Session GP11: Poster Session III: Reconnection; Laser-plasma Interactions and Acceleration; Plasma Technology; DIII-D I; Hohlraum and X-Ray Cavity Physics; Computation

Hard X-ray Bursts Observed in Association with Magnetic Reconnection in a Solar-Relevant Lab Experiment

Measurements by a plastic scintillator show transient emission of a sub-microsecond pulse of 6 keV X-rays by a cold, dense MHD-driven plasma jet having a collision mean free path much shorter than the jet dimensions so that acceleration of any particles to high energy was not expected. The X-ray pulse occurs when the jet undergoes a kink instability which accelerates the jet laterally so that a fast-growing secondary Rayleigh-Taylor instability is triggered that then breaks the jet. It is proposed that despite the short collision mean free path, an inductive electric field associated with this breaking accelerates a certain subgroup of electrons to 6 keV energy without any of these electrons undergoing collisions. It is further proposed that after being accelerated to high energy, the fast electrons are suddenly decelerated via collisions and radiate X-rays. Extrapolation to both the solar corona and chromosphere predicts the acceleration of a small subset of electrons to very large super-thermal energies by sub-Dreicer electric fields.   

Effective Ion Heating in Guide Field Reconnection

The energy conversion mechanism for ion perpendicular thermal energy is investigated by means of two-dimensional, full particle simulations in an open system. It is shown that ions gain kinetic energy due to the plasma potential drop, which is caused by the charge separation in the one pair of separatrix arms. Based on the force balance in the inflow direction, the strength of the normalized charge density can be expressed by electron Alfv\'{e}n velocity, which is measurable value in the laboratory experiment and/or satellite observation. Meanwhile, we found that the accelerated ions form a ring shape like distribution in $f(v_{1},v_{2})$, as a result, the ion perpendicular temperature $T_{i,perp}$ increases from inflow region. Here, both $v_{1}$ and $v_{2}$ are perpendicular to the magnetic field and $v_{2}$ is parallel to the in-plane. The mixing of particle populations is verified by means of tracing ions and it is shown three typical particle orbits and each orbit has different entry angle to the potential drop. This ring shape like distribution consists three different population due to the difference of the entry angles to the potential drop. This mixing process will thermalize ions and produce entropy without collisions.   

Plasma Transport at the Magnetopause in 3D Kinetic Simulations of MMS Reconnection Site Encounters with Varying Guide Fields

We present 3D fully kinetic simulations of asymmetric reconnection with plasma parameters matching MMS magnetopause diffusion region crossings with varying guide fields of $\sim$0.1 \footnote{Burch et al., Science (2016)}, $\sim$0.4 \footnote{Chen et al. JGR (2017)}, and $\sim$1 \footnote{Burch and Phan, GRL (2016)} of the reconnecting sheath field. For the weakest guide field case, drift turbulence at the magnetospheric separatrix was found to enhance transport and parallel electron heating \footnote{Le et al., GRL (2017)}. Here, we study how varying magnetic shear affects the morphology and plasma transport of the reconnecting magnetopause current sheet. Reconnection rates in 2D and 3D are compared using diagnostics based on particle mixing and on the magnetic field line-integrated parallel electric field. The role of magnetic shear in altering energization and particle distributions along the turbulent magneospheric separatrix is also studied. The PIC simulation results are compared to MMS observations.   

Time domain structures in a colliding magnetic flux rope experiment

Electron phase-space holes, regions of positive potential on the scale of the Debye length, have been observed in auroras as well as in laboratory experiments. These potential structures, also known as Time Domain Structures (TDS), are packets of intense electric field spikes that have significant components parallel to the local magnetic field. In an ongoing investigation at UCLA, TDS were observed on the surface of two magnetized flux ropes produced within the Large Plasma Device (LAPD). A barium oxide (BaO) cathode was used to produce an 18~m long magnetized plasma column and a lanthanum hexaboride (LaB$_6$) source was used to create 11~m long kink unstable flux ropes. Using two probes capable of measuring the local electric and magnetic fields, correlation analysis was performed on tens of thousands of these structures and their propagation velocities, probability distribution function and spatial distribution were determined. The TDS became abundant as the flux ropes collided and appear to emanate from the reconnection region in between them. In addition, a preliminary analysis of the permutation entropy and statistical complexity of the data suggests that the TDS signals may be chaotic in nature.   

