The Influence of Wire Speed on Phase Transitions and Residual Stress in Single Crystal Silicon Wafers Sawn by Resin Bonded Diamond Wire Saw

At a Glance

Metadata	Details
Publication Date	2021-04-14
Journal	Micromachines
Authors	Tengyun Liu, Peiqi Ge, Wenbo Bi
Institutions	Qilu University of Technology, Shandong University
Citations	10
Analysis	Full AI Review Included

Executive Summary

This research investigates the impact of wire speed on subsurface damage (SSD) and residual stress in single crystal silicon (sc-Si) wafers sliced using a Resin Bonded Diamond Wire Saw (RBDWS).

Phase Transition Confirmation: Raman spectroscopy confirmed the generation of amorphous silicon (Si-XII and Si-III phases) in the surface layer of all as-sawn wafers, characterized by distinct Raman peaks at 178.9 cm^-1 and 468.5 cm^-1.
Wire Speed Effect on SSD: Increasing the wire speed significantly increased the ratio of amorphous silicon generated. The depth of the amorphous silicon layer rose from 6.556 nm at 150 m/min to 22.73 nm at 270 m/min.
Wire Speed Effect on Stress: Both residual compressive stress and residual tensile stress increased with higher wire speed, correlating directly with the increased formation of amorphous silicon.
Stress Magnitude: Residual stress levels observed ranged from 0 to 300 MPa (compressive) and 0 to 200 MPa (tensile) across the wafer surfaces.
Slicing Mechanism Implication: The results suggest that higher wire speeds reduce the average cutting force per abrasive, promoting a ductile-regime material removal mode, which favors phase transition and subsequent high internal stress gradients.
Engineering Relevance: Optimizing wire speed is crucial for minimizing wafer warpage, as residual stress is the primary driver for deformation in thin sc-Si wafers.

Technical Specifications

Parameter	Value	Unit	Context
Wafer Material	P-type (111) Single Crystal Silicon	N/A	Ingot diameter: 56 mm.
Wire Saw Type	Resin Bonded Diamond Wire Saw (RBDWS)	N/A	Used on YJXQ120B multi-wire saw machine.
Wire Diameter	95 to 105	µm	Resin bonded wire specification.
Diamond Abrasive Size	8 to 16	µm	Abrasive density: 1200 grits/mm².
Wire Speed Range (Tested)	150, 210, 270	m/min	Three experimental conditions.
Feed Speed (Constant)	0.2	mm/min	Constant parameter across all tests.
Wire Tension (Constant)	12	N	Constant parameter across all tests.
Crystalline Si Raman Peak	521	cm^-1	Cubic diamond phase (Si-I).
Amorphous Si Raman Peaks	178.9 and 468.5	cm^-1	Corresponds to Si-XII and Si-III phases.
Stress-Raman Correlation	±3.2 cm^-1 = ±1 GPa	N/A	Calibration used for residual stress calculation.
Max Residual Compressive Stress	300	MPa	Observed range: 0 to 300 MPa.
Max Residual Tensile Stress	200	MPa	Observed range: 0 to 200 MPa.
Amorphous Layer Depth (150 m/min)	6.556	nm	Lowest observed depth (Raman ratio r=0.13).
Amorphous Layer Depth (270 m/min)	22.73	nm	Highest observed depth (Raman ratio r=0.59).

