Robert J. Mears
Robert J. Mears has a total of 12 co-authors. Here are the top co-authors by number of shared publications:
- Hideki Takeuchi
Atomera Inc. 2 , Los Gatos, California 95032
2 shared publications - Michael Charles Parker
School of Computer Science and Electrical Engineering, University of Essex, Colchester CO4 3SQ, UK
2 shared publications - Albert Lu
Department of Electrical Engineering, San José State University 1 , San José, California 95192
1 shared publication Daniel Connelly
Atomera Inc. 2 , Los Gatos, California 95032
1 shared publicationDonghun Kang
Atomera Inc. 2 , Los Gatos, California 95032
1 shared publication- Hiu Yung Wong
San Jose State University
1 shared publication
Latest Publications
- Open Access
- 25 March 2026
- Published by AIP Publishing in Journal of Applied Physics
- Vol. 139 (12)
Electron and hole surface roughness scattering-limited mobilities in oxygen-inserted (OI) Si channel
The electron and hole surface roughness scattering (SRS)-limited mobilities, deconvolved from 4.2 K measurements, were compared between oxygen-inserted (OI) and control Si channel for n- and pMOSFETs. It was found that electron SRS-limited mobility in the OI-Si channel is 55% higher, whereas hole SRS-limited mobility is comparable to the control. To elucidate the difference, the Si/SiO2 interface roughness and the nanoscale thickness variation of the gate oxide were analyzed in detail, via high-resolution transmission electron microscopy images, and fed into numerical calculation of the SRS-limited mobility. The power spectral densities (PSDs) based on the Yule–Walker autoregressive model revealed that the OI-Si has a smaller surface roughness at low-frequency components, corresponding to the electron Fermi surface energy level. In contrast, surface roughness improvement was found to be less at a higher frequency region of the PSDs corresponding to the hole Fermi surface energy level. Furthermore, an oxygen tracer experiment identified a dynamic exchange process between the Si/SiO2 interface and the OI-Si layer, suggesting an opportunity for improving hole SRS-limited mobility in the OI-Si channel via tuning post-oxidation thermal process steps.
- 25 March 2026
- Published by AIP Publishing in Journal of Applied Physics
- Vol. 139 (12)
Open AccessElectron and hole surface roughness scattering-limited mobilities in oxygen-inserted (OI) Si channel
The electron and hole surface roughness scattering (SRS)-limited mobilities, deconvolved from 4.2 K measurements, were compared between oxygen-inserted (OI) and control Si channel for n- and pMOSFETs. It was found that electron SRS-limited mobility in the OI-Si channel is 55% higher, whereas hole SRS-limited mobility is comparable to the control. To elucidate the difference, the Si/SiO2 interface roughness and the nanoscale thickness variation of the gate oxide were analyzed in detail, via high-resolution transmission electron microscopy images, and fed into numerical calculation of the SRS-limited mobility. The power spectral densities (PSDs) based on the Yule–Walker autoregressive model revealed that the OI-Si has a smaller surface roughness at low-frequency components, corresponding to the electron Fermi surface energy level. In contrast, surface roughness improvement was found to be less at a higher frequency region of the PSDs corresponding to the hole Fermi surface energy level. Furthermore, an oxygen tracer experiment identified a dynamic exchange process between the Si/SiO2 interface and the OI-Si layer, suggesting an opportunity for improving hole SRS-limited mobility in the OI-Si channel via tuning post-oxidation thermal process steps.
- 22 November 2024
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2024-02 (32) , 2364
(Invited) Applications of Oxygen Inserted Epitaxy
Silicon remains the semiconductor material of choice for most electronic applications of semiconductors and the silicon–silicon dioxide interface is one of the most studied interfaces. Over the past decade or more, epitaxial silicon-germanium (eSiGe) has found increasing applications in advanced logic, first for PMOS compressive source-drain (S/D) stressors, then as a channel material in high-k metal gate (HKMG) devices to help control the PMOS Vt, and more recently as a sacrificial layer to enable stacked nanosheets and even as a S/D liner to reduce phosphorus diffusion into the undoped nanosheets. Oxygen inserted (OI) silicon epitaxy was originally developed to complement and bridge the gap between silicon and silicon-germanium, and it has found applications ranging from improved figure of merit power devices to the most advanced nanosheet applications. In this paper we will provide an overview of OI silicon devices, their fundamental characterization and the electrical and physical benefits offered for a variety of semiconductor applications. Introduction Oxygen-inserted (OI) silicon had an unusual genesis for a semiconductor material, in that it was first explored using ab-initio quantum mechanical simulation. It was found that by carefully inserting a partial monolayer of oxygen during silicon epitaxy a stable layer could be manufactured where the individual oxygen atoms arranged themselves on a silicon-silicon bond, but that overall, the silicon retained its regular lattice, and the silicon-oxygen coordination was only one, as opposed to four in regular silicon dioxide. The idealized lattice configuration is shown in Fig. 1 (a). This configuration is technically thermodynamically metastable, but with a significant energetic barrier to formation of four-fold coordinated silicon dioxide, so that it can survive conventional semiconductor processing. Whereas a single oxygen atom is known to rapidly diffuse in