Secondary Instabilities in 3-D Magnetic Reconnection under a Strong Guide Field

3-D magnetic reconnection is investigated using the gyrokinetic electron and fully-kinetic ion (GeFi) particle simulation model. The simulation is carried out for a force free current sheet with a strong guide field $B_G$ as occurring in solar and laboratory plasmas. It is found that secondary instabilities are excited in the separatrix region of the primary reconnection due to the 3-D effects associated with the finite $k_z$, where $k_z$ is the wave number along the guide field direction. The instabilities are demonstrated as being of the MHD kink type, which lead to electron heating and acceleration in the parallel direction. The dependence of the growth rate of the secondary instabilities on the electron-ion resistivity, the ion-to-electron mass ratio $m_i/m_e$, and the half-width of the current sheet are also investigated.   

Understanding the dynamics and the energetics of magnetic reconnection in laboratory and space plasmas

We will present recent findings in the research of asymmetric and symmetric magnetic reconnection both in laboratory and space plasmas [1]. In spite of the huge difference ($10^6$-$10^7$) in physical scales, we find remarkable commonality between the properties and the dynamics of the reconnection layer in laboratory and space plasmas. The recent significant progress in diagnostics in the both fields made us possible to directly compare the observed physics processes. The experimental results on the energy conversion and partitioning are discussed and compared with quantitative estimates based on two-fluid analysis as well as space observations. We observed notable similarity in the energy partitioning in the reconnection layer of the MRX and space observations [1,2]. Furthermore, we have observed whistler waves and lower-hybrid frequency fluctuations at the lower density side of asymmetric reconnection layer on MRX [2]. The experimental results are remarkably consistent with the recent space observations from MMS [3]. We directly compare the data from the MRX and the recent MMS observations which show very similar power spectra. Ref.: [1] M. Yamada, et al, PoP 23, 055402 (2016), [2] J. Yoo et al, Submitted to J. G.R. (2017), [3] L. J. Chen et al, This conf.   

Studies of waves during asymmetric reconnection in laboratory and space

Wave activities during asymmetric reconnection have been studied by directly comparing data from the Magnetic Reconnection Experiment (MRX) with data from the Magnetospheric Multiscale (MMS). The power spectrum near separatrices on the low-density (magnetosphere) side shows remarkable similarities. Regarding fluctuations driven by lower hybrid drift instabilities (LHDI), it shows broad spectrum with the energy concentrated mostly below the lower-hybrid frequency. The free energy source of LHDI is the diamagnetic current by the strong density gradient near the separatrix region. The whistler waves show a power spectrum concentrated around half of the electron cyclotron frequency ($0.5f_{ce}$). In MMS, they propagate mostly parallel to the field line toward the X-line with a relatively high phase velocity ($\sim5\times10^7$ m/s). The whistler waves are also observed in the exhaust region and near the electron diffusion region (EDR). However, they become intermittent and with the frequency higher than $0.5f_{ce}$ in the exhaust region and higher than $0.5f_{ce}$ near the EDR. The dominant wave vector satisfies $kd_e\sim1$ for all cases. Possible excitation mechanisms for the observed whistler waves are discussed.   

Drift Kinetic Measurements of Plasma Gradients in MMS Data

In magnetic reconnection, magnetic stress energy is converted into particle energy through the topological rearrangement of magnetic field lines allowed by the breakdown of ideal MHD on small spatial scales. While many models exist to explain reconnection, the crucial physics lies in a thin current layer $\sim 1$ $d_e$ wide. NASA's Magnetospheric Multiscale (MMS) mission seeks to directly investigate magnetic reconnection in Earth's magnetosphere with a tetrahedral formation of four spacecraft. While gradients in plasma properties can be estimated through spacecraft positions, the spacing between each tends to be $\sim 10$ $d_e$, causing gradients at the crucial scale to be poorly resolved. We present a drift-kinetic method for obtaining gradients in the plasma distribution function with data from a single spacecraft. This model is derived from drift kinetics, verified with a VPIC fully-kinetic simulation, and applied to MMS data to infer the geometry of the reconnection region and demonstrate gradient-scale resolution superior to finite difference methods between the four spacecraft.   

