The present application is a non-provisional patent application claiming priority to European Patent Application No. EP 15199848.1, filed Dec. 14, 2015, the contents of which are hereby incorporated by reference.
The present disclosure is related to the inspection of patterns of geometric features produced on a semiconductor die by lithography and etching.
Inspection of feature patterns produced by litho and etch technology is helpful in order to detect defects in the applied lithographic masks and/or process defects. Inspection tools are based on making comparisons between two dies on a single wafer, or between a die and a reference pattern.
Many chip designs comprise an array of longitudinal features placed parallel to each other at regular distances thereby defining an array with a given pitch. This may be for example an array of fin-shaped structures for making finFET devices on a memory chip. As the size of such features has decreased below the available resolution of immersion lithography, double patterning techniques such as LELE (Litho-Etch Litho-Etch) or SADP (Self-Aligned Double Patterning) have been developed which allow the production of these feature arrays down to the 14 nm technology node. One problem that is typical to these techniques is known as ‘pitch walking’: the pitch of an array of features is not constant between dies printed on the same wafer. For example for a half pitch of 24 nm in the designed pattern, the pitch may change from the center to the edge of the wafer from 24 to 30 nm. Inspection of such dies may lead to a failure to detect small defects because they are masked by the pitch walking error.
The present disclosure is related to an inspection method as disclosed in the appended claims. The method helps avoid the negative impact of pitch errors.
The present disclosure is related to a method for detection of defects in a printed pattern of geometrical features on a semiconductor die, the pattern comprising an array of features having a nominal pitch, comprising the steps of: (1) determining deviations from the nominal pitch in the printed pattern, and (2) comparing the printed pattern with another version of the pattern, the other version having the same or similar pitch deviations as the printed pattern.
According to an embodiment, the semiconductor die is a first die of a plurality of dies comprising versions of the same printed pattern, and the method further comprises the step of determining deviations from the nominal pitch of the array in each of the plurality of dies, wherein the other version of the pattern is a pattern printed on a second die of the plurality of dies, the second die comprising an array with the same or similar pitch deviations as the first die. In some embodiments, the plurality of dies may be printed on the same semiconductor wafer.
According to an embodiment, an initial pattern is provided having no or minimal deviations from the nominal pitch of the array and the other version of the pattern is a reference pattern obtained by shifting one or more features of the array in the initial pattern, so that the pitch deviations in the reference pattern are the same or similar to the pitch deviations in the printed pattern. According to an embodiment, the initial pattern is a design intent pattern.
According to an embodiment, the semiconductor die is a first die of a plurality of dies, and the method further comprises the step of determining deviations from the nominal pitch of the array in each of the plurality of dies, wherein the initial pattern is a pattern produced on a second die of the plurality of dies, the second die comprising an array with minimal or no deviations from the nominal pitch.
In the latter embodiment, the plurality of dies may be printed on the same semiconductor wafer.
According to an embodiment, the deviations from the nominal pitch are determined by CD-SEM measurements.
The present disclosure is further related to a method for detection of defects in a plurality of dies printed on a semiconductor wafer, each die of the plurality of dies comprising a version of the same printed pattern of geometrical features, the pattern comprising one or more arrays of features having a nominal pitch, wherein
According to an embodiment, one or more of the defects are caused by misalignment of cut masks (10).
According to an embodiment, one or more of the defects are process defects or defects caused by errors in a lithographic mask.
The above-described pitch error problem and the negative consequences for defect detection are first described in more detail on the basis of the accompanying drawings. The upper and lower image of
The upper and lower image of
Some example inspection tools for die-to-die comparison of the lower image of
In the case of the arrays shown in the upper and lower image of
The same problem may occur when dies are compared to the design intent pattern, as illustrated in
Some examples methods of the present disclosure help fix these problems by first determining the actual pitch deviations in an array of features on the die or dies under inspection. This can be done by a technique such as CD-SEM. For example, the distance between neighboring features in the array may be measured at regular intervals along the longitudinal direction of the features, and these measured values may be averaged to result in the average distance between each pair of neighboring features. The pitch deviation can be measured with CD-SEM taking into account a proper anchoring before the measurement, i.e. determining a reference point with respect to which the position of the features is determined. The anchoring can be based on an additional feature used to do pattern recognition. This anchoring ensures the identification and positioning of a structure to start with (structure 1 for example). Any deviation of the position of feature 1 from the reference position is then to be taken into account when determining the pitch deviations.
From these measurements, the pitch deviations can be found between each pair of neighboring features of the array, by comparing the measured pitch with the nominal pitch. The nominal pitch of a regular array of features is the pitch as defined by the design intent pattern. So the nominal pitch is one single number for a particular array, defining, for example, the distance between the center lines of two neighboring features of the array.