silicon, there is a thermodynamic self-stabilization when the oxygen concentration has an aerial density of the order of 1E15cm-2, with an upper limit dependent both on the specific epitaxial recipe and the required defect density. If all available silicon bonds have attached oxygen, it is difficult to maintain high quality epitaxy. On the other hand, OI silicon has passed the most rigorous defect density requirements for advanced logic applications. Dopant engineering - simulation Oxygen is highly electronegative, and ab initio simulation shows that there is charge transfer in the silicon lattice in the near vicinity of the inserted oxygen as shown in Fig 1 (b). This leads to a change in the formation enthalpy for a single substitutional dopant atom or addition of a silicon point defect, vacancy or interstitial as shown in Fig 1 (c). The figure indicates that substitutional boron and phosphorus will have an energetic preference to be two or three silicon lattice atoms away from the inserted oxygen, whereas silicon point defects prefer to be coincident with the oxygen. As a result, the OI layers disrupt the diffusion of dopant-interstitial pairs, substantially reducing the dopant diffusion. Dopant engineering – experimental data OI silicon has been proven to create unique abrupt doping profiles for a variety of silicon and silicon germanium devices. In power devices, reducing the vertical and lateral diffusion below the OI silicon layer can be used to engineer significant improvement in critical figures of merit, such as the tradeoff between specific ON resistance (Rsp) and the breakdown voltage, and improved time constant (Ron∙ Coff) and reduced body current in RF power devices, while maintaining breakdown voltage. In advanced 3D applications, a 2nm thin liner of OI silicon has proven effective to create a diffusion barrier between highly doped epitaxial source/drain regions and undoped channels. By reducing dopant up-diffusion, thermally robust super-steep retrograde profiles can be created, which can be used to reduce local threshold voltage mismatch in both NMOS and PMOS devices by more than 50%. In gate-first HKMG planar CMOS applications, SiGe channel is used in the PMOS to facilitate Vt engineering. A thin OI silicon layer below the SiGe channel can be used to maintain a super-steep retrograde channel profile after HKMG formation and subsequent thermal processing. Interface interactions It has recently been discovered that OI silicon can also be used to engineer adjacent interfaces. For example, it has been shown by cryogenic measurements that OI silicon reduces surface roughness scattering of the silicon-dielectric interface. Also, OI silicon can reduce the interfacial dipole formed during HKMG processing, with benefits in reduced remote charge scattering and improved reliability. These and other examples of engineered silicon and silicon-germanium devices will be discussed in detail. Figure 1
- 22 November 2024
- Published by The Electrochemical Society in ECS Meeting Abstracts
- Vol. MA2024-02 (32) , 2364
(Invited) Applications of Oxygen Inserted Epitaxy
Silicon remains the semiconductor material of choice for most electronic applications of semiconductors and the silicon–silicon dioxide interface is one of the most studied interfaces. Over the past decade or more, epitaxial silicon-germanium (eSiGe) has found increasing applications in advanced logic, first for PMOS compressive source-drain (S/D) stressors, then as a channel material in high-k metal gate (HKMG) devices to help control the PMOS Vt, and more recently as a sacrificial layer to enable stacked nanosheets and even as a S/D liner to reduce phosphorus diffusion into the undoped nanosheets. Oxygen inserted (OI) silicon epitaxy was originally developed to complement and bridge the gap between silicon and silicon-germanium, and it has found applications ranging from improved figure of merit power devices to the most advanced nanosheet applications. In this paper we will provide an overview of OI silicon devices, their fundamental characterization and the electrical and physical benefits offered for a variety of semiconductor applications. Introduction Oxygen-inserted (OI) silicon had an unusual genesis for a semiconductor material, in that it was first explored using ab-initio quantum mechanical simulation. It was found that by carefully inserting a partial monolayer of oxygen during silicon epitaxy a stable layer could be manufactured where the individual oxygen atoms arranged themselves on a silicon-silicon bond, but that overall, the silicon retained its regular lattice, and the silicon-oxygen coordination was only one, as opposed to four in regular silicon dioxide. The idealized lattice configuration is shown in Fig. 1 (a). This configuration is technically thermodynamically metastable, but with a significant energetic barrier to formation of four-fold coordinated silicon dioxide, so that it can survive conventional semiconductor processing. Whereas a single oxygen atom is known to rapidly diffuse in silicon, there is a thermodynamic self-stabilization when the oxygen concentration has an aerial density of the order of 1E15cm-2, with an upper limit dependent both on the specific epitaxial recipe and the required defect density. If all available silicon bonds have attached oxygen, it is difficult to maintain high quality epitaxy. On the other hand, OI silicon has passed the most rigorous defect density requirements for advanced logic applications. Dopant engineering - simulation Oxygen is highly electronegative, and ab initio simulation shows that there is charge transfer in the silicon lattice in the near vicinity of the inserted oxygen as shown in Fig 1 (b). This leads to a