Observation of Characteristic Reconnection Regime Structures in the Terrestrial Reconnection Experiment (TREX)

Different regimes of collisionless asymmetric reconnection have different characteristic features and are accessed by varying the plasma parameters of the system [1]. Recently, the Terrestrial Reconnection Experiment (TREX), part of the Wisconsin Plasma Astrophysics Laboratory (WiPAL), used four coils in a 3m spherical vacuum vessel filled with plasma to drive reconnection with an external Helmholtz field. By varying the Helmholtz field, drive field, and plasma species, TREX entered experimentally unexplored reconnection regimes. Using magnetic probes, we observed the characteristic structures of these regimes, including exhaust jets and current layers with widths of only a few d$_{\mathrm{e}}$. These results show better agreement with kinetic simulation than had previously been reported [2] [3]. [1] Le, A. et al. Phys. Rev. Lett., 110, 135004. doi:10.1103/PhysRevLett.110.135004. [2] Le, A. et al. Geophys. Res. Lett., 44, 2096--2104. doi:10.1002/2017GL072522. [3] Ji, H. et al. Geophys. Res. Lett., 35, L13106. doi:10.1029/2008GL034538.   

Plasmoid Instability in Evolving Current Sheets and Onset of Fast Reconnection

A proper description for the onset of the plasmoid instability must incorporate the evolving process of the current sheet, as a high Lundquist number current sheet usually breaks apart before it approaches the Sweet-Parker width. We carry out two-dimensional simulations and theoretical analysis of the plasmoid instability in an evolving background. The plasmoid instability disrupts the current sheet and enters nonlinear regime when the sizes of plasmoids become comparable to the inner layer width of the tearing mode. The linear growth rate, current sheet width, and dominant wavenumber at current sheet disruption depend on not only the Lundquist number, but also the initial noise amplitude. The scalings obtained from simulations can be reproduced with a phenomenological model, which incorporates the effects of linear growth and advective losses due reconnection outflow. Our model predicts a critical Lundquist number below which disruption does not occur. The critical Lundquist number is not a constant value but has a weak dependence on the noise amplitude.   

3D magnetic reconnection studies in the Terrestrial Reconnection Experiment (TREX)

Recent experimental studies on the Terrestrial Reconnection Experiment (TREX) focus on nominally 2D asymmetric magnetic reconnection configurations. In 3D reconnection, however, regions of zero magnetic fields, called magnetic nulls, become important sites of potential particle acceleration. In order to study 3D null-point reconnection in TREX, an adjustable pulsed coil (null coil) is added to the experiment to introduce magnetic nulls. Tilting the null coil relative to the existing reconnection drive and Helmholtz coils forms complex 3D magnetic field geometries, vastly different from other 2D reconnection geometries explored on TREX. To measure the effects of the added null-points, a combination of magnetic probes and Langmuir probes are used to characterize the plasma.   

Endogenous Magnetic Reconnection and Associated High Energy Plasma Processes

The existence of an endogenous magnetic reconnection process in weakly collisional plasmas is proposed that relies on the presence of a significant electron temperature gradient [1] and a local current density gradient within the region where a drastic change of magnetic field topology is produced. The newly identified mode involves widths of the layer in which reconnection takes place that remain relevant even when large macroscopic distances are considered, such as those of interest to space and astrophysics. Given that plasmas in the Universe with considerable electron thermal energy contents are ubiquitous, it makes sense to rely on this process to generate, through magnetic field reconnection, high energy particle populations [1], momentum and angular momentum transport and, in any case, new magnetic field configurations. A depletion of magnetic energy is associated with a suppression of the current density gradient.\\ [1] B. Coppi, B. Basu and A. Fletcher, $\il{Nucl. Fusion}$, $\bf{57}$, 7 (2017).   

