Cerium Oxide (CeO2) or ceria is a rare earth oxide. In optics manufacturing it is a very popular polishing compound for a variety of different optical materials. 
Although ceria is designated a "rare earth" it is quite abundent. Rare earths make up approximately one-sixth of all naturally occuring elements.  Another name for this group of elements is the lanthanides. Specifically this consists of the 15 elements ranging from Lanthanum (57) to Lutetium (71).  Ceria is the most abundent of this group of elments. 
Pure cerium oxide is rarely ever used for polishing optical materials. Typically polishing compounds contain approximately 50% cerium oxide. The other 50% is usually varying concentrations of other rare earth oxides. Pure cerium oxide is almost white, and the addition of the other rare earth oxides produces color variations. Colors can range from nearly pure white to tan to dark rust brown.  The picture below is a picture of various commercially available cerium oxides to show the variety of colors available. This picture was taken at the Center for Optics Manufacturing.
There are three theories involving the mechanism for removal for glass polishing. They include the mechanical microscopic cutting theory, the flow theory and the chemical reaction theory.  This work is not intended to debate the differences between these theories, but it should be mentioned that polishing with cerium oxide is not a purely chemical or mechanical process. One important attribute of cerium oxide is the lattice defects that are introduced during the manufacturing process. These lattice defects help the "gripping" efficiency of the polishing particles and when in close contact with glass, friction and shear forces strech the ceria's grain surface and cause microfracture. This exposes the defects and the cerium oxide rips out nano size pieces of the glass surface. Continuous milling will continously expose more defects in the particles. 
The motivation behind this work was to determine if any difference could be seen between four batches of the same commerical cerium oxide product. These four batches had four different colors and when used to polish optical glass had varying removal rates. Removal rates are calculated by measuring the amount of material removed per minute. The first and second batches had removal rates similar to each other and to the expected result. Batch three had a removal rate 50% lower than the expected removal rate. Batch four was 20% lower than the expected value. The mean particle size for all of these batches is approximately 1.7mm. The picture below shows the four batches and their varying colors. Batch one is on the left and they are in increasing order.
The cerium oxide particles were adhered to double sided carbon tape that was attached to an aluminum sample stub. A compressed air gun was used to remove any particles that were not securly adhered to the tape. It was found after an initial trial that the particles would have to be sputter coated with gold in order to reduce charging. A picture of the four samples after they had been coated with gold can be found below. The number of dots correspond to the batch number. This identification was necessary to determine the batch number once the samples were placed in the SEM.
Scanning Electron Microscope (SEM) Techniques
Secondary Electron (SE) Imaging
It was found that charging was still occuring even after the particles were sputter coated with gold. In order to reduce this effect the accelerating voltage was decreased to 5kV and the aperature size was reduced to aperture 1. In order to increase resolution a short working distance of 3mm was used. A variety of magnifications were used ranging from 5,000x to 50,000x. The most information can be seen in the 20,000x and 50,000x images. The images were taken by mixing the in-lens SE detector and the in-chamber SE detector. As it was mentioned in the introduction defects in the structure of the particles are very important for the mechanical removal process. Tiny cracks were seen in batches one and two, and to a lesser extent in batch 4. The images for batch 3 did not show these tiny cracks. This feature or lack thereof may be influencing the mechanical polishing effiency of the polishing compound.
20,000x images - The cracks are difficult to see at this magnification.
50,000x images - Cracks can be seen on the particles in batch 1 and 2, whereas batches 3 and 4 look more amorphous
Back Scatter Electron Imaging
Back scatter electron (BSE) imaging is a great way to see contrast between areas of different atomic number. Unfortunately there was very little to see in my samples. It is interesting to see the differences between the particles and the carbon tape. A higher working distance was required for viewing these samples to ensure that the samples would not hit the BSE detector. I used a magnification of 10,000x in order to see multiple particles in contrast with the carbon tape background. These images can be seen below.
This technique was implemented in order to determine if variations in the different rare earth oxides could be seen between batches. Elements can be detected by the collection of the characteristic X-Rays. The software program is designed to automatically label the peaks with the element names. The figure below is an example of the spectrum output. Unfortunately for this project, the same elements were found in all four batches. At the start of this experiment I was assuming that the color variations were due to different rare earth oxides present in each of the samples. This turned out not to be the case. I was able to compare relative intensitites of the different elements, which can be seen below in the pie charts. These results can only be used for relative comparison because the software calculates these values using the assumption that the sample was completely flat. There does not seem to be any correlation between the concentration of any of the elements and the varying color between the batches.
Electron Interaction Modeling
The following is an electron interaction model of a 5kV accelerating voltage electron beam incident on to cerium oxide. This model is of bulk cerium oxide, which is only an approximation of what is occuring for my particles, because not not only are they particles and not a bulk material, but they also are not flat. A 5kV accelerating voltage was used to take all of the SEM images with both the SE and BSE detectors. This interaction modeling can be compared to the figure in our class notes seen below the model. It can be seen by comparing these to figures that the electron interaction model has the same tear drop appearance as what was given to us in the notes. For this project only three types of electrons were detected, the secondary electrons, backscattered electrons and characterisitc X-Rays.
1. H.H. Karow, Fabrication Methods for Precision Optics, John Wiley & Sons, Inc. New York (1993)
2. B.T. Kilbourn, Cerium, A guide to its role in chemical technology, Molycorp, Inc. White Plains NY (1992)
3. T.S. Izumitani, Optical Glass, American Institute of Physics, USA (1986)
4. A. Kaller, Properties of polishing media for precision
optics, Glastech. Ber. Glass Sci. Technol.,71, No. 6, 174-183 (1998)