Tool to provide integrated circuit masks with accurate dimensional compensation of patterns

• US 10,394,984 B2
• Filed: 11/04/2016
• Issued: 08/27/2019
• Est. Priority Date: 11/25/2015
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Rosenbluth, Alan E.
• Primary Examiner(s): Ngo, Brian
• Application Number:
  US15/329,654
• Publication Number:
  US 20170270230A1
• Time in Patent Office: 1,026 Days
• Field of Search: 716 51, 716 52, 716 53, 716 54, 716 55
• CPC Class Codes: G03F 1/36 Masks having proximity correction features; Preparation thereof, e.g. optical proximity correction [OPC] design processes
• CPC Class Codes: G03F 7/70091 Illumination settings, i.e. intensity distribution in the pupil plane or angular distribution in the field plane; On-axis or off-axis settings, e.g. annular, dipole or quadrupole settings; Partial coherence control, i.e. sigma or numerical aperture [NA]
• CPC Class Codes: G03F 7/70283 Mask effects on the imaging process
• CPC Class Codes: G03F 7/705 Modelling or simulating from physical phenomena up to complete wafer processes or whole workflow in wafer productions
• CPC Class Codes: G06F 30/367 Design verification, e.g. using simulation, simulation program with integrated circuit emphasis [SPICE], direct methods or relaxation methods

Abstract

Disclosed are mask definition tools, apparatus, methods, systems and computer program products configured to process data representing a semiconductor fabrication mask. A non-limiting example of a method includes performing a decomposition process on a full Transmission Cross Coefficient (TCC) using coherent optimal coherent systems (OCS) kernels; isolating a residual TCC that remains after some number of coherent kernels are extracted from the full TCC; and performing at least one decomposition process on the residual TCC using at least one loxicoherent system. The loxicoherent system uses a plurality of distinct non-coherent kernel functions and is a compound system containing a paired coherent system and an incoherent system that act in sequence. An output of the coherent system is input as a self-luminous quantity to the incoherent system, and the output of the incoherent system is an output of the loxicoherent system.

Claims

1. A tool configured to process input data, comprising:

an input to receive input data representing integrated circuit shapes within separate mask regions of a semiconductor fabrication mask for use in optical lithography; and

an output to provide output data representing a mask in which dimensions of mask shapes are compensated on the basis of image content in the vicinity of each mask shape when the mask is projected during optical lithography;

the tool being configured to match a partially coherent lithographic image by superposing a sum of images from a set of coherent systems and a sum of images from a set of loxicoherent systems, where the loxicoherent systems include a compound system comprised of a paired coherent system and an incoherent system that act in sequence, with an output of the coherent system being input as a self-luminous quantity to the incoherent system of the pair, and with the output of the incoherent system being an output of the loxicoherent system.

2. The tool as in claim 1, where the set of loxicoherent systems comprises a single loxicoherent system.

3. The tool as in claim 1, where an input for both the set of coherent systems and the set of loxicoherent systems is a frame of integrated circuit shapes comprising a portion of a mask region from the input data.

4. The tool as in claim 1, where lens apertures of the coherent system are Fourier transforms of optimal coherent systems kernels obtained by carrying out an eigendecomposition process on a full Transmission Cross Coefficient.

5. The tool as in claim 4, where lens apertures of constituent coherent and incoherent systems in each compound system of the loxicoherent system set are obtained by isolating a residual transmission cross coefficient that remains after a chosen set of coherent kernels in the coherent system set are extracted from the full transmission cross coefficient;

and then performing at least one decomposition process on the residual Transmission Cross Coefficient using at least one loxicoherent system.

6. The tool as in claim 5, where a first/primary loxicoherent system is selected to match portions of the transmission cross coefficient that are recalcitrant to matching by the chosen set of coherent kernels.

7. A method to process data representing a semiconductor fabrication mask, comprising:

performing a decomposition process on a full Transmission Cross Coefficient (TCC) using optimal coherent system (OCS) kernels;

isolating a residual Transmission Cross Coefficient (TCC) that remains after a chosen number of coherent kernels are extracted from the full Transmission Cross Coefficient (TCC); and

performing at least one decomposition process on the residual Transmission Cross Coefficient (TCC) using at least one loxicoherent system,

where the method comprising matching a partially coherent lithographic image by superposing a sum of images from a set of coherent systems and a sum of images from the at least one loxicoherent system, where the at least one loxicoherent system includes a compound system comprised of a paired coherent system and an incoherent system that act in sequence, with an output of the coherent system being input as a self-luminous quantity to the incoherent system of the pair, and with the output of the incoherent system being an output of the at least one loxicoherent system.

