Study on the Wear Mechanism of a Diamond AFM Tip During Scribing of a Single-Crystal Silicon
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Micromachines |
| Published | 2026-03-12 |
| Authors | X. Sheldon Lin, Litao Qi, Jinguo Han |
| DOI | 10.3390/mi17030344 |
Diamond AFM tip radius reached 2.51 µm after 900 µm of scribing at 100 µm/s and 50 nA. After the same 900 µm distance at 5 µm/s, the tip radius was 1.224 µm with the same 50 nA current. The experiments found that tip wear increased with current, speed, and distance. The simulations identified adhesive transfer and local graphitization as wear mechanisms.
Experimental setup and force calibration
Section titled “Experimental setup and force calibration”Tests used an NTEGRA AFM with a DCP20 diamond probe. The workpiece was a 10 mm × 10 mm × 0.6 mm single-crystal silicon wafer with a [100] direction and (100) plane. All tests used contact mode and feedback through cantilever deflection. The set current was also measured as a conductive-AFM signal, but it did not control feedback. Calibration on silicon dioxide gave sensitivity 13.2 ± 0.3 nm/nA and spring constant 35.48 ± 0.5 N/m. The fitted relation was F = 0.4684I, with F in µN, I in nA, and R² = 0.99. Thus 50 nA corresponded to 23.42 ± 0.94 µN, while 10 nA corresponded to 4.68 ± 0.19 µN. Five high-resolution images were collected at 100× magnification for each worn probe. Five angular measurements were averaged for each probe. Four independent manual scribings were made under the same test conditions.
Measured speed and distance effects
Section titled “Measured speed and distance effects”At a common distance of 300 µm, the experiment gave the following results.
| Scribing speed (µm/s) | Control current (nA) | Scribing distance (µm) | Mean tip radius (µm) |
|---|---|---|---|
| 5 | 50 | 300 | 0.921 |
| 52.5 | 50 | 300 | 1.251 |
| 100 | 50 | 300 | 1.384 |
| 200 | 50 | 300 | 2.021 |
The tip radius increased with distance at every tested speed. At 5 µm/s and 50 nA, tip radius was 0.463 µm after 60 µm. It rose to 0.834 µm after 180 µm and 0.921 µm after 300 µm. At 200 µm/s and 50 nA, tip radius increased from 0.994 µm after 60 µm to 1.404 µm after 180 µm and 2.021 µm after 300 µm. The 200 µm/s series showed the strongest growth with distance. The variation was smaller from 52.5 to 100 µm/s than from 5 to 52.5 µm/s.
Measured current effects
Section titled “Measured current effects”At 52.5 µm/s and 300 µm, tip radius increased with current through most of the tested range.
| Control current (nA) | Scribing speed (µm/s) | Scribing distance (µm) | Mean tip radius (µm) |
|---|---|---|---|
| 15 | 52.5 | 300 | 0.712 |
| 20 | 52.5 | 300 | 0.819 |
| 35 | 52.5 | 300 | 0.950 |
| 40 | 52.5 | 300 | 1.246 |
| 50 | 52.5 | 300 | 1.251 |
The smallest tip radius appeared at 15 nA, and the largest appeared at 50 nA. At 40 and 50 nA, tip radii were 1.246 and 1.251 µm, respectively. The current effect was weak between 40 and 50 nA. Because setpoint current controlled force through the calibration, this series also increased normal load. Across the 60, 180, and 300 µm tests at 15 nA, tip radius rose from 0.454 to 0.712 µm. At 50 nA, it rose from 0.737 to 1.251 µm across the same distances.
Molecular dynamics and the depth threshold
Section titled “Molecular dynamics and the depth threshold”The three-dimensional MD model used a deformable solid cone containing 23,302 carbon atoms. The cone had a 90° tip angle, a height of 50 Å, and a 3.57 Å lattice constant. The silicon block contained 123,981 atoms in a 300 × 90 × 150 Å cell at 298 K, oriented 100. The probe was divided into boundary, 298 K thermostat, and Newtonian layers, with wear occurring in the Newtonian region. Tersoff interactions described carbon bonds and carbon-silicon interactions. The simulation tested indentation depths of 10, 15, 20, 25, and 30 Å. At 10 Å, the tip showed reversible elastic deformation without atomic loss. At 15 Å, bonds broke at the apex and several carbon atoms adhered to silicon. At 20 Å, atomic ejection produced measurable probe loss. At 25 Å, microcracks and widespread shedding marked irreversible damage. At 30 Å, the tip fractured and collapsed plastically. No clusters appeared at 10 Å. Counts were 28 at 15 Å, 115 at 20 Å, 341 at 25 Å, and 742 at 30 Å. Cluster count increased monotonically and served as the model’s wear-severity metric.
