Mark D. Smith
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Mark D. Smith has a total of 53 co-authors. Here are the top co-authors by number of shared publications:
John J Biafore
KLA Tencor
3 shared publications- Chris A. Mack
Fractilia, Austin, Texas, United States
2 shared publications - Jeff D. Byers
KLA Tencor
2 shared publications Roel Gronheid
IMEC
2 shared publicationsAdy Levy
KLA-Tencor Corp. (United States)
1 shared publicationAlbert Jeans
Hewlett Packard Laboratories, Palo Alto, California, USA
1 shared publication
Latest Publications
- Open Access
- 4 August 2023
- Published by American Vacuum Society in Journal of Vacuum Science & Technology A
- Vol. 41 (5)
Process optimization for shallow trench isolation etch using computational models
The critical dimensions of advanced semiconductor manufacturing processes have decreased to a few tens of nanometers while the aspect ratios have increased beyond 100. The performance of plasma etch patterning processes as well as the cost and time of the development cycle are critical to the success of ramping a new technology node toward profitable high-volume manufacturing. In this paper, a computational patterning software, ProETCH®, has been developed with rigorous physics and advanced algorithms for modeling the etch patterning process, with the featured capabilities in calibrating the reaction mechanisms and optimizing the etch process. A shallow trench isolation etch process using self-aligned double patterning was investigated. A reaction mechanism of silicon etch by Ar/Cl2 plasma was developed to address the surface reactions, and a plasma hypermodel was introduced to correlate process operating conditions to plasma parameters at the wafer surface. The parameters of the reaction mechanism and the plasma hypermodel were calibrated with experimental data obtained from cross-sectional scanning electron microscope (XSEM) images. The calibrated model is used to identify the different fundamental pathways that contribute to the observed profile metrics in XSEMs. The model was then used for process development and optimization by solving the forward and inverse problems. In the forward problem, the model is used to predict the etching profile at different process conditions. Predictions for both interpolation conditions (process parameters within the range used for developing the model) and extrapolation conditions (process parameters outside of the range used for developing the model) agree well with the experimental data with the root mean square error less than 4 nm (1 nm resolution used for the mesh). In the inverse problem, the developed model is used to search for process conditions (e.g., values of bias power and pressure), which could result in desirable profiles. The solutions to the inverse problem demonstrate a degeneracy in process space of the etching process for a given target profile.
- 4 August 2023
- Published by American Vacuum Society in Journal of Vacuum Science & Technology A
- Vol. 41 (5)
Open AccessProcess optimization for shallow trench isolation etch using computational models
The critical dimensions of advanced semiconductor manufacturing processes have decreased to a few tens of nanometers while the aspect ratios have increased beyond 100. The performance of plasma etch patterning processes as well as the cost and time of the development cycle are critical to the success of ramping a new technology node toward profitable high-volume manufacturing. In this paper, a computational patterning software, ProETCH®, has been developed with rigorous physics and advanced algorithms for modeling the etch patterning process, with the featured capabilities in calibrating the reaction mechanisms and optimizing the etch process. A shallow trench isolation etch process using self-aligned double patterning was investigated. A reaction mechanism of silicon etch by Ar/Cl2 plasma was developed to address the surface reactions, and a plasma hypermodel was introduced to correlate process operating conditions to plasma parameters at the wafer surface. The parameters of the reaction mechanism and the plasma hypermodel were calibrated with experimental data obtained from cross-sectional scanning electron microscope (XSEM) images. The calibrated model is used to identify the different fundamental pathways that contribute to the observed profile metrics in XSEMs. The model was then used for process development and optimization by solving the forward and inverse problems. In the forward problem, the model is used to predict the etching profile at different process conditions. Predictions for both interpolation conditions (process parameters within the range used for developing the model) and extrapolation conditions (process parameters outside of the range used for developing the model) agree well with the experimental data with the root mean square error less than 4 nm (1 nm resolution used for the mesh). In the inverse problem, the developed model is used to search for process conditions (e.g., values of bias power and pressure), which could result in desirable profiles. The solutions to the inverse problem demonstrate a degeneracy in process space of the etching process for a given target profile.
- 24 March 2016
- Published by SPIE-Intl Soc Optical Eng
- Vol. 9778, 97781L-97781L-14
Lithography aware overlay metrology target design method
- 24 March 2016
- Published by SPIE-Intl Soc Optical Eng
- Vol. 9778, 97781L-97781L-14
Lithography aware overlay metrology target design method
- 17 March 2011
- Published by SPIE-Intl Soc Optical Eng
- Vol. 7969, 796904-796904-11
EUV secondary electron blur at the 22nm half pitch node
In this paper the Arrhenius behavior of blur upon EUV exposure is investigated through variation of the PEB temperature. In this way, thermally activated parameters that contribute to blur (such as acid/base diffusion) can be separated from non-thermally activated parameters (such as secondary electron blur). The experimental results are analyzed in detail using multi-wavelength resist modeling based on the continuum approach and through fitting of the EUV data using stochastic resist models. The extracted blur kinetics display perfectly linear Arrhenius behavior, indicating that there is no sign for secondary electron blur at 22nm half pitch. At the lowest PEB setting the total blur length is ~4nm, indicating that secondary electron blur should be well below that. The stochastic resist model gives a best fit to the current data set with parameters that result in a maximum probability of acid generation at 2.4nm from the photon absorption site. Extrapolation of the model predicts that towards the 16nm half pitch the impact on sizing dose is minimal and an acceptable exposure latitude is achievable. In order to limit the impact on line width roughness at these dimensions it will be required to control acid diffusion to ~5nm.