Flux-rope distribution function through a Maximum Entropy principle

The Principle of Maximum Entropy (MaxEnt) is utilized for inferring the distribution function of flux ropes formed through a resistive instability as a function of their mass, flux and velocity [1]. Our treatment is 3D (flux ropes) in nature, as opposed to previous works that have studied 2D structures (plasmoids) [2,3]. The distributions for the mass, width, flux and helicity are characterized by a power-law behavior with exponents of -4/3, -2, -3 and -2 respectively for small values, and display an exponential falloff for large values. The velocity distribution is shown to be flat at small values and becomes a power law for large values with an exponent of -7/3. A preliminary comparison with observational evidence suggests that the predictions of the theoretical model are consistent with the latter.\\ References: [1] M. Lingam, L. Comisso & A. Bhattacharjee, arXiv:1702.05782 (2017) [2] D. A. Uzdensky, N. F. Loureiro & A. A. Schekochihin, Phys. Rev. Lett., 105, 235002 (2010) [3] Y.-M. Huang & A. Bhattacharjee, Phys. Rev. Lett., 109, 265002 (2012)   

Survey of Waves in the Ion Diffusion Region of Asymmetric Reconnection in the Dayside Magnetopause

Wave modes in the ion diffusion region of an active dayside reconnection site are characterized using high time-resolution data from NASA's Magnetospheric Multiscale (MMS) mission. We observe high-amplitude, low-frequency ($\leq \omega_{LH}$) electrostatic fluctuations near the magnetospheric-side separatrices, which are most likely a result of the lower hybrid drift instability (LHDI). In the magnetospheric upstream region we find whistler waves with $\omega\approx \Omega_e/2$ and phase velocity near $5\times 10^7\hspace{2pt}m/s$ propagating toward the X-line. In the exhaust region we find a largely different set of electromagnetic waves, including an obliquely-propagating, high-frequency ($\omega>\omega_{LH}$), left-hand-polarized waves with phase velocities around $1\times 10^6\hspace{2pt}m/s$. Possible relationships between these waves and local distribution functions will be investigated through analysis of data from the fast plasma instrument (FPI). We will also discuss the possibility of mode conversion from electrostatic lower hybrid waves to the electromagnetic waves we observe.   

FLARE: A New User Facility for Studies of Multiple-Scale Physics of Magnetic Reconnection and Related Phenomena Through \textbf{\textit{in-situ}} Measurements

The FLARE device (Facility for Laboratory Reconnection Experiments; flare.pppl.gov) is a new laboratory experiment under construction at Princeton for the studies of magnetic reconnection in the multiple X-line regimes directly relevant to space, solar, astrophysical, and fusion plasmas, as guided by a reconnection phase diagram [Ji \& Daughton (2011)]. The whole device have been assembled with first plasmas expected in the fall of 2017. The main diagnostics is an extensive set of magnetic probe arrays, currently under construction, to cover multiple scales from local electron scales ($\sim$2 mm), to intermediate ion scales ($\sim$10 cm), and global MHD scales ($\sim$1 m), simultaneously providing {\it in-situ} measurements over all these relevant scales. The planned procedures and example topics as a user facility will be discussed.   

Onset of magnetic reconnection in a weakly collisional, high-$\beta$ plasma

In a magnetized, weakly collisional plasma, the magnetic moment of the constituent particles is an adiabatic invariant. An increase of the magnetic-field strength in such a plasma thus leads to an increase in the thermal pressure perpendicular to the field lines. Above a $\beta$-dependent threshold, this pressure anisotropy drives the mirror instability, which produces strong distortions in the field lines and traps particles on ion-Larmor scales. The impact of this instability on magnetic reconnection is investigated using simple analytical and numerical models for the formation of a current sheet and the associated production of pressure anisotropy. The difficulty in maintaining an isotropic, Maxwellian particle distribution during the formation and subsequent thinning of a current sheet in a weakly collisional plasma, coupled with the low threshold for the mirror instability in a high-$\beta$ plasma, imply that the topology of reconnecting magnetic fields can radically differ from the standard Harris-sheet profile often used in kinetic simulations of collisionless reconnection. Depending on the rate of current-sheet formation, this mirror-induced disruption may occur before standard tearing modes are able to develop.   

