Experiments Aim to Extend the Limits of Magnetic Separation

Magnetic separation has received much attention recently from industrial, medical, and governmental researchers interested in selective separation technologies. Currently, we are developing technologies for single-particle capture of submicron actinide particles from forensic samples of interest. The retrieval and concentration of transuranic and fission products from the local environment or sample sources can also be used to learn about the environment around nuclear facilities.

High-gradient magnetic separation (HGMS) utilizes large magnetic field gradients to effectively separate micron-sized paramagnetic particles of actinide compounds. HGMS can extract and concentrate micron-sized actinides and actinide-containing particulates with no sample destruction. Most host materials such as air, water, and organic matter are diamagnetic, which will allow for a physical separation of the paramagnetic actinides from these materials. Typically, HGMS separators consist of a high-field solenoid magnet, the bore of which contains a fine-structured ferromagnetic matrix material. The matrix fibers locally distort the magnetic field, generating high-field gradients at the surface of the filaments. These areas then become trapping sites for paramagnetic particles and are the basis for the magnetic separation. Using commercially available 45-micron-diameter stainless steel wool, the kaolin clay industry and soil remediation and water treatment processes have demonstrated HGMS of micron-sized materials.

The magnetic capture of smaller particles (submicron) can be obtained by increasing the magnetic force between the fibers and the paramagnetic particles. One approach towards increasing the magnetic force and improving separator efficiency is to reduce the size of the matrix. Others have examined submicron-sized spheres for the collection of nanometer-sized particles. In addition, fractionation methods for HGMS of nanometer particles are under development. We are investigating dendritic growth on surfaces and controlled corrosion of stainless steel wool in an effort to reduce the effective matrix capture diameter.

The matrix materials currently used for HGMS are inhomogeneous and have a complex cross section. In addition, the paramagnetic particles are nonspherical and include a range of particle sizes. All of these factors discourage precise analytical treatment. Therefore, our approach is semiempirical, combining a mechanistic analytical model with empirical input from controlled experiments. Previously we derived a rate model from a force balance on individual paramagnetic particles in the immediate vicinity of a matrix fiber. The model assumes that if the magnetic capture forces are greater than the competing viscous drag and gravity forces, the particle is captured and removed from the flow stream. The rate model depends on the magnitude of the capture cross section (separation coefficient) as determined by the force balance on the particle. In this work we have expanded the model to include the loading behavior of particles on the matrix fibers using the continuity equation, and we utilize capture-rate data from the experimental results. The resulting analytical model predicts the critical separation parameters for submicron, single-particle collection presented in this article. It thus provides a predictive tool for HGMS tests conducted here and can be used for designing both prototype and full-scale units for specific applications. We have conducted a series of HGMS experiments using particulate oxides of copper, praseodymium, or plutonium suspended in a liquid. The experimental data were correlated with the analytical model, which was developed previously with a systematic series of soil remediation tests.

Single-Particle Collection
Assuming a spherical particle with a known density, we can calculate the concentration of a specific-diameter particle. For a concentration of 4 parts per billion (ppb) of a 0.8-micron particle, there would be approximately 1300 plutonium oxide particles in 1 ml of solution. In the initial feed concentrations for our experiments, we typically use starting plutonium concentrations in the range of 2-15 ppb and an initial volume of 100 ml. This would translate to approximately 6500 particles with a 0.8-micron diameter in the feed solution. In these applications, we attempt to extract all the 6500 particles to get below the single-particle threshold.

To obtain a concentration of less than 3 x 10-11 M (~3 ppt) (parts per trillion) of plutonium oxide, the small but finite solubility of PuO₂ in water demanded that we examine alternate carrier fluids. Though the PuO₂ solubility constant is very low, in slightly basic conditions plutonium (IV) in water solutions at pH 8 can be found at concentrations up to 10-7 M. Preliminary HGMS tests conducted in pH 8 water failed to remove plutonium from the feed samples. For the remaining HGMS tests, we have examined dodecane as a carrier fluid to enhance the insolubility of the plutonium. Plutonium oxide particles sized from 0.2-0.8 microns were removed below 6 ppt. The separation efficiencies for these tests were over 99%. This value is the detection limit of the current analytical method (liquid scintillation counting) without a preconcentration step in the sample preparation. The data are in general agreement with the analytical model.