change in the formation enthalpy for a single substitutional dopant atom or addition of a silicon point defect, vacancy or interstitial as shown in Fig 1 (c). The figure indicates that substitutional boron and phosphorus will have an energetic preference to be two or three silicon lattice atoms away from the inserted oxygen, whereas silicon point defects prefer to be coincident with the oxygen. As a result, the OI layers disrupt the diffusion of dopant-interstitial pairs, substantially reducing the dopant diffusion. Dopant engineering – experimental data OI silicon has been proven to create unique abrupt doping profiles for a variety of silicon and silicon germanium devices. In power devices, reducing the vertical and lateral diffusion below the OI silicon layer can be used to engineer significant improvement in critical figures of merit, such as the tradeoff between specific ON resistance (Rsp) and the breakdown voltage, and improved time constant (Ron∙ Coff) and reduced body current in RF power devices, while maintaining breakdown voltage. In advanced 3D applications, a 2nm thin liner of OI silicon has proven effective to create a diffusion barrier between highly doped epitaxial source/drain regions and undoped channels. By reducing dopant up-diffusion, thermally robust super-steep retrograde profiles can be created, which can be used to reduce local threshold voltage mismatch in both NMOS and PMOS devices by more than 50%. In gate-first HKMG planar CMOS applications, SiGe channel is used in the PMOS to facilitate Vt engineering. A thin OI silicon layer below the SiGe channel can be used to maintain a super-steep retrograde channel profile after HKMG formation and subsequent thermal processing. Interface interactions It has recently been discovered that OI silicon can also be used to engineer adjacent interfaces. For example, it has been shown by cryogenic measurements that OI silicon reduces surface roughness scattering of the silicon-dielectric interface. Also, OI silicon can reduce the interfacial dipole formed during HKMG processing, with benefits in reduced remote charge scattering and improved reliability. These and other examples of engineered silicon and silicon-germanium devices will be discussed in detail. Figure 1
- Open Access
- 1 December 2005
- Published by Elsevier in Photonics and Nanostructures - Fundamentals and Applications
- Vol. 3 (2-3) , 139-147
A Fourier (k-) space design approach for controllable photonic band and localization states in aperiodic lattices
- 1 December 2005
- Published by Elsevier in Photonics and Nanostructures - Fundamentals and Applications
- Vol. 3 (2-3) , 139-147
Open AccessA Fourier (k-) space design approach for controllable photonic band and localization states in aperiodic lattices
- 26 October 2001
- Published by IOP Publishing in Journal of Optics A: Pure and Applied Optics
- Vol. 3 (6) , S171-S183
A Fourier transform theory for photon localization and evanescence in photonic bandgap structures
We propose an approach to the problem of describing photon behaviour near the bandgap region of photonic crystal structures using a Fourier transform formalism defined in optical reciprocal-space. A modified Debye-Waller approximation in conjunction with the use of a resonant refractive index provides an analytic solution to the coupled-mode equations describing photon propagation in a periodic high dielectric contrast structure. Overall, this represents a new technique for solving the inverse scattering problem. We use this method to consider photon behaviour at the Brillouin zone edges of a one-dimensional photonic crystal structure: our results showing agreement with conventional theory for the standard assumption of an infinitely long grating structure. However, when a photon is located in the bandgap region of a finite (five periods used as an example) high refractive index contrast grating structure, we find that photon velocity becomes superluminal as the photon wave becomes evanescent. This is in agreement with the basic precepts of the uncertainty principle. Additionally, as the bandgap strength increases, the probability of transmission through the dielectric barrier reduces exponentially, with an associated reduction in tunnelling time.
- 26 October 2001
- Published by IOP Publishing in Journal of Optics A: Pure and Applied Optics
- Vol. 3 (6) , S171-S183
A Fourier transform theory for photon localization and evanescence in photonic bandgap structures
We propose an approach to the problem of describing photon behaviour near the bandgap region of photonic crystal structures using a Fourier transform formalism defined in optical reciprocal-space. A modified Debye-Waller approximation in conjunction with the use of a resonant refractive index provides an analytic solution to the coupled-mode equations describing photon propagation in a periodic high dielectric contrast structure. Overall, this represents a new technique for solving the inverse scattering problem. We use this method to consider photon behaviour at the Brillouin zone edges of a one-dimensional photonic crystal structure: our results showing agreement with conventional theory for the standard assumption of an infinitely long grating structure. However, when a photon is located in the bandgap region of a finite (five periods used as an example) high refractive index contrast grating structure, we find that photon velocity becomes superluminal as the photon wave becomes evanescent. This is in agreement with the basic precepts of the uncertainty principle. Additionally, as the bandgap strength increases, the probability of transmission through the dielectric barrier reduces exponentially, with an associated reduction in tunnelling time.