M.S. Materials Science 14'
Zirconia coated iron powders are currently used as the slurry base in Magnetorheological Finishing (MRF). Recently, alterations to the zirconia coating process of the iron powders have been experimented with. The new coating method uses a higher concentration of zirconia to improve the corrosion resistance of the powder in acidic slurries. This project used measurement techniques focused on highlighting the differences and unique in several zirconia coated samples as well as concentrated zirconia particles extracted from the bulk powder. Through the use of a variety of sample preparation methods and electron microscopy techniques quantifiable differences in atomic composition, coating thickness and particle size were found.
Seven unique powder samples were analyzed. Four samples were selected for having different combinations iron powder size and zirconia coating thickness. The fifth and sixth powders were concentrated zirconia collected from two iterations of magnetic separation from the bulk powder of the seventh sample. Due to the wide range of information that was to be collected from these powders, several samples were prepared using multiple times using different techniques. To investigate particle cross sections and coating thicknesses some samples were polished after being embedded in a thermosetting epoxy. For powder imaging and determination of particle size, a monolayer of sample was spread on adhesive, followed by gold coating. Several samples were carbon coated for quantitative atomic composition analysis and X-Ray mapping, performed with EDaX.
Samples 31-B and HQZr-4 are composed of spherical coated iron powders of about 1-2 microns in diameter. 31-B, prepared using a newer production technique is believed to have a thicker coating of zirconia than HQZr-4. These powders are used as the base magnetic powders in MRF. The goal of investigating these two powders was to quantify difference between them: coating thickness, weight percentage of their components, or even the structure of their exterior. 325-N and 325-O are comprised of much larger, 44 micron, grains of iron and are also coated in zirconia. These two samples were chosen because it was thought that variations in coating may be more easily observed on larger particles, if this proved not to be the case, then possibly information could be gathered about the differences in coating behavior between large and small iron powders. The remaining three samples were collected from one experiment. Powder from the batch of 31-B was used for a magnetic separation experiment, with the goal of determining the quantity of free, or poorly attached zirconia in the bulk power. The separation experiment was performed in two iterations, resulting in two separate zirconia samples and post-separated 31-B. While the free zirconia had been previously identified, its structure and composition had not been observed. It was also of interest to determine if there was a discernible difference between pre and post separated 31-B. The following list summarizes each sample.
* 31-B: micron sized iron powder coated with a high concentration of zirconia
* HQZr-4: micron sized iron powder with a lower concentration of zirconia present during the coating process
* 325-N: 44 micron iron powder with high zirconia concentration
* 325-O: 44 micron iron power with low zirconia concentration
* PS-31-B: post-separation 31-B, the powder remaining after two magnetic separation procedures
* Zr-1: powder with low magnetic potential, collected from the first magnetic separation, expected to have high concentrations of zirconia
* Zr-2: powder collected from the second magnetic separation
Three methods were used, monolayer of powder on sample stubs either coated with gold or carbon, and a dispersion of powder in a thermosetting epoxy heated and compressed to form a puck. The monolayer powder samples were prepared by placing adhesive onto a sample stage, pouring powder onto the adhesive, using compressed air to remove execs powder, and then coating the monolayer to increase conductivity of the sample. These samples were used for: quantitative Energy Dispersive Spectroscopy (EDS), X-ray mapping, and imaging using secondary and backscatter detectors. The surface of the thermosetting epoxy pucks with dispersed powder samples were hand polished, using the lapping polishing machine located in the URnano Sample Prep room #216 of the Wilmot building. Abrasive pads ranging from 15 - 0.2 microns were used. The goal of this polishing procedure was to reveal cross sections of the coated particles, specifically the thickness of the zirconia coating, when viewed under an electron microscope. These polished pucks were coated with gold at 15 mA for 60 seconds to increase conductivity. The following list denotes which preparation methods were used for each sample.
* A low vacuum sputter coating of gold was used to coat each sample. Gold coating has the advantages of being conformal, quick, and easy to apply.
* Monolayer powder with carbon coating: 31-B, HQZr-4, 325-N, and 325-O. These samples were carbon coated to increase the resolution of quantitative mass fraction analysis. A low composition of zirconia in the powders, approximately 2-5%, was seen in preliminary scans of gold coated samples. There was concern that these values were low enough to be influenced by the gold coating. Gold, platinum, and zirconia all produce X-Rays of very similar energies, near 2.1 KeV. It was decided to use carbon coating, with a K alpha1 energy of 277 eV, to circumvent this issue.
* Thermosetting conductive epoxy pucks: 31-B, HQZr-4, 325-N, and 325-O. The zirconia powder samples were excluded due to their very small size, and lack of internal structure. PS-31-B was not included because it was thought that no visible differences would be seen between it and 31-B when viewing single particles.
Micrographs of the 1-2 micron powder samples, shown below in Figures 1 and 2, demonstrate a high variability in particle size and surface conditions. The texture on the surface of the powders comes from the zirconia coating. Due to the high amount of variability in size and coating texture in each of the powders, it is not possible to distinguis between the samples by viewing their exterior.
Figure 1:Micrographs taken with three different detectors, from left to right: an agglomeration of coated particles from sample 31-B taken with the InLens Detector, a backscatter image of particles with large clumps of zirconia from 31-B, and a close-up of particles in sample HQZr-4 taken with the SE2 detector.
