been prepared and tested for virus retention and results are presented. These data show high retention at high flow velocity—about 0.5-1 centimeters per second (cm/sec).

Processing & composition
Bioactive nanofibers are produced in kilogram quantities in three steps:

• Sol-gel reaction
• Filtration of sol from water
• Heating to 200-400°C

The product is a white, free-flowing powder consisting of fibers, again, approximately 2 nm in diameter and tens to hundreds of nanometers long, collected in aggregates. X-ray diffraction shows the fibers are principally boehmite (AlOOH) with minor phases of gamma alumina and aluminum hydroxide, or Al(OH)3. The amount of Al(OH)3 decreases above 200°C as gamma alumina content increases. Beyond approximately 400°C, the principal phase is gamma alumina. The fiber shape is maintained to about 1,050°C, where the principal phase becomes alpha alumina. Figure 1 shows the surface area as a function of heat treatment. A peak in surface area is seen at around 250°C.

Dynamic testing
Our first virus test data are seen in Table 1. The test involved passing 60 milliliters (ml) of water containing 105 plaque-forming units per milliliter (PFU/ml) of virus in buffered pH 7.5 water through 25 millimeter (mm) diameter controls (using Millipore HA 0.4 mm membrane) at a flow rate of 112 ml/minute (min). We then challenged a thin bed (0.7 grams, gm) of the nanofibers that had been spread over the control with the same virus mixture. With a single filter there was sufficient scatter indicating bypass of the sorbent, so further testing was done with a pair of filters. The virus was detected as plaque-forming units using a double-layer assay with E. coli as the host. The results (see Table 1) show very high efficiency for virus retention by the nanofiber vs. the control.

Static ‘capacity’ testing
Experiments were done to determine the degree of attachment of different virus to the bioactive nanofiber. The primary purpose was to determine which of the different process variants best adsorbed virus. Bacteriophages MS-2 and PRD-1 were diluted in 150 ml of 0.02 molar imidazole/0.02M glycine buffer (pH 7.2) to give an effective concentration of approximately 8×105 PFU/ml. The nanofibers (0.5 gm) were placed into a 15-ml conical centrifuge tube containing 10 ml of the bacteriophage solution. The powder was completely dispersed by vortex mixing and was shaken on a rocking table for 15 minutes. The mixture was centrifuged at low speed (2,500 rpm) to collect the powder at the bottom of the tube and the supernatant—the usually clear liquid overlying material deposited by settling, precipitation or centrifugation—was assayed for the presence of virus. An initial and final aliquot (fraction) was taken from the buffer solution containing bacteriophage to verify that virus concentration remained constant throughout the time of the experiment. Table 2 shows that heating to 450°C had no noticeable effect on the attachment of either virus. The data also demonstrate that the fibers will clear virus from solutions to greater than log 2 in one cycle. Experiments are ongoing with respect to determining the static capacity of virus on the nanofiber.

We then prepared fibrous filter media containing modified nanofibers in conjunction with supporting fibers such as cellulose/glass fiber mixes. The tests (see Table 3) involved a solution at pH 7.5 containing approximately 107 PFU/ml virus at a flow of approximately 1 ml/sec through a 2.5-cm diameter filter. The (duplicate) data show that there is greater than 6 log (99.9999 percent) retention starting at approximately 20 weight percent (wt-%) bioactive nanofiber. It’s likely there’s increasing retention capability with loadings greater than 20 wt-%, but the virus assay technique is currently limiting. More recent data show about 8 log retention.

Testing with simulated salt water shows no reduction of virus retention, suggesting such filters may be used as a pre-filter for reverse osmosis (RO) membranes. Virus retention was also not affected when raw sewage was substituted for water in preparing the virus challenge solution.

Filtration mechanisms
The two basic mechanisms for filtration pathogens from water are via sieving and electrokinetic adhesion (adsorption, absorption, etc.). Ultrafilters and RO membranes are generally polymeric and screen particles using pores smaller than the particle. The particles accumulate on the surface rather than in the depth of the filter. The capacity of such filters suffers because they easily clog at the surface. Also, any defect in a surface pore could result in lowering of pressure drop and allow by-pass, compromising the filter.

