Heterogeneous computing system with both CPUs and GPUs has become a class of widely used hardware architecture in supercomputers. As heterogeneous systems delivering higher computational performance, they are being built with an increasing number of complex components. This is anticipated that these systems will be more susceptible to hardware faults with higher power consumption. Numerical linear algebra libraries are used in a wide spectrum of high-performance scientific applications. Among numerical linear algebra operations, one-sided matrix decompositions can sometimes take a large portion of execution time or even dominate the whole scientific application execution. Due to the computational characteristic of one-sided matrix decompositions, they are very suitable for computation platforms such as heterogeneous systems with CPUs and GPUs. Many works have been done to implement and optimize one-sided matrix decompositions on heterogeneous systems with CPUs and GPUs. However, it is challenging to enable stable and high performance one-sided matrix decompositions running on computing platforms that are unreliable and high energy consumption. So, in this thesis, we aim to develop novel fault tolerance and energy efficiency optimizations for one-sided matrix decompositions on heterogeneous systems with CPUs and GPUs.

To improve reliability and energy efficiency, extensive researches have been done on developing and optimizing fault tolerance methods and energy-saving strategies for one-sided matrix decompositions. However, current designs still have several limitations: (1) Little has been done on developing and optimizing fault tolerance method for one-sided matrix decompositions on heterogeneous systems with GPUs; (2) Limited by the protection coverage and strength, existing fault tolerance works provide insufficient protection when applied to one-sided matrix decompositions on heterogeneous systems with GPUs; (3) Lack the knowledge of algorithms, existing system level energy saving solutions cannot achieve the optimal energy savings due to potentially inaccurate and high-cost workload prediction they rely on when they are used in one-sided matrix decompositions; (4) It is challenging to apply both fault tolerance techniques and energy saving strategies to one-side matrix decompositions at the same time given that their current designs are not naturally compatible with each other.

To address the first problem, based on the original (Algorithm Based Fault Tolerance) ABFT, we develop the first ABFT for matrix decomposition on heterogeneous systems with GPUs together with the novel storage errors protection and several optimization techniques specifically for GPUs. As for the second problem, we design a novel checksum scheme for ABFT that allows data stored in matrices to be encoded in two dimensions. This stronger checksum encoding mechanism enables much stronger protection including enhanced error propagation protection. In addition, we introduce a more efficient checking scheme. By prioritizing the checksum verification according to the sensitivity of matrix operations to soft errors with optimized checksum verification kernel for GPUs, we can achieve strong protect to matrix decompositions with comparable overhead. For the third problem, to improve energy efficiency for one-sided matrix decompositions, we introduce an algorithm-based energy-saving approach designed to maximize energy savings by utilizing algorithmic characteristics. Our approach can predict program execution behavior much more accurately, which is difficult for system level solutions for applications with variable execution characteristics. Experiments show that our approach can lead to much higher energy saving than existing works. Finally, for the fourth problem, we propose a novel energy saving approach for one-sided matrix decompositions on heterogeneous systems with GPUs. It allows energy saving strategies and fault tolerance techniques to be enabled at the same time without brings performance impact or extra energy cost.

Phonon, a quantized lattice vibration, plays an important role in the materials’ physical and mechanical properties. This makes investigations on phonon dynamics an indispensable subject for better understanding the physical world. For decades, people have been seeking ways to manipulate phonon dynamics, which are closely related to atomic structures and electromagnetic properties of materials, for the development of novel materials with desired properties and more insights into the fundamental physics. This dissertation discusses the studies of phonon engineering in the spatially confined silicon systems, which have been widely used in the biomedical filed and have great potential applications in optoelectronic industry, as well as bulk sapphire systems using inelastic neutron scatterings.The dissertation starts with introducing the basic information about the phonon dynamics and neutron scatterings which are the two primary subjects of my graduate research. In Chapters 2 to 4, the inelastic neutron scattering experiments and discussions on the effects of particle size, temperature, surface oxidization, and surface functionalization on the phonon dynamics of silicon nanocrystals are presented. These effects are found to be greater on the transverse acoustic phonon modes than the optical phonon modes in silicon nanocrystals. In Chapter 5, the atomic structures of 3-nm spherical silicon nanocrystals were measured with elastic neutron scattering for the first time. The diffraction spectra show huge anisotropic structure variations inside the silicon nanocrystals compared to their bulk counterpart. In Chapter 6, the effect of the low concentration dopants on the phonon dynamics of sapphire was studied using inelastic neutron scattering. This dissertation sheds light on the phonon dynamics, as well as their dependence on the intrinsic and extrinsic effects, of materials with great potential applications and will contribute to further investigations on the phonon engineering of various materials.