- 17 March 2011
- Published by SPIE-Intl Soc Optical Eng
- Vol. 7969, 796904-796904-11
EUV secondary electron blur at the 22nm half pitch node
In this paper the Arrhenius behavior of blur upon EUV exposure is investigated through variation of the PEB temperature. In this way, thermally activated parameters that contribute to blur (such as acid/base diffusion) can be separated from non-thermally activated parameters (such as secondary electron blur). The experimental results are analyzed in detail using multi-wavelength resist modeling based on the continuum approach and through fitting of the EUV data using stochastic resist models. The extracted blur kinetics display perfectly linear Arrhenius behavior, indicating that there is no sign for secondary electron blur at 22nm half pitch. At the lowest PEB setting the total blur length is ~4nm, indicating that secondary electron blur should be well below that. The stochastic resist model gives a best fit to the current data set with parameters that result in a maximum probability of acid generation at 2.4nm from the photon absorption site. Extrapolation of the model predicts that towards the 16nm half pitch the impact on sizing dose is minimal and an acceptable exposure latitude is achievable. In order to limit the impact on line width roughness at these dimensions it will be required to control acid diffusion to ~5nm.
- 17 March 2011
- Published by SPIE-Intl Soc Optical Eng
- Vol. 7969, 79691T
Impact of mask line roughness in EUV lithography
Resist line edge/width roughness is one of the most critical aspects in EUV lithography for the 32 nm technological node and below. It is originated by the uncertainties which characterize the lithographic process: source speckle effect, mask line and surface roughness, mirror roughness, flare effect and resist pattern formation all contribute to the final roughness. In this paper mask and resist line edge roughness were compared by means of frequency analysis on top-down SEM images: it was found that low frequencies mask roughness are well correlated with the Power Spectral Density of the resist roughness. Mask high frequencies components resulted less critical due to the natural cut-off of the optical system. Experimental data for both mask and resist were implemented in the PROLITH Stochastic Resist Model simulator to quantify the mask line edge roughness contribution to the final resist roughness: the results showed that 16% of the low frequency resist roughness component is originated at the mask level. For that reason, mask impact was set as 0.6 nm of the overall line edge roughness resist budget.
- 17 March 2011
- Published by SPIE-Intl Soc Optical Eng
- Vol. 7969, 79691T
Impact of mask line roughness in EUV lithography
Resist line edge/width roughness is one of the most critical aspects in EUV lithography for the 32 nm technological node and below. It is originated by the uncertainties which characterize the lithographic process: source speckle effect, mask line and surface roughness, mirror roughness, flare effect and resist pattern formation all contribute to the final roughness. In this paper mask and resist line edge roughness were compared by means of frequency analysis on top-down SEM images: it was found that low frequencies mask roughness are well correlated with the Power Spectral Density of the resist roughness. Mask high frequencies components resulted less critical due to the natural cut-off of the optical system. Experimental data for both mask and resist were implemented in the PROLITH Stochastic Resist Model simulator to quantify the mask line edge roughness contribution to the final resist roughness: the results showed that 16% of the low frequency resist roughness component is originated at the mask level. For that reason, mask impact was set as 0.6 nm of the overall line edge roughness resist budget.
- 1 May 2010
- Published by Wiley in SID Symposium Digest of Technical Papers
- Vol. 41 (1) , 1151-1154
77.3: Invited Paper : Roll‐to‐Roll Manufacturing of Backplanes for Paper‐Like Displays
HP and Phicot are planning the world's first R2R (roll‐to‐roll) manufacturing line for display backplanes based on the SAIL (Self‐Aligned Imprint Lithography) process. Economic benefits for R2R compared to batch, cost comparisons of different R2R processes, comparison of substrate options, and necessary supply chain infrastructure developments are presented.
- 1 May 2010
- Published by Wiley in SID Symposium Digest of Technical Papers
- Vol. 41 (1) , 1151-1154
77.3: Invited Paper : Roll‐to‐Roll Manufacturing of Backplanes for Paper‐Like Displays
HP and Phicot are planning the world's first R2R (roll‐to‐roll) manufacturing line for display backplanes based on the SAIL (Self‐Aligned Imprint Lithography) process. Economic benefits for R2R compared to batch, cost comparisons of different R2R processes, comparison of substrate options, and necessary supply chain infrastructure developments are presented.
