Chemically Vapour Deposited (CVD) Silicon Carbide (SiC) is a candidate material for high quality mirror substrates and high quality reflective optics. Structures made of ceramics most often fail due to flaws on the object’s surface. Typically, the flaw is small (atomic dimensions) and cannot be easily detected. Failures occur at stresses that are influenced by the statistical distribution of surface flaws, thus, the relative unpredictability of the failure conditions. The position, relative intensity and bandwidth of Raman spectra are closely related to stacking order (polytypes), crystal size and defects/faults in the structure. E.g., the shift in of about 5-6 cm-1 Raman Stokes line is an indication of the stress and a possibility of crack (or future crack).
The scope of the project is to develop a system for scanning the surface of the sample in horizontal and vertical axes and detection of the possible cracks using Raman technique. The surface scanning bench with x, y & z axes to hold and position the probe, assembling the Raman spectrometer subsystem, data collection and analysis are part of the scope of the project. The system has to be extensible, low cost with high resolution requirements.
Dispersive Raman Spectrographs are considered for this application as this is sufficient and most suitable for such studies. While many off-the-shelf Raman Spectrometers are available from well known sources and come with sophisticated software, housing and control capabilities, these are expensive, and built for specific industries and laboratories and did not suit this application. Thus, it was decided to assemble the spectrometer from readily available modules. This offered flexibility in choosing components and assembling to suit the application needs where the costs can also be optimized with future expansion capability.
The Raman spectra for crystalline, film and bulk form for Si, SiC is between 500 cm-1 to 900 cm-1. A monochromator with a spectral range of about 400-900 cm-1 would meet the requirement. The shift in Raman peaks due to residual stress is expected to be around 5-6 cm-1. However, the resolution required to detect the shift is specified to be about 1 cm-1. The collection optics consists of a Raman probe that would have all the necessary splitters with edge & notch filters built into it.
Standard CCD detector, working at room temperature would meet the requirement as the sensitivity is not very critical in this application.
Like most general Raman applications, a standard Solid state Diode Laser source with high stability, low noise with a Wavelength of 514/532 nm, and a spot size of 2-5 mm wide operating between 10-50mW is considered. Considering the application needs, costs and future expansion options, OceanOptics HR4000 was considered. It was decided that a readily available Raman probe could be used for the initial experimental phases.
The mechanical arrangement for the system consists of an X-Y table that holds the probe and is also capable of vertical movement. The X-Y arms are moved using a stepper motor controlled by a drive and a microcontroller. The commands to move to a position are sent by the PC based on the size of the sample. The probe can be positioned vertically such that it can be very close to the surface to minimize the loss of scattered light.
Data collection is done by the software that can scan based on external triggers. A custom application is also developed to interface with the spectrometer. The data collected will be analyzed to detect the shift in the peak with reference to standard spectra.
An initial prototype was first built to validate the approach. The final system is under construction.