There is a trend for ever-higher levels of integration which is driven by the demand for enhanced functionality in integrated circuits. This has resulted in the utilization of increasingly three-dimensional structures as a method to gain additional functions from a given area.
This is seen in a number of areas of both interconnect and device technology. The continuing trend for higher aspect ratios in vias, memory capacitor structures, and contacts is one example of this. The adoption of vertically-oriented transistor channels within FinFET logic devices at advanced technology nodes is another example.
The utilization of through-silicon via (TSV) structures to generate 3D interconnects, permits vertical stacking of multiple dies on a much bigger length scale. All three of these developments lead to fresh challenges in metrology and process control, and a common theme should be established between them in order to measure depths and profiles of etched structures.
Diagnostic methods, such as AFM, SEM, and SPM are an important part of the characterizing processes. Optical metrology techniques are highly desired, because they supply quick measurements on product wafers, which permits advanced process control and routine monitoring.
Figure 1. 3D array of square trenches in silicon (left) and corresponding simulated infrared spectra (right), illustrating agreement between RCWA and EMA calculation methods in the long wavelength limit.
Model-Based Infrared Reflectometry
Model-based infrared reflectometry (MBIR) combines a spectral range of 1-20 μm with a photometrically accurate FTIR measurement system, plus a model-based analysis of reflectance spectra from multilayered films and structures.
Unique advantages are provided by the infrared wavelength range for the measurement of 3D-etched structures. These have already been exploited in MBIR applications for a number of structures, including isolation trenches, power device trenches and memory capacitors.
The simplified modeling of complex periodic structures is a key benefit of the infrared metrology. At wavelengths bigger than the pitch of a structure, the light propagates through the structure like it is a homogeneous medium with an effective refractive index.
Effective Medium Approximation
By employing an effective medium approximation (EMA), this can be measured from the geometry of the structure and the refractive indices of its component materials. If the structure parameters – for example, trench width – vary with depth, it is modeled as a stack of many layers, each one characterized by its own effective refractive index.
So, the challenge of modeling the optical response of a complex etched structure can be reduced to the easier challenge of modeling a multilayer stack.
The example presented in Figure 1 illustrates this point. Simulated spectra can be observed for 45 degrees incidence and S-polarization for a number of square trenches etched in silicon. The trenches are 1 micron deep, have a width of 0.125 microns, and have a pitch of 0.25 microns. An exact calculation is shown in the figure utilizing rigorous coupled-wave analysis (RCWA) and a simplified calculation is also shown using a Maxwell-Garnett type EMA.
The “diffraction threshold” is shown by a vertical line corresponding to about 10,000 cm-1 (that is a wavelength of 1 micron), this is where the first transmitted diffraction order emerges. In the infrared range to the left of the threshold, the diffraction is absent. In the short-wavelength range to the right of the threshold there is a particularly complex spectrum which is usual for scatterometry.
Coming from the interference of light reflected from the top and bottom of the trench structure, the spectrum comprises a regular pattern of interference fringes. As shown in Figure 1, the fringes can be approximated accurately by utilizing the EMA method. The structures parameters, such as trench width and depth, are easily established from the amplitude and period of the fringes.
So, the long-wavelength method simplifies the measuring of 3D-etched structures greatly. It is worth noting that the advantage is not just in the ease of modeling, but also the fact that infrared spectra themselves are fairly simple. Their relationship to the parameters of the structure can be understood readily.
This information has been sourced, reviewed and adapted from materials provided by Semilab Semiconductor Physics Laboratory.
For more information on this source, please visit Semilab Semiconductor Physics Laboratory.