[Paper Review] Polishing of {100} and {111} single crystal diamond through the use of Chemical Mechanical Polishing

Evan L. H. Thomas, Soumen Mandal+2 more|arXiv (Cornell University)|Feb 14, 2014

Diamond and Carbon-based Materials Research3 references21 citations

TL;DR

This study demonstrates that Chemical Mechanical Polishing (CMP) using a polyurethane/polyester pad and colloidal silica slurry effectively reduces surface roughness on {100} and {111} single crystal diamond wafers. The method achieves RMS roughness reductions to 0.23 nm and 0.09 nm respectively, without diamond grit or elevated temperatures, offering a simplified, IC-compatible approach to achieving atomically flat, defect-free diamond surfaces for advanced electronic and optical applications.

ABSTRACT

Diamond is one of the hardest and most difficult to polish materials. In this paper, the polishing of {111} and {100} single crystal diamond surfaces by standard Chemical Mechanical Polishing, as used in the silicon industry, is demonstrated. A Logitech Tribo Chemical Mechanical Polishing system with Logitech SF1 Syton and a polyurethane/polyester polishing pad was used. A reduction in roughness from 0.92 to 0.23 nm root mean square (RMS) and 0.31 to 0.09 nm RMS for {100} and {111} samples respectively was observed.

Motivation & Objective

• To develop a low-damage, scalable polishing method for single crystal diamond suitable for high-power and high-frequency electronic and optical devices.
• To address the limitations of traditional mechanical polishing, which induces anisotropic surface damage, nano-grooves, and subsurface fractures.
• To adapt standard Chemical Mechanical Polishing (CMP) techniques from the silicon industry for use on diamond, avoiding high temperatures and diamond grit.
• To evaluate the effectiveness of CMP in achieving atomically flat, defect-free surfaces on both {100} and {111} diamond crystal orientations.
• To characterize the wear and conditioning of polishing pads during prolonged diamond CMP to optimize process longevity and performance.

Proposed method

• Used a Logitech Tribo CMP system with a SUBA-X polyester/polyurethane polishing pad and alkaline colloidal silica slurry (Syton, pH 9.2–10.1, 4–5% ethylene glycol).
• Polished {100} and {111} single crystal diamond samples at room temperature with 4 psi down pressure and 60 rpm rotation in opposite directions.
• Prepared samples via SC-1 cleaning (H₂O₂:NH₄OH:DI H₂O, 1:1:5 at 75 °C for 10 min) followed by ultrasonic DI H₂O bath, then mounted in a polymer holder with cyanoacrylate and Crystalbond.
• Conditioned the polishing pad in-situ using electroplated diamond grit to increase surface asperities and improve slurry distribution and polishing efficiency.
• Performed post-polishing cleaning with SC-1 and HF solutions to remove residual slurry and prevent contamination.
• Characterized surface topography using Atomic Force Microscopy (AFM) over 25 μm² areas, measuring RMS roughness and line traces perpendicular to original polishing grooves.

Experimental results

Research questions

• RQ1 Can standard Chemical Mechanical Polishing (CMP) using colloidal silica and a polyurethane/polyester pad effectively reduce surface roughness on {100} and {111} single crystal diamond?
• RQ2 How does CMP compare to traditional mechanical polishing in terms of surface damage, nano-groove removal, and subsurface defect formation?
• RQ3 What is the impact of diamond’s extreme hardness on polishing pad wear and longevity during prolonged CMP processes?
• RQ4 Can the CMP process be adapted for use with diamond without requiring elevated temperatures or diamond grit, making it compatible with IC fabrication standards?
• RQ5 To what extent does the polishing mechanism differ between {100} and {111} crystal orientations in terms of removal rate and surface quality?

Key findings

• CMP successfully removed nano-grooves from mechanically polished {100} and {111} diamond surfaces, resulting in significantly smoother topographies.
• RMS surface roughness on the {100} plane decreased from 0.92 nm to 0.23 nm after polishing, indicating a 75% reduction in roughness.
• RMS surface roughness on the {111} plane decreased from 0.31 nm to 0.09 nm after polishing, indicating a 71% reduction in roughness.
• The polishing process was effective across both hard and soft mechanical polishing directions, with no evidence of fracture damage or groove reformation.
• Post-polishing cleaning with SC-1 and HF solutions effectively removed residual slurry and debris, ensuring clean, contamination-free surfaces.
• Significant pad wear was observed after 7 hours of polishing, with polyester strands severed and polyurethane foam abraded, indicating the need for in-situ conditioning and pad replacement strategies.