Experimental Investigation of Magnetic Reconnection in Weakly Ionized Plasmas

Magnetic reconnection is a fundamental process in magnetized plasmas wherein stored magnetic energy is rapidly released. While commonly studied in fully ionized plasmas, many important systems (e.g., the chromosphere or interstellar medium) are weakly ionized, which can significantly modify reconnection physics. Recent IRIS observations have enabled detailed studies of chromospheric reconnection, highlighting its in partially ionized systems [1]. Previous experiments on the Magnetic Reconnection Experiment have qualitatively shown that reconnection is slow in partially ionized systems, in contrast to theoretical predictions, although the underlying physics is unclear [2]. Here, new experiments are performed to examine the detailed role of neutrals. An in-situ filterscope has been developed to simultaneously measure Helium line emission at 668, 706, and 728 nm from a localized, 1 cm$^3$ volume with high time-resolution. Collisional radiative modeling is used to determine the neutral density, as well as the electron density and temperature. By measuring the neutral density, the detailed neutral-plasma coupling during reconnection is studied in detail. [1] De Pontieu, Bart, et al. Solar Physics 289.7 (2014) [2] Lawrence, Eric E., et al. Physical review letters 110.1 (2013)   

Electron phase space structure of asymmetric reconnecting current sheets with varying guide fields

Magnetic reconnection in asymmetric current sheets is known to produce highly structured electron velocity distributions, including crescents within the diffusion region [Bessho et al., GRL 2016] and crescents and mixed populations along the separatrices [Egedal et al., PRL (2016), Hesse et al., PRL (2017)]. Using a series of 2D particle-in-cell simulations of asymmetric magnetic reconnection with varying guide magnetic fields, we investigate electron distribution functions in both the upstream and exhaust regions of reconnecting current sheets. Electric fields, especially those parallel to the local magnetic field, play an important role in shaping and producing beams in velocity space. We relate the electron distributions to macroscopic field and plasma structures and compare our PIC results to recent Magnetospheric Multiscale (MMS) Mission measurements during dayside magnetopause reconnection.   

How Alfv\'en waves set the large scale structure of magnetic reconnection.

A PIC simulation of anti-parallel reconnection shows the formation of the out-of-plane or Hall magnetic field that extends hundreds of inertial lengths from the X-line.This structure is generated by field-aligned electron currents that flow outside the magnetic separatrices when ion and electrons decouple on length scales less than $d_{i}$. We observe that this Hall field propagates from the X-point to far downstream into the exhaust along the magnetic field lines at Alfv\'enic speed.Thus the propagation of this large scale reconnection structure can be associated with a Alfv\'en wave generated in the inner electron diffusion region, specifically near the X-line.   

Kubo Resistivity of magnetic flux ropes

Magnetic flux ropes are bundles of twisted magnetic fields and their associated current. They are common on the surface of the sun (and presumably all other stars) and are observed to have a large range of sizes and lifetimes. They can become unstable and resulting in coronal mass ejections that can travel to earth and indeed, have been observed by satellites. Two side by side flux ropes are generated in the LAPD device at UCLA. Using a series of novel diagnostics the following key quantities, \textbf{B,} \textbf{u}, Vp, n, Te have been measured at more than 48,000 spatial locations and 7,000 time steps. Every term in Ohm's law is also evaluated across and along the local magnetic field and the plasma resistivity derived and it is shown that Ohms law is non-local. The electron distribution function parallel and antiparallel to the background magnetic field was measured and found to be a drifting Kappa function. The Kubo AC conductivity at the flux rope rotation frequency, a 3X3 tensor, was evaluated using velocity correlations and will be presented. This yields meaningful results for the global resistivity. Frequency spectra and the presence of time domain structures may offer a clue to the enhanced resistivity.   

