Introduction
By far, the most users of CO2 snow cleaning are in the optical fields and applications. CO2 snow cleaning has successfully cleaned individual optical pieces, systems, substrates of all kinds, mirrors, gratings, and other kinds of optics. Cleaning applications include types of glasses, both coated and uncoated, IR and UV optics, Si and Ge based optics, 3-5 and 2-6 optics, diamond turned optics, mirrors, gratings, telescope mirrors, and more. After some examples below, we discuss proper cleaning methods.
1 – Glass Lens - These two before and after images demonstrate CO2 snow cleaning of a small camera lens. Sample size was about 15 mm2. Clips held the sample, with thermal input of 50 - 60 C before and during cleaning. Cleaning time was 3 seconds and working distance was 4 cm. .
Before Cleaning After Cleaning
2 - ITO Coated Glass – In 2012, Wang et al. published a paper on cleaning ITO coated glass substrates for organic photovoltaic cells. In their conclusions, they state: “We have demonstrated that CO2 snow cleaning of ITO-coated glass substrates is effective at removing particles, thereby improving the yield by reducing shorts of large-area OPV cells. Using snow cleaning, 1.44 cm2 devices exhibit a yield of 70% compared to zero yield for solvent-cleaned-only devices.” In a later paper by Xiao et al. in 2015 they state “The discrete cell yield is 100%, mainly owing to the snow-cleaning process that efficiently removes particles on the substrate surface, and therefore reduces the risk of shorting”
3 – X-Ray Optic – Below is a comparison of a before and after cleaning of an optical piece used in X-ray spectroscopy. The optic was heated while cleaning and it took multiple cleanings to remove the stains. The outer rings of water spots are difficult and may take several tries but still not come off. We can, however, offer suggestions.
X-Ray Optic
CO2 Snow Cleaned Uncleaned
4- Diamond-Turned Optics – Metal based diamond-turned optics are easy to clean, requiring possible sample support and occasional thermal input. Increased working distances or a low velocity nozzle may be needed for extremely soft surfaces such as OFHC copper. Manufacturers using CO2 snow cleaning on diamond turned optics do not publicize this application; however, scientists have published more complex examples. Goss et al., and Grazulis et al. used CO2 snow cleaning on diamond machined III-V semiconductors. Goss pointed out that CO2 snow cleaning reduced the need for hazardous cleaning chemicals.
5- Cleaning optics in-situ – For dusting and removing small particles – easy, just increase the working distance to reduce moisture issues. If you need hydrocarbon removal, you may need thermal input (e. g. a hair dryer).
6 - Solvent, Water, and Other Stains – Solvent stain removal is straightforward, and usually requires thermal input on glass samples. CO2 gas or liquid feed work well, though liquid CO2 feed is always faster for stain removal. Issues arise if there is a chemical reactions (monolayer interactions), porosity, or just too thick a residue. Water stains are possible to remove, but not always simple or effective. The central stain areas are usually removed, but the outer ring is the challenge. Sections of the ring come off with repeated tries and, while sometimes 100% removal is achieved, that level is not always possible. Possible chemical interactions between the water stain and the optical substrate or coating can prevent total removal.
Sharpie smiley faces on a silicon wafer before and after cleaning demonstrate organic removal. Thermal input is required. Room temperature liquid CO2 snow feed is a better option here.
7– Large Optics – Particle removal from large optics, such as telescope mirrors, is covered on the telescope page https://www.co2clean.com/telescopes. Similar ideas apply to particle removal from large glass plates or even polymer reel to reel applications.
8- Quality Control - Indirectly, CO2 snow cleaning of coated optics can be a quality control step. This was used for a thin Mg coating on an Al X-Ray anode. A poor coating would delaminate. The same concept will work on thin films on glass.
9 – Optical Mold Cleaning – It has been done. There are papers on the topic, but details and methods are hidden. The process works best on metal molds to remove residual polymers, monomers, and lubricants. Sometimes, a high flow nozzle is needed.
10 – Hydrocarbon Removal – In 1990, Sherman studied hydrocarbon removal from filthy optical neutral and bandpass filters using surface analysis (XPS). Hydrocarbon based reductions were over 50 to 60% and inspection of the filters indicated success. CO2 snow cleaning of new glass plates showed about a 50% hydrocarbon reduction.
Tips for Cleaning Optics -
With proper attention to details, cleaning can be done. There are two issues when you clean glass or other optics – sample support and thermal management. Samples, especially small samples, must be securely held in place. Samples should be cleaned in a dry environment, with a purge gas covering the CO2, or with thermal input from a hot plate or hot air gun. This is needed to avoid moisture condensation caused by the cold CO2 snow – see https://www.co2clean.com/cleaning-issues .
Dusting and removing small particles, is straightforward - increasing the working distance reduces and even eliminates moisture issues. If you need hydrocarbon removal, you may need thermal input like a hair dryer. This method works with all nozzles, though the working distances will vary.
Anti-static devices are useful, as is grounding the sample and their supports.
Potential Damage on Coatings and Substrates
Metal, dielectric, or insulating thin film coatings on glass or other substrates can be cleaned without any damage to the deposited coating. A properly deposited thin film onto a clean substrate will withstand CO2 snow cleaning. The presence of ionic, covalent, and metallic bonds between the coating and substrate implies the coating can withstand cleaning without delamination. Even a self-absorbed monolayer will withstand CO2 snow cleaning.
If the thin film was deposited on a dirty surface, one without proper precleaning, the thin film may be removed. Thin films bound by van der Waals forces can be removed.
Certain crystals are susceptible to thermal shock, though much less than expected. As for glasses, to our knowledge, only one bulk glass has shown thermal effects - NPK52. Bulk fluoride crystals, such as LiF can be damaged, but users have successfully cleaned MgF2 and LiF thin films. For these thin films, a lower velocity nozzle or a greater working distance will assist Testing is required for synthetic fluoride glasses (FPL53) and low velocity nozzles may work. Testing is always suggested for every application.