Adhesive transfer and local graphitization
Section titled “Adhesive transfer and local graphitization”At the diamond-silicon interface, local distortion and partial coordination loss enabled temporary bonds between carbon and silicon. Atoms then detached from the probe, transferred to silicon, or reattached to carbon sites on the probe. The simulations describe this as adhesive wear, including both interfacial transfer and carbon-carbon reattachment. The Tersoff potential represents these events as a proxy for adhesion. It does not explicitly model electron sharing or orbital hybridization, so the interpretation remains potential-dependent. RDF analysis detected a weak 1.42 Å peak at 15 Å, indicating local bond disorder and the start of sp² bonding. At 20 Å, that peak grew and a 2.46 Å graphite-like neighbor peak appeared. No 3.35 Å graphite interlayer peak appeared through 30 Å, so diamond remained the main phase. Carbon atoms with three neighbors increased from zero at 10 Å to 32 at 15 Å. The count rose to 187 at 20 Å, 1,648 at 25 Å, and 6,360 at 30 Å. This supports local graphitization rather than complete conversion. At fixed 15 Å depth and 100 Å travel, clusters rose from 29 at 50 m/s to 53 at 200 m/s. At 200 m/s, the tip showed the largest radial expansion and a deformed, atomically mixed layer. At fixed 15 Å depth and 100 m/s, clusters rose from 24 to 30 to 38 for 30, 60, and 100 Å. Earlier diamond-tip simulations cited by the authors attributed degradation mainly to thermally assisted graphitization during GaAs and 3C-SiC processing. The present study highlights adhesive transfer in diamond-silicon scribing, while local graphitization develops as indentation deepens.
Measurement checks and model limits
Section titled “Measurement checks and model limits”FE-SEM at 300 µm measured 0.992 µm at 5 µm/s and 1.41 µm at 100 µm/s, both at 50 nA. The optical and FE-SEM values differed by no more than 7.2%. The optical method includes geometric asymmetry from uneven edge wear. The authors accepted current drift below 5% over 300 µm for trend analysis because wear slightly increased contact area. MD used idealized conditions and high strain rates, so it complements rather than directly reproduces experimental wear. Scale, time resolution, and damage criteria also differed between experiment and simulation. The methods table lists simulation speeds of 10, 50, 100, and 150 m/s, while the results also report 200 m/s. The experimental work used the DCP20 probe type and did not compare diamond grades, coatings, or growth methods. Material selection for these probes cannot therefore be resolved by the reported data.
Implications for AFM nanofabrication
Section titled “Implications for AFM nanofabrication”Tip radius affects machining resolution, depth, and subsurface damage, so probe condition belongs in the process record. The tested guidance for paths within 300 µm is segmented scribing or speed no greater than 52.5 µm/s. The simulation results led the authors to identify 15 Å as an upper limit for stable, high-fidelity nano-scribing. Engineers should recalibrate force as contact area grows, because tip wear can shift the current-to-force relation. The study supports parameter control, but it did not establish a superior diamond grade or coating. Each process record should include current, force, speed, path length, and post-test tip radius.
Materials and support
Section titled “Materials and support”Conductive diamond AFM tips for single-crystal silicon scribing need an electrical path and an apex that preserves machining resolution during wear. The diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond with Au, Pt, Pd, Ti, W, or Cu metallization. The engineering support page can be used for custom probe fabrication, but the paper does not specify the DCP20 coating or diamond structure.
View Original Abstract
To elucidate the wear mechanisms of diamond AFM tips during nanoscale scribing of single-crystal silicon, this study combines controlled experiments with atomistic molecular dynamics (MD) simulations. Scribing tests were conducted under systematically varied bias current, scribing speed, and scribing distance. Tip morphology evolution was quantitatively characterized. Concurrently, a three-dimensional MD model reproduced probe-silicon interactions to analyze bond breaking, atomic detachment, and structural transformation at the atomic scale. The results show that increasing current, speed, and distance significantly accelerate tip blunting. Simulations reveal a progressive transition in deformation behavior from elastic response to atomic attrition, plastic damage, brittle cracking, and catastrophic fracture as indentation depth increases, and cluster analysis establishes a quantitative correlation between process parameters and wear severity. This integrated experimental simulation framework provides mechanistic insight into diamond tip degradation and offers quantitative guidance for improving probe durability and process reliability in AFM-based nanofabrication.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →Cvd Diamond Coating On Silicon Wafer
View Specs →
High Concentration Quantum Grade Single Crystal Diamond
View Specs →Original source
Section titled “Original source”References
Section titled “References”- 2008 - Effects of scratching directions on AFM-based abrasive abrasion process [Crossref]
- 2023 - On the deformation mechanism of SiC under nano-scratching: An experimental investigation [Crossref]
- 2010 - Effect of material type and tip radius of AFM probes on nanosheets groove machining accuracy
- 1995 - Wear of the atomic force microscope tip under light load, studied by atomic force microscopy [Crossref]
- 1999 - Deformation and wear of pyramidal, silicon-nitride AFM tips scanning micrometre-size features in contact mode [Crossref]