Pressure Anisotropy Measurements on the Terrestrial Reconnection Experiment

The Terrestrial Reconnection Experiment (TREX) at the Wisconsin Plasma Astrophysics Laboratory (WiPAL) studies collisionless magnetic reconnection. In this regime, electron pressure anisotropy should develop, deviating from Hall reconnection dynamics and driving large-scale current layer formation [1]. A multi-tip version of the M-probe of Shadman [2], containing 32 Langmuir probe tips and two magnetic coils, measures this anisotropy. Each tip is biased to a different potential, simultaneously measuring discrete parts of the I-V characteristic. Pulsing the coil locally increases the magnetic field near the tips, inducing a magnetic mirror force to reflect electrons with large values of $v_{\perp}/v$. The change in velocity modifies the I-V characteristic and can be used to infer $p_{\parallel}/p_{\perp}$. Results and analysis from the probe are presented. [1] J. Egedal \emph{et al.}, Nature Phys. (2012). [2] K. Shadman, Phys. Plasmas (2004).   

Overview and first results of experiments on magnetic reconnection between colliding magnetized plasmas at the National Ignition Facility

Expanding laser-produced plasmas naturally self-generate magnetic fields by the Biermann battery effect, and the collision of two plumes can drive magnetic reconnection. The National Ignition Facility at LLNL occupies a unique position for laser-driven magnetic reconnection experiments by simultaneously allowing very large plasma temperature, low plasma resistivity, and large system size, which allows observation of secondary instabilities driven during magnetic reconnection and particle acceleration relevant to astrophysical plasmas. Magnetic reconnection experiments have been conducted on the NIF through the NIF Discovery Science program with the first experimental shots conducted in May 2017. We will present the design of the experimental platform and results from the first experimental day. Magnetic reconnection data is obtained from proton radiography based on a DHe3 backlighter, x-ray self-emission, and a new low-energy particle spectrometer (NIF EPPS-300G) developed by the NIF Facility and Engineering and fielded for the first time on these experiments.   

The kinetic structure of the electron diffusion region observed by MMS during asymmetric reconnection

During asymmetric magnetic reconnection in the dayside magnetopause in situ spacecraft measurements by NASA's MMS mission provide new detailed information on the electron dynamics within the electron diffusion region. In particular, we report here on observations by MMS4 which traveled the closest on the topological X-line [1] in the event on October 16, 2015, first reported by Burch et al., [2]. In addition to crescent shaped electron distributions [2,3], the measurements include electron beams flowing in toward the diffusion region. These beams of incoming electrons are formed by $E_{\parallel}$ acceleration along the high-density side separatrices. They penetrate across the electron diffusion region, where their directions are nearly unaffected by the rapid changes in the magnetic field geometry. Matching electron beam features are observed in 2.5D kinetic simulations, revealing their role in breaking the electron frozen-in-law through contributions to the off-diagonal stress in the electron pressure tensor.\\[1ex] [1] Denton et al., Geophys.~Res.~Lett., 43, 5589–5596, (2016).\\[1ex] [2] Burch et al., Science 352, 2939, (2016).\\[1ex] [3] Egedal et al., PRL 117, 185101 (2016).\\[1ex]   

Impact of compressibility and a guide field on Fermi acceleration during magnetic island coalescence

Previous work has shown that Fermi acceleration can be an effective heating mechanism during magnetic island coalescence, where electrons may undergo repeated reflections as the magnetic field lines contract. This energization has the potential to account for the power-law distributions of particle energy inferred from observations of solar flares. Here, we develop a generalized framework for the analysis of Fermi acceleration that can incorporate the effects of compressibility and non-uniformity along field lines, which have commonly been neglected in previous treatments of the problem. Applying this framework to the simplified case of the uniform flux tube allows us to find both the power-law scaling of the distribution function and the rate at which the power-law behavior develops. We find that a guide magnetic field of order unity effectively suppresses the development of power-law distributions.   

