NO. SC - 803
SC TECHNOLOGY APPLICATION NOTE
Pre-Exposure Reflectivity Optimization For Improved CD Control
Overview:
Reflective interference from photoresist layers and substrate layers influence the effective exposure energy "coupled" within the photoresist film during the exposure process. In turn, these interference effects play a major role in critical dimension (CD) control for the process.
By controlling the pre-exposure reflectivity during the substrate layer deposition process and the photoresist coat/bake process, significant improvements in CD control can be achieved. Alternatively, the precise exposure energy required to yield the correct CD may be calculated based upon the pre-exposure reflectivity and layer thickness measurements. By these means significant improvements in CD yield can still be realized despite changes in layer thickness and pre-exposure reflectivity inherent to a non-optimized process.
The following discusses the fundamental principles and the practice of pre-exposure reflectivity control using the INS 800 for process control and CD yield improvements.
Photoresist and Semi-transparent Substrate Layer Reflectivity Effects:
Figure 1 illustrates the effects of layer thickness upon reflectivity. Photoresist films and semi-transparent substrate layers (such as polysilicon, oxide, nitride, etc…) exhibit similar trends in reflectivity control with respect to layer thickness.
Figure 1: Percent Reflectivity Versus Layer Thickness
Figure 2 illustrates the correlation of CD and reflectivity with respect to layer thickness. Figure 3 illustrates the general correlation of reflectivity and the exposure energy required to yield the nominal CD relative to layer thickness. The pre-exposure reflectivity is the predominant controlling factor for the resulting CD and the exposure required to yield the correct CD. Layer thickness is the predominant controlling factor for the reflectivity. Therefore, by controlling layer thickness, the pre-exposure reflectivity may be precisely controlled, and hence, the exposure required to yield the correct CD’s.
Figure 2: Correlation Of 0.80µm CD and Pre-Exposure Reflectivity with Layer Thickness
Figure 3:
Relationship of Reflectivity and The Exposure Required To Yield A
0.80µm
CD Versus Layer Thickness
Pre-exposure Reflectivity Optimization: Multi-layer Reflectivity Co-optimization
Generally, processes are designed in a manner to optimize photoresist thickness with respect to exposure energy. In fact, what actually occurs during these process development procedures is the optimization of pre-exposure reflectivity with respect to the effective exposure energy. Generally, when the process is transferred into manufacturing, a specific photoresist thickness and a specific exposure energy are specified for the process. Providing sub-layer thickness does not change, the process will remain stable and will yield the desired CD’s. However, if the sublayer thickness was not optimized to maximize the reflectivity latitude, even very small changes in sub-layer thickness will cause either under-exposure or over-exposure conditions for the process and the process will not yield optimal CD’s.
Many process layers utilize semi-transparent sub-layers in addition to the photoresist layer. For these layers it is imperative to co-optimize the sub-layer thickness in conjunction with the photoresist layer thickness to yield a stable pre-exposure reflectivity which, in turn, will maximize the exposure latitude and the ability to control the process.
Co-optimization is performed in two steps. First, thickness of the semi-transparent sublayer(s) is adjusted to center upon either a maximum or minimum inflection point of the INS 800-measured reflectivity vs. thickness curve (Figure 4). The wavelength used to optimize reflectivity should be either the actinic wavelength (i.e. exposure wavelength) or an integer multiple of the actinic wavelength. For highly reflective substrates, it is advisable to select a minimum reflectivity inflection point to help minimize "notching" and "step coverage necking" effects.
Second, after the optimal thickness has been determined for the sub-layer, the photoresist coating thickness is matched with the sub-layer thickness to center reflectivity at either a maximum or minimum inflection point of the reflectivity versus thickness curve (using optimal sub-layer thickness substrates).
Figure 4: Maximized Process Latitude and Exposure Control Range
Co-optimization of the layer thickness at a maximum or minimum reflectivity inflection point will provide maximum process latitude for: a) sub-layer thickness, b) photoresist layer thickness and c) exposure. Thus, significant improvements in process and CD control are realized.
Exposure Optimization: Compensation for Layer Thickness and Reflectivity Changes
With many established processes, changing layer thickness to optimize reflectivity may not be a practical alternative to improve CD control due to the impact upon existing process specifications and the time required to "requalify" the process. In these cases, it is generally viable to "adjust" exposure energy to optimize the CD yield.
By measuring layer thickness and pre-exposure reflectivity of the combined photoresist and substrate layers(s) prior to exposure, the exposure energy required to yield the correct CD may be determined as shown in Figure 5.
Figure 5:
Exposure Required To Yield the Correct CD and Pre-Exposure Reflectivity
Versus Layer Thickness
When the exposure is adjusted to compensate for "shifts" in pre-exposure reflectivity, the CD yield can be dramatically improved. Alternatively, the photoresist layer thickness may be adjusted to yield a consistent pre-exposure reflectivity (and CD’s), however, this "compensation" process is far more complex and difficult when compared to a simple exposure adjustment even though both have relatively the same CD yield improvement effect.
The primary objective of layer co-optimization of pre-exposure reflectivity is to ensure that the "composite" pre-exposure reflectivity (i.e. substrate layers plus photoresist layer reflectivity) is replicated wafer-to-wafer and lot-to-lot. Processes developed by means of the co-optimization technique will seldom require adjustments to exposure energy to yield the optimal CD. Non-optimized processes, on the other hand, will frequently require exposure energy adjustments to maintain consistent CD’s.