Depth filters use fibers to create a random array of irregularly shaped pores generally larger than the particle. Thin layers of such filter media are ineffective in filtering small particles. By thickening the filter (many “unit” layers), however, a subsequent layer coincides with the larger pore of the previous layer so the average pore size declined as the depth filter is thickened. For most efficient capture, the particle and fiber diameter (and resultant surface area) should be of similar size. Depth filters generally are capable of removing approximately 85-95 percent of particles, but their use for complete sterilization (e.g., greater than 4-log removal) for virus hasn’t been achieved nor anticipated. Structurally, filters produced from the nanofibers are typical of depth filters. Their fiber diameter is much smaller than previously available fibers, though, so there’s far greater opportunity for interception and capture of small particles.

Streaming potential, which is generated in dilute solutions moving at high velocity through a filter, is useful in determining the filter’s electrostatic properties. Such measurements showed that our fibers are highly positive even in neutral water. The high density of positive charge is the result of the large population of hydroxyl groups on the surface, which are readily distributed across the fiber. Thus, capture is enhanced if the particles are negatively charged. Bacteria and their fragments exhibit negative charge due to their cell wall chemistry, and virus are also negatively charged so they are attracted to and adsorbed by the nano-alumina fibers.

Flow velocity data
One of the concerns with using fibers as small as 2 nm is the expected high pressure drop, particularly during high loading in the depth filter. A series of bioactive nanofiber filters similar to those shown in Table 3 were tested for flow properties at a pressure of 0.2 atmospheres (0.2 bar). Table 4 shows that increasing the loading of nanofibers causes an increase in pressure drop. Nevertheless, the attainable flow rates are substantially greater by about two orders of magnitude as compared to equivalent ultraporous membranes. For instance, Millipore VM (50-nm rating) and VS (25-nm rating) flow velocities are respectively 0.01 and 0.0025 cm/sec at a pressure differential of 0.7 bar. Note that the data were developed at lower 0.2-bar for pressure, so that the flow performance of filters is even more favorable. Part of the reason for the high flow capability of the nanofibers is that it’s hydrophilic, while membranes composed of cellulose esters, plastics and PTFE are generally hydrophobic—water repelling. Water can be forced through hydrophobic filters, but the pressure drop required to force it through pores smaller than 1 µm becomes prohibitively high, and the risk of uneven wetting increases. Thus, flow rates of membrane filters suitable for virus retention are generally in the range of 300-600 liters per square meter per hour (L/m2/hr). Nanofiber filters range between 10,000-20,000 L/m2/hr.

We’re currently studying retention of 30-nm latex spheres and obtaining breakthrough capacity data. These mirror results we’ve gotten with virus. These results also show substantial retention without excessive pressure drop at least to the break point of the filter. Testing is under way with E. coli and 3-µm size latex spheres.

Conclusion
The features and benefits of bioactive nanofibers can be summarized as follows:

• High (about 7 log) virus clearance—Can be increased by thickening filter.
• Higher flow rate—Several orders of magnitude greater than equivalent membrane filters.
• Higher capacity for particulates.
• Higher resistance to clogging.
• Robust—Point defects in a depth filter are compensated by fibers randomly spaced behind such defects, while there’s no such redundancy in a membrane filter.

Suggested applications include:

• Purification of water for residential, medical, dental and military purposes;
• Filter sanitization of medical serums and biological fluids;
• Filter sanitization of other pharmaceutical products;
• Separation of macromolecules such as proteins, DNA, etc., on the basis of charge characteristics of the particles;
• Concentration for the purpose of bio analysis;
• Tissue engineering—Compacts of nanofiber have been shown1 to attract and retain osteoblast (bone) cells and allow their proliferation at higher rates than hydroxyapatite; or
• As a substrate for biosynthesis.

Acknowledgment
The authors would like to thank the U.S. Department of Energy and NASA for their support in the development of these filters.

References

1. Gutwein, L.G., F. Tepper and T.J. Webster, “Increased Osteoblast Function on Nanofibered Alumina,” Proceedings of the 26th International Conference on Advanced Ceramics and Composites, Cocoa Beach, Fla., Jan. 13-18, 2002.

About the authors
Fred Tepper and Dr. Leonid Kaledin are with Argonide Corp., of Sanford, Fla. Argonide is a manufacturer of nano-powders and Nano Ceram™ nano-alumina fibers. Introductory sales of NanoCeram filters are planned for this fall. They can be reached at (407) 322-2500, email: fred@argonide.com or website: www.argonide.com