The Nikon eReview: Nikon Works With Customers and EDA Vendors to Bridge the k1 Gap.

Although the industry has already invented elaborate ways to stretch optical lithography and reduce the k₁ factor, at the recent LithoVision symposium Dr. Stephen Renwick, Principal Engineer at Nikon Precision, highlighted the importance of scanner makers partnering with optical proximity correction (OPC) vendors to provide flexible imaging solutions to bridge the k₁ gap for customers. Dr. Renwick discussed solutions to support various steps of the lithography process, including the Nikon Scanner Signature File (NSSF). This is used to improve OPC accuracy by providing detailed scanner information such as illuminator pupil fill, vectorial lens aberrations, flare, stage information, chromatic aberrations, and laser bandwidth to the OPC vendors.

Figure 1. Nikon is partnering with OPC vendors to provide flexible imaging solutions to bridge the k₁ gap for customers.

A LithoVision 2009 poster presented by Hsu-Ting Huang from Cadence Design Systems demonstrated the benefits associated with the use of NSSF for a 42 nm half-pitch NAND flash layout. Using NSSF version 1.5 and incorporating information specific to a particular scanner model type in the OPC reduced the maximum CD error from 11.91 nm to 1.88 nm and RMSE from 2.8 to 0.69 nm. Edge placement errors greater than 1 nm were also reduced by a factor of 10 by using the NSSF. Additionally, Dr. Jacek Tyminski from Nikon Precision presented a poster studying the impacts of individual scanner-based imaging models at the 45 nm half-pitch and beyond. Dr. Tyminski reported that through the majority of the performance range, predictions of average scanner model type-based and individual scanner-based models differed by 1 to 2 nm. However, near the scanner's resolution limit, accuracy of the imaging model was more affected by the individual scanner signature with differences between the models approaching 5 nm. Tyminski concluded that when imaging aggressive patterns with highly coherent illuminators, signatures of the individual scanners must be considered in OPC tolerancing.

Following LithoVision, Synopsys Inc. announced a joint collaboration with Powerchip Semiconductor and Nikon to deploy the Nikon Scanner Signature Files as a means to increase Proteus ProGen model accuracy on 42 nm flash memory designs. Nelson Lai, OPC Department Manager for the Nano-Printing Technology Group at Powerchip, commented that they "…expect the stepper-specific NSSF parameters can provide the additional modeling accuracy required in this highly competitive memory market."

Renwick also discussed a method for OPE Automatching, which would enable customers to operate a Nikon scanner using the same OPC solution as a competitor's scanner by automatically providing Nikon settings to match the other scanner's through-pitch behavior. The user would input starting information to the software that includes a set of test features (e.g. lines with varying CD through pitch), either a pupilgram from the competitor's tool or resist process information, and LNA/sigma settings of the competitor's tool. Then, making use of Nikon scanner performance, especially in the illuminator, this method would utilize nonlinear optimization to determine the optimal solution. With appropriate software the automatching method is completely self-contained and eliminates the need for customers to release any proprietary information to Nikon.

Figure 2. The OPE Automatching method discussed would enable customers to operate a Nikon scanner using the same OPC solution as a competitor's scanner by automatically providing Nikon settings to match the other scanner's through-pitch behavior.

Renwick finished his presentation with a discussion of Source Mask Optimization (SMO) development efforts. SMO consists of altering the mask and illumination source to optimize imaging performance. Simply stated, this is achieved by using the "good light" and removing the "bad light." Renwick discussed a method that would consist of inputting criteria such as image contrast, edge slope, or process window, in addition to the target features, whereby the SMO system would output the optimized source intensity distribution and mask pattern. Such a method would enable unique illumination solutions to be realized by custom illumination sources.

Figure 3. Source Mask Optimization consists of altering the mask and illumination source to optimize imaging performance. The method described by Renwick would enable unique illumination solutions to be realized by custom illumination sources.

Renwick summarized his presentation noting that solutions such as NSSF, as well as the other methods under development, must be flexible and allow the customer to employ their own preferred vendor or internal software, and stressing that Nikon continues to partner with customers and EDA vendors to provide advanced imaging solutions to overcome next-generation challenges.