Figure 2: Variety in surface structure of coated particles, from left to right: an InLens micrograph of a single particle from 31-B, two particles from powder PS-31-B taken with the SE2, another micrograph of HQZr-4 taken with the InLens detector, note that significant differences are also found in the same powders.
Micrographs of the 44 micron powder, shown in Figure 3, are substantially different from those shown in Figures 1 and 2. There is no consistancy to particle size and a high variability in shape. As expected from the investigation of the smaller powders, there was no way to determine what coating method had been used on the samples from observation of the surface alone.
Figure 3: Micrographs of samples 325-N and 325-O. From left to right: a large field of view showing a typical distribution of particle size and shape, a close up of a particle found in 325-N, a similar micgrograph was collected of a particle in 325-O.
The micrographs of the separated zirconia, Zr-1 and Zr-2, appear to have slightly different structures than the zirconia on the powder. The free zirconia appears to be slightly less angular, note the rounded structres in Figure 4.c below. These two powders are also visually indistinguishable.
Figure 4: Micrographs of samples Zr-1 and Zr-2. From left to right: A small clump of isolated Zr-1, a similar structure found in the Zr-2 sample, a close up of the free zirconia structure.
The pollished surface of the thermosetting epoxy pucks was successful in revealing cross sections of the powders. Finding a cross section of a bisected particle did prove to be difficult, as material removal of the powders only occured in very localized boundary areas of the pucks. The interface between the HQZr-4 (white dots) and large dark areas (Si) shown in Figure 5,a were the best locations to find powder cross sections. This method could be improved through better dispersion of the coated powder into the epoxy, and through a longer and more careful polishing procedure. It should also be noted that even when good cross sections were found, distinguishable coating thickness, like those in Figure 5.b-c, it was difficult to determine the true thickness the zirconia coating appeared to have smeared over the surface of the neighboring Si and Fe. From the few cross sections that were found, 31-B did have a thicker coating than HQZR-4, this agrees with our expectations.
Figure 5: Images of powder cross sections. From left to right: a large field of view on the HQZr-4 puck using BSD to highlight the differences in elements. A coated HQZr-4 particle, diameter and coating thickness were calculated in ImageJ: the diameter is approximately 1.5 microns with a coating thickness of 50 nm. A coated 31-B particle with a diameter of 2 microns and a coating thickness of 150 nm.
The coating of larger particles unfortunately did not aid in distinguishing or determining coating thickness. The unexpected irregular and rough shape of the particles was one part of the problem. Another unexpected issue was the similarity in size to the copper found in the epoxy. This made it surprisingly difficult to be sure that iron was even in the field of view. Even use of the BSD didn't help that much, due to the similarity in atomic weight between copper and iron.
Figure 6: Three cross sections of 325-N particles. No quantifiable data was obtained.
A particle size analysis was performed on PS-31-B using ImageJ. While it was successfully executed, the lack of precision in this calculation limited its use to only this example. Average particle diameter was determined to be 1.71 microns.
Figure 7:Micrographs of the initial image and resulting ImageJ map used to calculate grain size.
EDS of the powders was very succesfull once carbon coatings were applied. The results of these scans did show that the zirconia separation was largly successful, and that 31-B has significantly more zirconia present than HQZR-4. However, there were a few unexpected results, Zr-2 registered having 2% more zirconium than Zr-1 and PS-31-B has slightly more zirconia than 31-B.
Frigure 8: Above are the EDS spectra and qantitative weight% of the elemental composition for each of the powders. Each spectrum was run for two minuts to increase precision of results and normalize them to each other.
The initial hope for X-Ray mapping was to produce maps of particle cross sections showing an iron core and zircionia coating. Due to the large interaction volume of the 20 KeV beam and the small mass fraction of zirconia, this was not achieved. However, some very interesting maps were still produced.
Figure 9: X-Ray map of carbon coated 31-B: from left to right: InLens image, Carbon, Iron, Oxygen, Zirconium.
Figure 11: X-Ray map of carbon coated 325-N: from left to right: SE2 image, Carbon, Iron, Oxygen, Zirconium.
Figure 12: X-Ray map of carbon coated Zr-2: from left to right: SE2 image, Carbon, Iron, Oxygen, Zirconium.
Figure 14: X-Ray map of gold coated 325-N polished puck: from left to right: BSD image of an iron particle cross section embeded in thermosetting epoxy, Au, C, O, Si, Zr.
Created by adjusting the hue and saturation of two images taken of the same location with different detectors. This method can create a very different perspective on the micrograph.
Figure 16: a BSD micrograph over InLens with hue, saturation and opacity adjusted.
Figure 17: An InLens micrograph over a BSD. BSD micrograph colors were inverted while the InLens image had hue and opacity adjusted.
I would like to take this opportunity to thank Brian L. McIntyre for imparting some of his extensive knowledge and expertise to me semester. I would also like to thank Dr. Stephen Jacobs for technical guidance in this project.
Shafrir, Shai., Henry J. Romanofsky, Michael Skarlinski, Mimi Wang, Chunlin Miao, Sivan Salzman, Taylor Chartier, Joni Mici, John C. Lambropoulos, Rui Shen, Hong Yang, and Stephen D. Jacobs. “Zirconia Coated Carbonyl Iron ParticleBased Magnetorheological Fluid for Polishing.” Applied Optics 48.35 (2009): 6796810. Web. 01, Mar. 2014.