We have shown that when values from the literature are used for the model-insensitive parameters, the number of calibrated parameters can be reduced for Model A from four to two, for Model C from four to one, for Model B from six to four, and for Model D from six to three. The reduced number of calibrated parameters leads to a reduced complexity of the calibration task and a reduced uncertainty in optimized parameters, especially in connection with the HYDRUS programs. The small number of additional calibration parameters makes the modeling approach of temperature-dependent root growth a convenient add-on to the HYDRUS models. However, the implementation of the influence of other stress factors on the vertical root penetration, as well as on the root length density distribution, into the HYDRUS models still needs to be validated against experimental data.The purpose of this project is to develop the technology necessary for a fiber based laser plasma accelerator to be used for electronic brachytherapy treatment of prostate cancer. Such a device would deliver MeV electrons suitable for medical applications in a small form factor machine. The laser pulse is transported through a hollow core fiber optic that terminates at an embedded gas target where ionization and acceleration occurs. LPAs operate with energy gradients 1000 times larger than conventional accelerators allowing for the energy gain to occur within a millimeter of the treatment area, increasing the therapeutic dose to the tumor site while minimizing collateral irradiance of adjacent healthy tissue. These milestones include, design of a gas jet capable of providing 1-10 MeV electrons, development of a hollow core fiber suitable for transporting the necessary laser energy,ebb and flow trays and establishing a program to evaluate the efficacy of LPA produced electrons as a radiation source. The gas jet target uses a gas profile featuring a sharp density transition to reduce the laser energy needed for electron injection.
Laser plasma simulations will be conducted to establish the gas parameters and physical dimensions necessary to produce MeV electrons. This will be followed by design and fabrication of a gas jet able to deliver the requisite profile. Accomplishments Simulations have successfully shown 1-10 MeV electron beam generation with about 10 mJ of laser energy. This is achieved by relying on breaking of a laser excited plasma wave at a sharp density transition. To realize this density profile with step, a special gas jet has been developed with two distinct gas regions. The second gas region has half the electron density of the first region, and the density transition occurs in a few 10’s of µm that, according to the simulations, is expected to allow electron injection with small amounts of laser energy. The total length of the gas cell is 350 µm, sufficiently small to act as an internal target during treatment. Guided by the simulation results, a design was developed to physically achieve such a density profile. This jet design uses two converging diverging nozzles that are pressure matched at the nozzles’ exit. The resulting flow is known as a slip line where two regions of disparate density can exist in steady state. In order to demonstrate the initial design, the gas jet was constructed at a scale that is fourteen times larger than the original design. This enables easier manufacturing and a rather straightforward measurement of the density drop. Furthermore, the enlarged device would act as an initial laser target for comparison to simulation. A provisional patent was submitted for this novel two-stream gas jet technique. Currently, we are in the process of integrating the target into an existing target chamber that is powered by an existing10 TW laser system. This includes modifying several components of the system to accommodate the new target. In initial experiments, high power laser beams have been focused on the gas target with the longitudinal profile of plasma recombination light showing evidence of a density drop. Quantifying the longitudinal gas density and studies to establish the minimum amount of laser energy that is needed to generate e-beams are scheduled for FY15. Additionally, a second round of simulations that includes ionization effects for the scaled up target is being performed to benchmark experimental results.This proposal is laying the groundwork for the development of computer codes that will be ultimately capable of virtual prototyping and virtual computer experiments on the next generation of supercomputers. The advanced design of future particle accelerators requires consideration of a variety of physical processes that are not appropriately modeled with current techniques and computational power.
We are pursuing the development of a new, highly accurate and scalable solver combining the accuracy of spectral methods with the scalability of finite-difference methods. We propose a crucial change to the scalable algorithm, namely changing the solver portion from FDTD to PSATD solver, to mitigate significant unphysical effects from discretization errors. Traditionally, such a move would compromise the scalability of the PIC method and render the calculation intractable for the required accuracy and spatial resolution. Key to our approach is the design and high-performance implementation of a new solver to take advantage of multilevel parallelism in emerging systems by naturally subdividing the computational domain and workload in a way that is optimally assigned to the heterogeneous computational resources. Our algorithmic research will provide high numerical accuracy and stability. One of our most significant accomplishments is the development of a novel method for highfidelity modeling of the propagation of electromagnetic waves on a supercomputer. The unprecedented level of flexibility in the tuning of accuracy and locality of our method enables very high accuracy while preserving scalability to a very large number of computer cores. Our new solver also offers larger time steps than conventional solvers by removing the standard limitations due to the well-known Courant-Friedrich-Levy condition. Another significant accomplishment is the analysis of the so-called “numerical Cherenkov” instability for our solver in various modes that has lead to the successful development of novel methods that mitigate the instability. The mitigation method has been automated in our ParticleIn-Cell code Warp and is now successfully used by collaborators at other institutions. Strong scaling of the new method was demonstrated on NERSC’s supercomputers Hopper and Edison, yielding near-linear scaling up-to 50,000 cores. During the scalability studies, we also worked to analyze the performance variability of Edison at larger core counts and developed methods to ascertain scalability in the time-sharing environment with the newer Cray Aries interconnect with Dragonfly topology. The scalability of our solver was also contrasted to the scaling of standard global FFTs. The results were presented at the international SC’14 conference and will be documented in a journal publication. We are beginning to develop and study the scalability of the solver on test-bed next generation computers through portable implementations of the solvers, via heterogeneous and emerging programing models. The solver is critical to the further development of the Warp code, which has been selected in a highly competitive competition as a first tier NERSC NESAP application. We are also in the process of extending our methods to three dimensions and applying them to large-scale simulations of laser plasma accelerators.