- 11 March 2010
- Published by SPIE-Intl Soc Optical Eng
- Vol. 7636, 76360R-76360R-10
Resist pattern prediction at EUV
Accurate and flexible simulation methods may be used to further a researcher's understanding of how complex resist effects influence the patterning of critical structures. In this work, we attempt to gain insight into the behavior of a state-of-the-art EUV resist through the use of stochastic resist simulation. The statistics of photon and molecule counting are discussed. A discrete, probabilistic ionization and electron scattering simulator for acid generation at EUV is discussed. At EUV, acid generators are hypothesized to be activated by secondary electrons yielded by ionization of the resist upon absorption of photons. Model fit to experimental data of mean CD and LWR for a state-of-the-art EUV resist is shown.
- 11 March 2010
- Published by SPIE-Intl Soc Optical Eng
- Vol. 7636, 76360R-76360R-10
Resist pattern prediction at EUV
Accurate and flexible simulation methods may be used to further a researcher's understanding of how complex resist effects influence the patterning of critical structures. In this work, we attempt to gain insight into the behavior of a state-of-the-art EUV resist through the use of stochastic resist simulation. The statistics of photon and molecule counting are discussed. A discrete, probabilistic ionization and electron scattering simulator for acid generation at EUV is discussed. At EUV, acid generators are hypothesized to be activated by secondary electrons yielded by ionization of the resist upon absorption of photons. Model fit to experimental data of mean CD and LWR for a state-of-the-art EUV resist is shown.
- 1 January 2006
- Published by Springer Nature
- p. 235-253
The Maralinga Range after 1963
- 15 July 2002
- Published by SPIE-Intl Soc Optical Eng
- Vol. 4691, 125-137
3D lumped parameter model for lithographic simulations
Simplified resist models are desired for fast simulation of resist profiles over large mask areas. The Lumped Parameter Model was originally developed as one such model. However, the LPM model has been limited to 2D resist simulations of 1D aerial image slices with positive tone resists. In this paper we present a modified Lumped Parameter Model applicable to 3D resist simulations of both positive and negative tone resists. In addition several new LPM parameters are introduced that further improve accuracy. The derivation of the 3D LPM model, rationale for including the new parameters, and simulation results using the new model are given.
- 15 July 2002
- Published by SPIE-Intl Soc Optical Eng
- Vol. 4691, 125-137
3D lumped parameter model for lithographic simulations
Simplified resist models are desired for fast simulation of resist profiles over large mask areas. The Lumped Parameter Model was originally developed as one such model. However, the LPM model has been limited to 2D resist simulations of 1D aerial image slices with positive tone resists. In this paper we present a modified Lumped Parameter Model applicable to 3D resist simulations of both positive and negative tone resists. In addition several new LPM parameters are introduced that further improve accuracy. The derivation of the 3D LPM model, rationale for including the new parameters, and simulation results using the new model are given.
- 15 July 2002
- Published by SPIE-Intl Soc Optical Eng
- Vol. 4691, 1199-1210
Comparison between the process windows calculated with full and simplified resist models
While numerical simulation is generally regarded as indispensable for wavefront engineering tasks such as OPC decoration and phase-shift mask design, full resist models are rarely used for this purpose. By full resist models, we mean models derived from a physical, mechanistic description of the chemical response of the photoresist to exposure and the subsequent PEB and develop processes. More often, simplified models such as an aerial image threshold model or the Lumped Parameter Model (LPM) are used because these models are much faster and make optimization of optical extension technology more tractable. Simplified resist models represent a compromise between computational speed and simulation accuracy. The purpose of this study is to quantify the differences between the process windows calculated with simplified and full resist models. Our approach is first to fit the parameters in the simplified models to match results obtained with a full resist model, and then to compare the predictions of the simplified resist models with those obtained with the full model. We take two approaches to model tuning: mathematical derivation of relationships between the models, and least-squares fitting of FE matrix data for isolated and dense lines.
- 15 July 2002
- Published by SPIE-Intl Soc Optical Eng
- Vol. 4691, 1199-1210
Comparison between the process windows calculated with full and simplified resist models
While numerical simulation is generally regarded as indispensable for wavefront engineering tasks such as OPC decoration and phase-shift mask design, full resist models are rarely used for this purpose. By full resist models, we mean models derived from a physical, mechanistic description of the chemical response of the photoresist to exposure and the subsequent PEB and develop processes. More often, simplified models such as an aerial image threshold model or the Lumped Parameter Model (LPM) are used because these models are much faster and make optimization of optical extension technology more tractable. Simplified resist models represent a compromise between computational speed and simulation accuracy. The purpose of this study is to quantify the differences between the process windows calculated with simplified and full resist models. Our approach is first to fit the parameters in the simplified models to match results obtained with a full resist model, and then to compare the predictions of the simplified resist models with those obtained with the full model. We take two approaches to model tuning: mathematical derivation of relationships between the models, and least-squares fitting of FE matrix data for isolated and dense lines.