8. The method of claim 7, where performing the decomposition process on the data using the at least one loxicoherent system comprises decomposing the Transmission Cross Coefficient (TCC) into kernels that are separated along axes rotated in a Hopkins domain.

9. The method of claim 8, where the axes are rotated in the Hopkins domain by 45°

.

10. The method as in claim 7, where the loxicoherent system uses a plurality of distinct non-coherent kernel functions.

11. The method as in claim 7, where the coherent OCS kernels are derived from at least one Hopkins bilinear model, and where at least one kernel decomposition is made along an axis that is slanted between two domains of the Hopkins bilinear model.

16. The method as in claim 7, operable in at least one dimension of wafer image data.

12. A method to process data representing a semiconductor fabrication mask, comprising:

performing a decomposition process on a full Transmission Cross Coefficient (TCC) using optimal coherent system (OCS) kernels;

isolating a residual Transmission Cross Coefficient (TCC) that remains after a chosen number of coherent kernels are extracted from the full Transmission Cross Coefficient (TCC); and

performing at least one decomposition process on the residual Transmission Cross Coefficient (TCC) using at least one loxicoherent system,

where a region wherein the intensity frequency is about zero results in a presence of a fin structure in the processed mask data that is approximated as a product of two functions and {tilde over (T)}, where yields a mask filter in a loxicoherent term after further decomposition and {tilde over (T)} yields an intensity filter in the loxicoherent term.

13. A method to process data representing a semiconductor fabrication mask, comprising:

performing a decomposition process on a full Transmission Cross Coefficient (TCC) using optimal coherent system (OCS) kernels;

isolating a residual Transmission Cross Coefficient (TCC) that remains after a chosen number of coherent kernels are extracted from the full Transmission Cross Coefficient (TCC); and

performing at least one decomposition process on the residual Transmission Cross Coefficient (TCC) using at least one loxicoherent system,

where a Transmission Cross Coefficient residual (TCC^(r)) exists after the OCS coherent kernels are subtracted and results in a presence of a fin structure in the processed mask data, and where the at least one loxicoherent system is used to mitigate the presence of the fin structure.

14. A method to process data representing a semiconductor fabrication mask, comprising:

performing a decomposition process on a full Transmission Cross Coefficient (TCC) using optimal coherent system (OCS) kernels;

isolating a residual Transmission Cross Coefficient (TCC) that remains after a chosen number of coherent kernels are extracted from the full Transmission Cross Coefficient (TCC); and

performing at least one decomposition process on the residual Transmission Cross Coefficient (TCC) using at least one loxicoherent system,

where a Transmission Cross Coefficient residual (TCC^(r)) exists after the OCS coherent kernels are subtracted and results in the presence of a fin structure in the processed mask data, further comprising;

expressing TCC^(r)in terms of a mean mask frequency given by f≡

(f₁+f₂)/2 and an intensity frequency given by Δ

f≡

f₁-f₂;

expanding TCC^(r)as a separated series in rotated coordinates f and Δ

f;

expanding into separated terms to enable {tilde over (T)} to be revised as an adjusted function T″

; and

obtaining functions T′

as the mask filter and T″

as the intensity filter to comprise the loxicoherent system.

15. A method to process data representing a semiconductor fabrication mask, comprising:

performing a decomposition process on a full Transmission Cross Coefficient (TCC) using optimal coherent system (OCS) kernels;

isolating a residual Transmission Cross Coefficient (TCC) that remains after a chosen number of coherent kernels are extracted from the full Transmission Cross Coefficient (TCC); and

performing at least one decomposition process on the residual Transmission Cross Coefficient (TCC) using at least one loxicoherent system,

where the loxicoherent system operates based on intensity and mitigates a DC-like impact on a wafer image when two interfering frequencies are equal or about equal resulting in a slope discontinuity in the TCC.

17. An article of manufacture comprising a tangible computer readable medium having information stored therein or thereon, the information being configured to convert and transform a first object, embodied as a starting mask, into a second object, embodied as a final mask, configured to be used during fabrication of semiconductor circuits and structures, where the information is configured to perform a decomposition process on a full Transmission Cross Coefficient (TCC) using optimal coherent systems (OCS) kernels;

isolate a residual Transmission Cross Coefficient (TCC) that remains after a chosen number of coherent kernels are extracted from the full Transmission Cross Coefficient (TCC); and

perform at least one decomposition process on the residual Transmission Cross Coefficient (TCC) using at least one loxicoherent system, and matching a partially coherent lithographic image by superposing a sum of images from a set of coherent systems and a sum of images from the at least one loxicoherent system, where the at least one loxicoherent system include a compound system comprised of a paired coherent system and an incoherent system that act in sequence, with an output of the coherent system being input as a self-luminous quantity to the incoherent system of the pair, and with the output of the incoherent system being an output of the at least one loxicoherent system.