Overlay errors

The LELE approach can in principle be applied to arbitrary patterns by appropriately splitting into mask levels. While the LELE approach has been demonstrated to work in production environments, it has the disadvantage of doubling the cost of patterning for the critical layers. Another limitation is the effect of overlay error on the space CDs overlay errors translate directly into CD errors, as shown in Figure 3.8. This places severe demands on the overlay accuracy of future exposure tools. Self-aligned spacers for double patterning could avoid this issue. ... 

Dielectric constants of metals, semiconductors and insulators can be detennined from ellipsometry measurements [38, 39]. Since the dielectric constant can vary depending on the way in which a fihn is grown, the measurement of accurate film thicknesses relies on having accurate values of the dielectric constant. One connnon procedure for detennining dielectric constants is by using a Kramers-Kronig analysis of spectroscopic reflectance data [39]. This method suffers from the series-tennination error as well as the difficulty of making corrections for the presence of overlayer contaminants. The ellipsometry method is for the most part free of both these sources of error and thus yields the most accurate values to date [39]. 

Computational accuracy can be dramatically improved by dynamically adding elements where they minimize the error. For example, more elements ean be added in the neighborhood of a strong gradient in the velocity to help resolve shocks and vortex sheets. Elements may be removed from regions of smooth flow to minimize the computational cost without degrading the overall accuracy. The concept is shown in Fig. 9.6 where a finer mesh overlays the original mesh. This mesh refinement can be carried out to as many levels as necessary [15], [16], [17]. 

Some forms of disorder common in adsorbed layers are islands of clustered adsorbates leaving patches of bare substrate domains in which different patches of the overlayer have identical structure but do not match at their junction because of an error in registry, i.e., an error in relative positioning parallel to the surface periodicity errors, in which individual adsorbates do not fit in the periodic arrangement of the surrounding... 

If distortions of the sample are truly isotropic, the only errors that remain after magnification is corrected are residual distortions in the optics, and mask errors. The easiest way to reduce mask errors is to go to 5X or 10X step and repeat systems, but eventually, it should be possible with electron beams to reduce mask errors to insignificant levels. Distortion in the optics of the model 500 is already below 0.1 pm and 3[Pg.15]

As dimensions get smaller it is increasingly important to improve overlay and this places stringent requirements on the distortion of the lenses. The distortion of the best lenses remains at about 0.1 i m and this error contributes significantly to final overlay when different cameras are used for different layers. 

In addition to alignment errors, overlay accuracy depends on mask errors and mask and wafer distortions. Errors and distortions are likely to be relatively serious because the mask is fragile and has the same dimensions as the wafer. With the conventional sources, this problem is somewhat alleviated because isotropic changes in wafer or mask size can be corrected by changing the mask-wafer gap, but adequate overlay for sub-micron devices has yet to be demonstrated. 

To achieve that, an object reconstruction algorithm detects connected components (blobs) in the binary images of each material class. In a first step, blobs that do not meet pre-defined size restrictions are considered as classification errors and filtered out (e.g. the black line due to a faulty camera pixel in object D). The binary image of each material class (material and overlay) is then morphologically dilated... 

Ru (001) substrate in Fig. 2.11 [18J. Each trace corresponds to a different initial silver coverage. Systems of low melting metals such as Cu, Ag and Au, adsorbed on high melting metals such as Ru and W, are very suitable for fundamental desorption studies, because, firstly, readsorption on the sample surface does not occur and, secondly, the rate of desorption is measured without any systematic error due to background pressures, as would be the case with gases such as H2, CO and N2. Metallic overlayers have recently attracted attention as models for bimetallic surfaces in reactivity studies [19,20]. We use the data of Fig. 2.11 to show that TDS gives information on ... 

Follow the similar procedures to overlay two structures for structural comparison. Specify, at least, three similarity atom pairs in order to obtain the acceptable overlay. It is not necessary to give the bond distance between the two atoms in the pair (assume to be zero) for the overlay. Choose Overlay command from the Tools menu to set both RMS error and gradient to 0.01 and click start to initiate the overlay computation. 

---