Deployment of the OSIRIS EM-PIC code on the Intel Knights Landing architecture

Electromagnetic particle-in-cell (EM-PIC) codes such as OSIRIS [1] have found widespread use in modelling the highly nonlinear and kinetic processes that occur in several relevant plasma physics scenarios, ranging from astrophysical settings to high-intensity laser plasma interaction.~Being computationally intensive, these codes require large scale HPC systems, and a continuous effort in adapting the algorithm to new hardware and computing paradigms. In this work, we report on our efforts on deploying the OSIRIS code on the new Intel Knights Landing (KNL) architecture. Unlike the previous generation (Knights Corner), these boards are standalone systems, and introduce several new features, include the new AVX-512 instructions and on-package MCDRAM. We will focus on the parallelization and vectorization strategies followed, as well as memory management, and present a detailed performance evaluation of code performance in comparison with the CPU code. [1] R. A. Fonseca et al., Lecture Notes in Computer Science \textbf{2331}, 342-351 (2002)   

End-to-end plasma bubble PIC simulations on GPUs

Accelerator technologies play a crucial role in eventually achieving exascale computing capabilities. The current and upcoming leadership machines at ORNL (Titan and Summit) employ Nvidia GPUs, which provide vast computational power but also need specifically adapted computational kernels to fully exploit them. In this work, we will show end-to-end particle-in-cell simulations of the formation, evolution and coalescence of laser-generated plasma bubbles. This work showcases the GPU capabilities of the PSC particle-in-cell code, which has been adapted for this problem to support particle injection, a heating operator and a collision operator on GPUs.   

Simulating The Prompt Electromagnetic Pulse In 3D Using Vector Spherical Harmonics

We describe a new, efficient code for simulating the prompt electromagnetic pulse. In SHEMP ("Spherical Harmonic EMP"), we extend to 3-D the methods pioneered in C. Longmire's CHAP code [1]. The geomagnetic field and air density are consistent with CHAP's assumed spherical symmetry only for narrow domains of influence about the line of sight, limiting validity to very early times. Also, we seek to model inherently 3-D situations. In CHAP and our own CHAP-lite [2], the independent coordinates are r (the distance from the source) and $\tau =$ t-r/c; the pulse varies slowly with r at fixed $\tau $, so a coarse radial grid suffices. We add non-spherically-symmetric physics via a vector spherical harmonic decomposition. For each (l,m) harmonic, the radial equation is similar to that in CHAP and CHAP-lite. We present our methodology [3] and results on model problems. [1] C. L. Longmire, IEEE Trans. Electromag. Compatibility 20 no. 1, 3 (1978). [2] W. A. Farmer, et al., IEEE Trans. Nucl. Sci. 63, 1259 (2016). [3] A. Friedman, et al., submitted to J. Rad. Effects Research and Engr.; also report LLNL-JRNL-732043 (2017).   

Reduced 3d modeling on injection schemes for laser wakefield acceleration at plasma scale lengths

Current modelling techniques for laser wakefield acceleration (LWFA) are based on particle-in-cell (PIC) codes which are computationally demanding. In PIC simulations the laser wavelength $\lambda_0$, in $\mu$m-range, has to be resolved over the acceleration lengths in meter-range. A promising approach is the ponderomotive guiding center solver (PGC)~[1, 2] by only considering the laser envelope for laser pulse propagation. Therefore only the plasma skin depth $\lambda_p$ has to be resolved, leading to speedups of $(\lambda_p/\lambda_0)^2$. This allows to perform a wide-range of parameter studies and use it for $\lambda_0 \ll \lambda_p$ studies. We present the 3d version of a PGC solver in the massively parallel, fully relativistic PIC code OSIRIS~[3]. Further, a discussion and characterization of the validity of the PGC solver for injection schemes on the plasma scale lengths, such as down-ramp injection, magnetic injection~[4] and ionization injection, through parametric studies, full PIC simulations and theoretical scaling, is presented. References: [1] D. Gordon et. al., IEEE Trans. Plasma Sci. 28, 1135 (2000) [2] A. Helm et. al., submitted (2017) [3] R. A. Fonseca et. al., Lect. Notes Comp. Sci., 2331, 343 (2002) [4] J. Vieira et al., Phys. Rev. Lett. 106, 225001 (2011)   