The goal of this proposal is to develop the scientific user community for the next generation soft x-ray spectro-microscopy and tomography based on x-ray ptychography. X-ray ptychography is a robust diffraction imaging method that can take full advantage of the high brightness of the ALS and which is not limited in resolution by the x-ray optics. The technique we have developed enhances the already high resolution scanning transmission x-ray microscope with the addition of diffraction data and is able to probe chemical species, molecular orientation, magnetization,4×8 flood tray and structural morphology at <10 nm resolution. We will apply this method to difficult problems in the material sciences with a focus on materials for energy and carbon cycle research. In particular, we will study chemical phase transformations in nano-crystals of active battery electrodes, hydration reactions in calcium silicate hydrate, pore evolution in sintered zirconia, and magnetic domain structure in thin magnetic films. Finally, we will further develop in situ methods for high resolution microscopy of electrochemical and hydration reactions in these materials. These studies require the ability to visualize 10 nm scale features embedded in a micron-scale matrix with sensitivity to the chemical species however there currently is no imaging technique with such capabilities. We will develop soft x-ray scanning diffraction microscopy with the study of these systems and provide a microscopy program that will directly and efficiently benefit from the very high brightness of the new COSMIC beamline at the ALS and future beamlines at the ALS-U. This program will readily be extended to other problems in the materials sciences such as the study of porous zeolites for carbon capture and catalysis and the characterization of materials by design like self-assembled nanoparticle aggregates for replacement of non-renewable resources in batteries and solid state devices. Finally, this project will leverage the computational resources at NERSC for the reconstruction of Giga-element datasets and potentially provide experimental feedback to material computation efforts. Using the unique spatial resolution and chemical sensitivity of our imaging method we have directly visualized a two phase chemical transformation and the complicated chemical domain pattern present in the smallest available nano-crystals of LiFePO4. Our measurements were also correlated with high resolution transmission electron microscope images of many such particles to show that these small particles do not suffer from material fracture but still display a complicated domain pattern which does not correlate with the material crystallographic axes as was expected. These results seem to indicate that the superior electrochemical performance of the small particles is due primarily to mechanical issues rather than the fundamental character of the phase transformation itself and provides critical insight into how to engineer higher performing batteries. We have also used our high resolution imaging technique to measure the magnetic domain wall width distribution in a thin film of CoPd, the pore size distribution in sintered Yttria stabilized zirconia, and the x-ray magnetic circular dichroism spectrum in magnetotactic bacteria all with spatial resolutions which are not achievable at any other x-ray facility worldwide.
Storage ring based light sources have been extremely successful over the years, enabling forefront science in many areas. Recent developments in accelerator technology and lattice design open the door for very large further increases in brightness – more than 100 times, particularly by reducing the horizontal emittance. This can greatly enhance the capability of light sources for imaging and spectroscopy. While there are no showstoppers, work was necessary to reduce risk and cost, enable a timely execution, and optimize the properties of the source to the needs of the science case and prospective user community. The goal of this project was to investigate, demonstrate and improve critical technologies necessary for diffraction limited storage ring light sources at the subsystem level. At a later stage, tests will be carried out with some of those new systems using beams in existing light sources. roject Description There is a great need in advanced diffraction gratings at the ALS and other synchrotron facilities. To provide high spectral resolution and high diffraction efficiency the gratings are required to have a variable line spacing , perfect blazed grooves, and smooth groove surface. For Reflective Zone Plates elliptical lines are required to provide two dimensional focusing. Due to limitations of conventional techniques one can only design gratings within a very limited set of variables. New approaches for fabrication of high quality grating in a cost effective manner should be found to make advanced diffraction gratings available. This work aims development of a new grating fabrication process based on Direct Write Lithography technique which is used in semiconductor industry for making high resolution MEMS structures, nanostructures, and consumer microelectronics. A grating pattern can be written by scanning a focused laser or electron beam over a substrate surface coated with a photoresist. The DWL technology will be used as a pattering technique for making diffraction gratings of arbitrary groove density variation and groove shape complexity. The best available DWL systems will be used for making grating prototypes which will be characterized in terms of groove position accuracy, imaging performances, and diffraction efficiency. DWL process will be optimized and eventually combined with wet anisotropic etch technique for making highly efficient blazed gratings.