Figure 1. Scanning electron micrograph of extruded stainless steel matrix material. Results of experiments show that maximum separation effectiveness is obtained when the matrix fiber diameter approaches the diameter of the particles to be captured.

Matrix Material
The experiments discussed here used a matrix material based on extruded stainless steel fibers. Figure 1 shows a scanning electron micrograph of the extruded material. The multiple ridges in the material form a continuum of sharp axial edges that generate large magnetic field gradients and serve as trapping sites for the submicron particles. The extruded sample developed for this program contains a much higher density of edge locations compared with traditional HGMS shaved stainless steel wool materials. The extrusions produce small matrix fiber diameters and high matrix loading capacities.

The effect of matrix diameter was examined experimentally by determining HGMS effectiveness for four different matrix materials with different cross-sectional areas. The results indicate that for selective small-particle capture, very fine matrix fibers are required. For particles below 0.2 µm, HGMS separation is much more effective with smaller wire diameters (5 µm). The effect of changing the applied magnetic field for the same matrix size from 7 T to 2.5 T is not nearly as significant as the effect of reducing the matrix size. At 2.5 T the matrix material is nearly saturated magnetically, and any further increase in the applied field has a small effect on the separation. Increasing the matrix residence time will also increase the separation effectiveness, but at extremely low superficial velocities the resulting particle settling can be a significant problem.

To greatly enhance performance and to achieve a single-particle collection threshold, it has become apparent through model predictions and experimental data that the matrix fiber diameter is a crucial separation variable. Tests have been initiated to develop new matrices with submicron-sized fiber diameters. Two different approaches are being studied to generate these materials. One approach is controlled corrosion studies on 25-µm-sized drawn stainless steel wool material, and the second approach is the growth of micron-sized nickel dendrites on stainless steel wool fibers by chemical vapor deposition.

Nickel dendrites were deposited on the stainless steel matrix material using a hot-wall reactor. Nickel dendrite structures were uniformly deposited on the surface throughout the wool structure. The length of the dendrites is 1-2 orders of magnitude smaller than the diameter of the wool fiber substrate, shown in Figure 2. The resulting matrix material has a very high surface area containing fine points that may greatly enhance the high field gradients. In addition to magnetic characterization of the material, HGMS testing on these materials is planned in the future. This material has the potential to advance HGMS towards subnanometer particle collection. We are currently evaluating the HGMS performance of the new materials.

Figure 2. Scanning electron micrograph of extruded stainless steel matrix material with nickel dendrites grown on the surface of the material. This is one approach in the development of new matrices with submicron-sized fiber diameters. The other is controlled corrosion studies on 25-µm-sized drawn stainless steel wool material.

Summary
We have derived a model from a continuity equation that uses empirically determined capture cross-section values. The model allows us to predict high-gradient magnetic separator performance for a variety of materials and applications. The model can be used to optimize the capture cross-section values and thus increase the capture efficiency. Results show that maximum separation effectiveness is obtained when the matrix fiber diameter approaches the diameter of the particles to be captured.

Experimentally, we obtained a single-particle capture limit with 0.8-µm PuO₂ particles with dodecane as a carrier fluid. The finite solubility of plutonium in water prevented the complete removal of the contaminants when using water as the carrier fluid.

The development of new matrix materials is being pursued through the controlled corrosion of stainless steel wool and the deposition of nickel dendrites on the existing stainless steel matrix material. The new materials are promising for the submicron collection of paramagnetic particles. HGMS experiments on the new materials are planned.

This article was contributed by Laura A. Worl. Other researchers on the project are David Devlin, Dallas Hill, and Dennis Padilla, all of NMT-6, and F. Coyne Prenger, (ESA-EPE).

NMT | LANL | DOE
Phone Book | Search | Help/Info

L O S  A L A M O S  N A T I O N A L   L A B O R A T O R Y
Operated by the University of California for the US Department of Energy