Date of Award


Degree Name

MS in Engineering - Materials Engineering


Materials Engineering


Linda Vanasupa


The ability to work with small amounts of fluids is an emerging technology that can greatly benefit the biomedical industry, diagnostics, and society as a whole. Typically, these microfluidic devices are fabricated using polydimethylsiloxane (PDMS). Although it is optically clear, relatively inert, and easy to manipulate, PDMS does have its limitations. These include its tendency to swell when it comes in contact with certain chemicals and its hydrophobicity, which makes it difficult to analyze aqueous samples. Glass is an alternative material that addresses both issues. Etching is used to create these channels in glass. Wet etching procedures are typically isotropic and can lead to contamination. Dry etching is capable of producing anisotropic profiles, which is a desired trait. The purpose of this thesis was to characterize the process for the dry etching of borosilicate glass for microfluidic channels. Etch rate and surface roughness were studied, with partial pressure ratio (SF6:O2) and RF power as the factors. After formulating a DOE, the glass wafers were etched, with aluminum as the etching mask. The etch rate and roughness were measured using a stylus profilometer and an ANOVA was generated to reveal any statistical significance between the treatments. There was a definite increase in etch rate with an increase in the SF6:O2 ratio as an increase in fluorine atoms etched more of the material. An increase in RF power led to an increase in etch rate due to ionic bombardment. From the ANOVA analysis, partial pressure ratio and RF power did not have a significant effect on roughness. This may have been due to the high variability from the small sample size. From the sample means, there may have been a trend present. An increase in SF6:O2 may have led to a higher roughness due to the amount of non-volatile compounds generated as more F atoms were available to react with the surface. For RF power, the sample means suggested that a higher RF power led to a lower roughness. If this were the case, it may have been due to the increase in ionic bombardment which was able to remove the non-volatile products that accumulated on the etched surface. Microscopic images of the etched surface revealed possible damage to the aluminum mask. The cause is unknown and could have occurred from various sources.