In some examples, the pattern under inspection is then compared to another version of the same pattern, having an array with the same or similar pitch deviations as the pattern under inspection. The comparison may result in a defect ranking. Because arrays are compared which have the same or similar pitch deviations, these pitch deviations are no longer dominating the defect ranking, and smaller defects, such as the mask defects 5 and process defect 6 are more likely to be detected. At the same time, the pitch deviations as such are detected because of the preliminary measurement of these deviations. The method therefore allows for detecting pitch deviations caused by pitch walking, for example, while at the same time allowing for the detection of smaller mask or process defects.
According to a first embodiment, the ‘other version’ of the pattern is a printed pattern printed on the same wafer as the pattern under inspection. In some examples, the pitch deviations are measured in the above-described way for a plurality of dies printed on a single wafer. The wafer may for example be a focus and dose modulated test wafer, on which a plurality of versions of the same pattern are printed under varying conditions of the de-focus and dose of the lithography tool. Dies printed at a given de-focus value may then be compared to dies printed at zero de-focus, in order to establish a process window for the tool's focus settings.
According to the first embodiment, the pattern of a particular die under inspection can be compared only to the pattern on a second die with the same or similar pitch deviations as the die under inspection. For example when pitch walking increases from the center of the wafer to the edge of the wafer, dies near the edge are consistently compared to dies that are also near the edge, so that the pitch walking error between compared dies is similar.
As illustrated in
Instead of comparing the pattern to a pattern printed on another die with similar pitch error, the pattern may be compared to a reference pattern, obtained by starting from an initial pattern having no or minimal pitch errors in the array, and shifting the features of the array, so that the pitch deviations in the reference pattern become the same or similar to the pitch deviations observed in the die under inspection. The initial pattern may be the design intent pattern, wherein the features of the array or arrays are shifted so as to correspond with the pitch deviations observed in the pattern under inspection. This approach is illustrated in
The reference pattern may be obtained by shifting features in the design intent pattern as described above. Alternatively, the reference may be obtained by shifting features in an actual pattern printed on a die. For example, when the array 1-4 is printed on one die with no or minimal pitch error, this printed array may be used as the initial pattern, wherein the features on the array in this initial pattern are shifted according to the measured pitch errors on the die under inspection.
In the second case (die-to-design intent), the design intent is adapted before each comparison is made, by shifting one or more of the features in the arrays on the design intent pattern according to the recorded X and Y deviations, so that these arrays have the same pitch deviations as the arrays under inspection in each die. Optionally, a third option may be provided (not shown), in which a die is first selected in which for one or more arrays the pitch deviations are minimum. The shift operation is then performed on contours of printed features of this die prior to each comparison. This third option is therefore identical to the die-to-design intent option, except that an actually printed die is used instead of the design intent, as the starting point before effectuating the shifts.
While examples have been illustrated and described in detail in the drawings and foregoing description, such illustration and description are to be considered illustrative or exemplary and not restrictive. Other variations to the disclosed embodiments can be understood and effected by those skilled in the art in practicing the claims, from a study of the drawings, the disclosure and the appended claims. In the claims, the word “comprising” does not exclude other elements or steps, and the indefinite article “a” or “an” does not exclude a plurality. The mere fact that certain measures are recited in mutually different dependent claims does not indicate that a combination of these measures cannot be used. Any reference signs in the claims should not be construed as limiting the scope.
Number | Date | Country | Kind |
---|---|---|---|
15199848 | Dec 2015 | EP | regional |
Number | Name | Date | Kind |
---|---|---|---|
6121156 | Shamble | Sep 2000 | A |
6185324 | Ishihara | Feb 2001 | B1 |
6301008 | Ziger et al. | Oct 2001 | B1 |
6587581 | Matsuyama | Jul 2003 | B1 |
7248732 | Kuwabara | Jul 2007 | B2 |
8036447 | Hayakawa | Oct 2011 | B2 |
8526710 | Nakagaki | Sep 2013 | B2 |
9170209 | Chang | Oct 2015 | B1 |
20020187582 | Satya | Dec 2002 | A1 |
20040190008 | Mieher et al. | Sep 2004 | A1 |
20060280358 | Ishikawa | Dec 2006 | A1 |
20080241970 | Winkler | Oct 2008 | A1 |
20110311126 | Sakai | Dec 2011 | A1 |
20130140457 | Minekawa | Jun 2013 | A1 |
20150176985 | Shchegrov et al. | Jun 2015 | A1 |
20150233844 | Kohli | Aug 2015 | A1 |
20150268164 | Amir | Sep 2015 | A1 |
20170167992 | Halder | Jun 2017 | A1 |
Number | Date | Country |
---|---|---|
2015136533 | Sep 2015 | WO |
Number | Date | Country | |
---|---|---|---|
20170167992 A1 | Jun 2017 | US |