Date of Award


Embargo Period


Degree Type


Degree Name

Doctor of Philosophy (PhD)


Mathematical Sciences


Burak Ozdoganlar


The demand for miniature devices and parts has increased significantly over the past decade. During this period, research and development of mechanical micromachining processes, including micromilling and microdrilling, have enabled fabrication of intricate three-dimensional micro- and meso-scale components and features on a broad range of materials. One of the key concerns in micromilling is related to the cutting tool. The currently available carbide micro-endmills have the following issues: (a) relatively large cutting edge radius (typically 2 μm - 5 μm) compared to commonly used uncut chip thickness levels, (b) poor tolerance in diameter (typically +/- 10%), and (c) rapid wear when used for machining of hard materials or at high-temperatures. Such non-ideal tools affect the machining process significantly and lead to both poor dimensional accuracy and rough machined surfaces, thus limiting the wide-spread application of micromachining.

This Ph.D. research addresses aforementioned issues by developing a mandrel-based precision polishing process using ultra-high-speed (UHS) miniature spindles, to fabricate single-crystal diamond and tool-grade ceramic micro-endmills which will be superior to the existing carbide microendmills in terms of accuracy and sharpness. The presented work has two specific aspects: The first involves the development of the mandrel-based polishing process and experimental analysis of the polishing characteristics of single-crystal diamond and tool-grade ceramics. And the second involves the design and analysis of precision polishing equipment for the mandrel-based polishing process. Together, these two aspects are aimed to provide experimental understanding of the mandrel-based polishing process and to enable identification of favorable polishing conditions that will allow accurate fabrication of micro-tools from single-crystal diamond and ceramics. A majority of the work is devoted to analyzing the (unwanted) motions of UHS spindles used for the mandrel-based polishing process, with the aim of identifying a favorable set of spindle parameters that would allow for accurate and repeatable fabrication of the micro-tools. This included developing spindle-metrology and analysis techniques applicable to measurement of axial and radial error motions of UHS spindles that currently do not exist in literature.

Initially, the effectiveness of the mandrel-based polishing process in removing single-crystal diamond is demonstrated by polishing and shaping diamond to create smooth surfaces and sharp edges (≤ 1 μm edge radius). Among others, an important issue that was identified in the mandrel-based polishing process was the poor dimensional and form accuracy during material removal. To address this issue, a dual-stage polishing test-bed was designed and constructed to include 1) a large-wheelbased traditional diamond polishing system with high material removal rates for “rough” polishing, and 2) a rigid, mandrel-based polishing configuration with capability to create intricate micro-scalefeatures and high-aspect-ratio structures on single-crystal diamond and ceramics.

Next, polishing characteristics of various tool-grade ceramics were experimentally analyzed to evaluate their applicability for micro-scale cutting. Almost all the ceramic materials tested yielded a better surface roughness than sub-micron grade carbide that is commonly used for micro-tools. All ceramic materials were capable of being sharpened to edge radii less than 2 μm, which is less than the edge radii of sub-micron grade carbide.

One of the most important factors governing the effectiveness of the mandrel-based polishing process in creating accurate features is the speed-dependent axial and radial error motions of the UHS spindle. Undesired motions of the UHS spindles have a direct influence on the dimensional and form accuracy, as well as the surface finish, of the polished surfaces. A thorough quantitative analysis of these motions for the specific UHS spindle used on the dual-stage polishing test-bed is essential to understand their influence on the polishing characteristics. However, there is no existing metrology technique to quantify the error motions of UHS spindles.

To address this need, a laser Doppler vibrometry (LDV)-based methodology was developed to measure the axial and radial error motions of UHS spindles from the surface of a custom-fabricated sphere-on-stem precision artifact. The measured axial and radial motions were post-processed to obtain different components of the error motions, including synchronous and asynchronous components of the axial and radial error motions in both fixed-sensitive and rotating-sensitive directions. The sources and amounts of uncertainties in measuring the motions and in calculating the error motions were then analyzed. The developed methodology is then applied to analyze the radial and axial motions of the electrically-driven hybrid-ceramic-bearing UHS spindle used on the dual-stage polishing test-bed. The measured axial and radial motions were seen to be strongly dependent upon the spindle speed, thermal-state of the spindle, and the over-hang length of the artifact (tool). Certain speed/over-hang length combinations were identified that could potentially induce significant dimensional errors, shape distortions, and surface roughness to the polished surfaces.

The developed UHS spindle-metrology technique was advanced further by implementing errorseparation methods to remove the artifact form error and quantify the true spindle error motions. Two different error separation techniques were developed - Multi-Orientation Technique and a modified Donaldson Reversal Method. Both techniques were successfully demonstrated to remove artifact form error from radial motions measured at speeds up to 150 krpm.

The thesis concludes with a discussion of future work that is needed for successful fabrication of accurate single-crystal diamond and ceramic micro-endmills in a predictable fashion. Specific tasks that should be completed to ensure that the potential high-impact nature of this work is realized have been identified and described in detail.

The specific contributions of this research include: (1) Design and development of a two-stage high-precision polishing test-bed to enable accurate fabrication of micro-scale tool geometries; (2) Development of laser Doppler vibrometry (LDV)-based methodology for measurement of axial and radial error motions when using miniature ultra-high-speed (UHS) spindles; (3) An experimental characterization of the radial and axial error motions of a typical UHS spindle with hybrid-ceramic bearings, identifying the various sources of error motions and quantifying them; (4) Implementation of two different error-separation techniques (Multi-orientation technique and Donaldson reversal method) to remove the artifact form error and obtain the true spindle error motions, and (5) An experimental understanding of the mandrel-based polishing process and the polishing behavior of single-crystal diamond and various tool-grade ceramics.