Nanoparticles represent a novel class of material consisting of hundreds to a few thousand atoms each, and their physical and chemical properties are significantly different than the bulk materials. The electronic structure, chemical and optical properties of nanoparticles can be tuned by the size, shape, surface modification and interaction with supporting materials, to fulfil the potential specific applications in catalysis, imaging and electronic devices. In the preparation of nanoparticles, protecting ligands play a crucial role in the dispersion, size control, and shape control of particles. Here in this thesis, ligand functionalization of metal nanoparticles and engineering of carbon nanomaterials were manipulated, and the influence of metal-organic interaction on the chemical, optical and electrochemical properties of nanoparticles and their applications in fuel cell electrocatalysis were studied.

Ruthenium nanoparticles protected by ferrocenecarboxylates (RuFCA) were synthesized. The carboxylate group were bound onto the nanoparticle surface via Ru–O bonds in a bidentate configuration which is highly polarized, leading to the diminishment of the electron density of the iron centers and the increase of formal potential of the ferrocenyl moieties by 120mV in electrochemical measurements. In addition, galvanic exchange reactions of the RuFCA nanoparticles with Pd(II) followed by hydrothermal treatment at 200 °C led to (partial) decarboxylation of the ligands such that the ferrocenyl moieties were now directly bonded to the metal surface, as manifested in voltammetric measurements that suggested intervalence charge transfer between the nanoparticle-bound ferrocene groups.

In a further study, decarboxylation was also found happened at the metal-ligand interface in the hydrothermal treatment of 2-naphthalenecarboxylate protected ruthenium nanoparticles at higher temperature 250°C, and the naphthalenyl moieties became directly bonded to the metal cores, which was confirmed by infrared and X-ray photoelectron spectroscopic measurements. In comparison with the as-produced RuCOONA nanoparticles, the decarboxylated nanoparticles (RuNA) exhibited markedly different optical and electronic properties due to electronic coupling between the particle-bound naphthalene groups. The intraparticle charge delocalization led to spilling of nanoparticle core electrons to naphthalene moieties, resulting in the negative shift of the formal potential.

In addition, the electron-transfer properties of the ruthenium nanoparticles protected by 1-dodecyne, laurate and 1-dodecanethiol were examined by scanning tunneling spectroscopic (STS) measurements. Ruthenium-vinylidene (Ru=C=CH–), –oxygen (Ru-O), and –thiolate (Ru-S) interfacial bonds were formed when protected by alkyne, carboxylate and thiol ligands, respectively, and the polarization of the interfacial bonds was found to increase in the order of Ru=C=CH– < Ru-S < Ru-O. The relatively large nanoparticles (dia. ∼ 3 nm) were found to show clearly-defined Coulomb staircase; and with diminishing particle core dimensions to below 1 nm, Coulomb blockade started to emerge. The nanoparticle molecular capacitance and effective nanoparticle dielectric constants were estimated and the dielectric constants increased inversely with the nanoparticle core dimensions; and at any given particle size, the dielectric constants varied with the specific metal-ligand interfacial bonds, increasing in the order of Ru-S < Ru=C=CH– < Ru-O.

Part of the dissertation research was devoted to carbon nanomaterials. In one of the study, nitrogen-doped graphene quantum dots (NGQDs) were prepared by a facile hydrothermal method and incorporated with ruthenium metal ions by exploiting the unique complexation of selected nitrogen dopants with ruthenium ions. Complexation of NGQDs with ruthenium ions likely occurred through the pyridinic nitrogen dopants, leading to the incorporation of multiple metal centers within the conjugated graphitic C sp2 scaffolds (Ru-NGQDs). Intervalence charge transfer between embedded Ru ions was studied electrochemically, the Ru-NGQD compounds exhibited two pairs of voltammetric waves, with a peak spacing of 150 mV, suggesting Class II delocalized system. Near-infrared spectroscopic measurements demonstrated an absorption band emerged at ca. 1450 nm at mixed-valence metal charge transfer, by using Ce(SO4)2 as the oxidizing reagent.

