Big Update

Although results from previous trials had shown that 1:1 EtOH/Water solvent gave promising results, more recent trials lead us to believe that a 2:1 ratio of ethanol to water gives more continuous coverage of nanospheres on the surface of the silica film, much of it being continuous monolayer with sporadic interruptions.  A concern which has been previously discussed is the evaporation of the alcohol in the solvent during the sonication and sphere deposition steps.  The greater amount of ethanol in the solvent makes for faster evaporation from the surface of the substrate as the spheres adsorb to the surface.  This leads to the conclusion that faster solvent evaporation leads to better monolayer formation, while at the same time making keeps the volume of the solvent constant.
Evaporation being primarily a mass transport matter has lead me to question the effects of air currents on the layer deposition phase.  To try and reduce the effects of these air currents, as well as any contamination in the air that may be settings on the sample or in the solution during the bead layering, a plastic cover has been fitted for the modified Langmuir-Blodgett trough apparatus.  Since using the cover the streakiness of the sample to the naked eye has been slightly reduced and monolayers have been forming consistently, which sporadic interruptions and multi-layers still, but the results are promising.  While the formation of monolayers is crucial, the formation of multi-layers may not cause problems for the final nano-pillar formation.
It is hypothesized that during the Reactive Ion Etching (RIE) step, only the areas covered in monolayer will allow ions to the surface on the substrate through the gaps between the spheres.  Areas where the beads are absent will be etched as well but will not have nanopillars and should be distinct from areas with nanopillar formation.  Where multi-layers have formed should have enough coverage with spheres that the areas between spheres will be masked by the above layers of spheres, leading to an overall masking of the area and prevention of etching.  We think that these areas will also be distinct from nanopillar regions under AFM and we will be able to test this hypothesis using SEM to see if the overall masking takes place.
Another method of sphere layering was tested for better results.  The method is referred to as the scooping technique (1).  The theory behind it is that by adding the sphere diffusion dropwise to a solution of water and ethanol that a monolayer of the spheres will form on the surface of the solution.  Then   a solution of SDS (Sodium Dodecyl Sulfate) is added to the solution with the floating monolayer to help the crystallization of the monolayer.  Next, the silica substrate that we are using would be dipped horizontally through the surface of the solution and allowed to soak.  Then the substrate is withdrawn at a constant speed and pick up the monolayer which has formed on the surface of the solution and it will adsorb as is to the surface of the substrate.
A few difference between the Langmuir-Blodgett technique and the scooping technique, to clear up any confusion, is that only a few drops of sphere diffusion is added to the solvent solution and the concentration of SDS is much lower because only a few drops of a dilute solution are added.  Then in the LB method the substrate is withdrawn vertically and the monolayer adsorbs at a constant rate and in the scooping method it is supposed to adsorb all at once. 
The technique did produce some samples that had intermittent layer formation but not in the quantity that the LB method produced and therefore the LB technique was used still.
The velocity of the barriers of the LB trough was increased to test their effect on the monolayer formation.  The increased speed of the barriers increased the coverage of spheres, in roughly the same ratio of monolayer to multi-layer, which is good because it meant that we had increased the amount of monolayers with which to create nanopillars without creating an excess of multilayers in the process. 
Four samples of 200 nm spheres on silica were produced, one for SEM prior to RIE and three for use with RIE, using these parameters.
Etching was performed with CF4 and O2. Prior to etching an unlayered sample was cleaned using ultrasonic methanol baths in 5 minute intervals.  This piece was used to determine the etch rate.  The clean sample was etched with an area of the sample covered to prevent etching in this area and the difference in height between the etched and unetched areas were determined using a profilometer and from this the etch rate was determined to be roughly 100 nm/min etching.  The rates of gas were 10 sccm CF4 and 5 sccm O2 at a pressure of 15 mTorr.  The RIE apparatus was a Trion Minilock II.  Power settings were 30 W for RIE and 350 W for inductively coupled plasma.    Images of the pre and post etching were taken on a Nikon Optical Microscope with above incident light.
The images showed that steps to create a cleaner surface prior to sphere deposition need to be taken in order to create a layered sample clear of debris.  A possible source of this debris is the oven in which oxide layer formation was done, as well as the scribing of the silicon wafer. Both will be investigated to determine if either can be modified to provide a cleaner sample.
 One of the layered samples was etched with the same parameters as above for 30 seconds.  The spheres were then removed in a ultrasonic methanol bath.  A second layered sample was etched under the same conditions, but a metal plate was attached that that it hung over half of the sample and it was etched for 5 minutes.  The metal plate blocked the path of direct etching ions but not ones coming from an angle.  This allowed for an etching gradient along the sample.  The idea was that the area of the sample fully exposed would be fully etched and the farther under the metal plate the spheres would be etched less and less, thus providing areas of different etching environments that could help lead us to the conditions that would be optimal for the etching  that we need.  When this sample was removed, it appeared that the spheres that were under the metal plate might have been removed and/or melted from naked eye inspection.  A concern brought up is that using a solution with too much surfactant may prevent even etching due to any etching prevention being caused by a surfactant film being present.  We will be able to do SEM next week on both samples in order to gain a better idea of what may have happened, as well as what conditions we will need and if anything about the process of sphere deposition may need to altered.

1. Y.J. Zhang, Journal of Alloys and Compounds, 450 (2008), 512-516