Key Methodologies

Slicing Equipment and Material: Single crystal silicon ingots (P-type, 56 mm diameter) were sliced using a YJXQ120B multi-wire saw machine utilizing a commercial Resin Bonded Diamond Wire Saw (RBDWS).
Process Parameter Variation: Three distinct wire speeds (150, 210, and 270 m/min) were tested to isolate the speed effect. Feed speed (0.2 mm/min) and wire tension (12 N) were held constant.
Cutting Fluid Application: A cutting liquid (basic solution mixed 1:400 with water) was sprayed from a nozzle to provide lubrication and cooling during the reciprocating sawing motion.
Raman Spectroscopy: A Microscope Raman spectrometer was used to characterize the as-sawn wafers. The system utilized a 633 nm excitation laser (1.23 mW power) and had a sensitivity of ±0.2 cm^-1.
Surface Mapping: Thirteen specific points were measured on each wafer surface along both the feed direction and the wire movement direction to map the residual stress distribution.
Data Analysis and Fitting: Raman spectra were analyzed. Gaussian fitting was applied to amorphous peaks, and the shift from the standard 521 cm^-1 sc-Si peak was used to calculate residual stress (based on the ±3.2 cm^-1 = ±1 GPa correlation).
Amorphous Layer Depth Calculation: The depth (d_a) of the amorphous layer was derived from the Raman Intensity Ratio (r), which compares the surface area under the amorphous silicon peaks to the surface area under the crystalline silicon peak.

Commercial Applications

Photovoltaic (PV) Wafer Production: Direct application in optimizing diamond wire sawing parameters for high-efficiency solar cell substrates, focusing on minimizing subsurface damage (SSD) to reduce subsequent etching requirements and material loss.
Semiconductor Wafer Manufacturing: Relevant for producing ultra-thin silicon wafers where warpage induced by residual stress must be tightly controlled to ensure compatibility with advanced lithography and device fabrication processes.
Brittle Material Processing: The methodology for correlating process speed, phase transition, and residual stress via Raman spectroscopy is transferable to the slicing and grinding of other hard, brittle materials like Silicon Carbide (SiC) and Sapphire.
Equipment Design and Tooling: Provides critical feedback for manufacturers of diamond wire saws and diamond wire, guiding the development of wires and machines optimized for high throughput (speed) without sacrificing wafer quality (low SSD and stress).
Quality Control and Metrology: Establishes Raman spectroscopy as a reliable, non-destructive technique for rapid assessment of subsurface damage and residual stress in as-sawn silicon surfaces.

View Original Abstract

Lower warp is required for the single crystal silicon wafers sawn by a fixed diamond wire saw with the thinness of a silicon wafer. The residual stress in the surface layer of the silicon wafer is the primary reason for warp, which is generated by the phase transitions, elastic-plastic deformation, and non-uniform distribution of thermal energy during wire sawing. In this paper, an experiment of multi-wire sawing single crystal silicon is carried out, and the Raman spectra technique is used to detect the phase transitions and residual stress in the surface layer of the silicon wafers. Three different wire speeds are used to study the effect of wire speed on phase transition and residual stress of the silicon wafers. The experimental results indicate that amorphous silicon is generated during resin bonded diamond wire sawing, of which the Raman peaks are at 178.9 cm−1 and 468.5 cm−1. The ratio of the amorphous silicon surface area and the surface area of a single crystal silicon, and the depth of amorphous silicon layer increases with the increasing of wire speed. This indicates that more amorphous silicon is generated. There is both compressive stress and tensile stress on the surface layer of the silicon wafer. The residual tensile stress is between 0 and 200 MPa, and the compressive stress is between 0 and 300 MPa for the experimental results of this paper. Moreover, the residual stress increases with the increase of wire speed, indicating more amorphous silicon generated as well.

Tech Support

Original Source

DOI: https://doi.org/10.3390/mi12040429

References

2021 - Investigation of rapid composite plating of core wire magnetized electroplated diamond wire saw [Crossref]
2014 - Comparison of diamond wire cut and silicon carbide slurry processed silicon wafer surfaces after acidic texturisation [Crossref]
2016 - Effect of growth rate and wafering on residual stress of diamond wire sawn silicon wafers [Crossref]
2016 - Wire sawing technology: A state-of-the-art review [Crossref]
2016 - Phase and stress evolution in diamond microparticles during diamond-coated wire sawing of Si ingots [Crossref]
2015 - Determination of the impact of the wire velocity on the surface damage of diamond wire sawn silicon wafers [Crossref]
2017 - Mechanisms of material removal and subsurface damage in fixed-abrasive diamond wire slicing of single-crystalline silicon [Crossref]
2012 - Study of ductile-to-brittle transition in single grit diamond scribing of silicon: Application to wire sawing of silicon wafers [Crossref]