Covalently crosslinking of GQDs were accomplished by forming ensembles of a few hundred nanometers in size by McMurry deoxygenation coupling reactions of peripheral carbonyl functional moieties catalyzed by TiCl4 and Zn powders in refluxing THF. Photoluminescence measurements showed that after chemical coupling, the excitation and emission peaks blue-shifted somewhat and the emission intensity increased markedly, likely due to the removal of oxygenated species where quinone-like species were known to be effective electron-acceptors and emission quenchers.

Metal nanoparticles were also prepared and tuned for the applications as high-performance catalysts. In one of the research, gold core@silver semishell Janus nanoparticles were prepared by interfacial chemical etching of Au@Ag core–shell nanoparticles at the air/water interface. The resulting Janus nanoparticles exhibited an asymmetrical distribution of silver on the surface of the gold cores. Interestingly, the Au@Ag semishell Janus nanoparticles exhibited enhanced electrocatalytic activity in oxygen reduction reactions, as compared to their Au@Ag and Ag@Au core–shell counterparts, likely due to a synergistic effect between the gold cores and silver semishells that optimized oxygen binding to the nanoparticle surface. In another research, cysteine-stabilized Ag–Cu hollow nanoshells are prepared by the co-reduction of silver nitrate and cupric nitrate with sodium borohydride in the presence of sodium thiocyanate. When capped with 1-dodecanethiol, the hollow nanoshells become dispersible in apolar organic solvents and the cavity may be exploited for the effective phase-transfer of target molecules such as rhodamine 6G between water and organic media. The Ag–Cu nanoshells also show apparent catalytic activity toward the reduction of 4-nitroaniline by sodium borohydride, a performance that is markedly better than that of the solid counterparts and comparable to leading results in recent literature based on relevant metal catalysts.

Bone defects have become a challenging health issue after extensive trauma, osteoporosis, and tumor extirpation worldwide. To enhance bone healing, bone grafts have been employed in clinical treatment for many years. Autologous grafts have been considered as the “gold standard” for the bone graft material, but their application is limited by the donor site morbidity. Demineralized bone matrix (DBM), one of the alternations of autologous graft, has been widely used for craniofacial bone repair. However, the application of DBM is compromised due to its rapid dispersion after the implantation and low osteoinductivity. In order to deliver and localize DBM, carriers with high bioactivity are in need to be developed.

Chitosan has been proved to be attractive for researchers in the field of tissue bioengineering due to its high biocompatibility, biodegradability, and low immunogenicity. Besides, heparin, a highly sulfated polysaccharide in the extracellular matrix, possesses the typical structure which can interact with multiple osteoinductive factors to enhance and stabilize the biological activities of growth factors. Therefore, we synthesized a chitosan-based hydrogel with heparinized conjugation (Hep-MeGC), which can not only deliver DBM but also enhance its osteogenic activity. Moreover, upon the stimulation of BMPs (major osteogenic factors in DBM), BMP efficacy is significantly reduced due to increased expression of BMP antagonists like Noggin. Thus, the potency of BMPs in DBM may be enhanced through Noggin suppression. This study aims to investigate the new method by using Noggin suppression and Hep-MeGC hydrogel to enhance DBM-mediated bone repair.

To suppress Noggin, human bone marrow stem cells (hBMSCs) were transfected with lentiviral particles that encode Noggin-shRNA. The methacrylate glycol chitosan (MeGC) and heparinized chitosan hydrogel (Hep-MeGC) hydrogel, were then prepared to deliver DBM and BMSCs with Noggin-suppression. The effect of DBM and Noggin suppression on osteogenesis was evaluated in vitro, and the ability of DBM and Noggin suppression in Hep-MeGC hydrogels to bone regeneration was investigated through in vivo study.

Noggin suppression + DBM greatly enhanced the osteogenesis of BMSCs in Hep-MeGC hydrogels, as shown by the increased ALP production, calcium deposition, and expression of osteogenic genes. The mechanistic study showed that Noggin suppression enhanced the DBM-based osteoinductivity through the stimulation of BMP-Smad signals. Detected by microCT and histological/immunohistochemical analysis, the combinational treatment further displayed a robust bone healing capacity in the critical-sized mouse calvarial defect model. To conclude, the incorporation of DBM and Noggin suppression in the Hep-MeGC hydrogel significantly promoted bone healing.