Convergence of the Ponderomotive Guiding Center approximation in the LWFA

Plasma accelerators arose as potential candidates for future accelerator technology in the last few decades because of its predicted compactness and low cost. One of the proposed designs for plasma accelerators is based on Laser Wakefield Acceleration (LWFA). However, simulations performed for such systems have to solve the laser wavelength which is orders of magnitude lower than the plasma wavelength. In this context, the Ponderomotive Guiding Center (PGC) algorithm for particle-in-cell (PIC) simulations is a potent tool. The laser is approximated by its envelope which leads to a speed-up of around 100 times because the laser wavelength is not solved. The plasma response is well understood, and comparison with the full PIC code show an excellent agreement. However, for LWFA, the convergence of the self-injected beam parameters, such as energy and charge, was not studied before and has vital importance for the use of the algorithm in predicting the beam parameters. Our goal is to do a thorough investigation of the stability and convergence of the algorithm in situations of experimental relevance for LWFA. To this end, we perform simulations using the PGC algorithm implemented in the PIC code OSIRIS. To verify the PGC predictions, we compare the results with full PIC simulations.   

Two-stage acceleration of externally injected electrons in plasma bubble derived from the combination of DLA and LWFA.

Simultaneous interactions of accelerated electrons directly with a laser pulse and with a laser wakefield are studied using a novel quasistatic 3D particle-in-cell code. Relativistic electrons externally injected into the plasma bubble's decelerating phase can gain significant energy through the direct laser acceleration (DLA) mechanism from the driving laser pulse [1-3], increasing the amplitude of betatron oscillations. With time, the resonant interaction condition is violated, leading to gradual dephasing between electrons and laser wave, and to eventual slipping of the electrons to the back of the plasma bubble. After that, the oscillating electrons experience the second stage of acceleration gaining energy only from the bubble wakefield. We analyze each stage of acceleration and show that electrons undergoing two stages emits much more X-ray radiation compared with those accelerated during one wakefield stage. This work was supported by DOE grant DESC0007889 and by AFOSR grant FA9550-16-1-0013. [1] X. Zhang, V. N. Khudik and G. Shvets, Phys. Rev. Lett. 114, 184801 (2015). [2] X. Zhang, V. N. Khudik, A. Pukhov and G. Shvets, Plasma Phys. Control. Fusion 58 034011 (2016). [3] V. N. Khudik, X. Zhang and G. Shvets, arXiv: 1610.0945 (2015).   

Ultra-low emittance electron beam generation using ionization injection in a plasma beatwave accelerator

Ultra-low emittance beams can be generated using ionization injection of electrons into a wakefield excited by a plasma beatwave accelerator. This all-optical method of electron beam generation uses three laser pulses of different colors. Two long-wavelength laser pulses, with frequency difference equal to the plasma frequency, resonantly drive a plasma wave without fully ionizing a gas. A short-wavelength injection laser pulse (with a small ponderomotive force and large peak electric field), co-propagating and delayed with respect to the beating long-wavelength lasers, ionizes a fraction of the remaining bound electrons at a trapped wake phase, generating an electron beam that is accelerated in the wakefield. Using the beating of long-wavelength pulses to generate the wakefield enables atomically-bound electrons to remain at low ionization potentials, reducing the required amplitude of the ionization pulse, and, hence, the initial transverse momentum and emittance of the injected electrons. An example is presented using two lines of a CO$_2$ laser to form a plasma beatwave accelerator to drive the wake and a frequency-doubled Ti:Al$_2$O$_3$ laser for ionization injection.   

Emittance preservation in plasma-based accelerators with ion motion

In a plasma-accelerator-based linear collider, the density of matched, low-emittance, high-energy particle bunches required for collider applications can be orders of magnitude above the background ion density, leading to ion motion, nonlinear focusing fields, and, hence, to beam emittance growth. By analyzing the response of the background ions to an ultra-high density beam, analytical expressions, valid for nonrelativistic ion motion, are derived for the transverse wakefield and for the final (i.e., after saturation) bunch emittance. Analytical results are validated against numerical modeling. A class of initial beam distributions are derived that are equilibrium solutions, which require head-to-tail bunch shaping, enabling emittance preservation with ion motion.   

Plasma Channel Lenses and Plasma Tornadoes for Optical Beam Focusing and Transport

Shaped plasmas offer the possibility of manipulating laser pulses at intensities far above the damage limits for conventional optics. An example is the plasma channel, which is a cylindrical plasma column with an on-axis density minimum. Long plasma channels have been widely used to guide intense laser pulses, particularly in laser wakefield accelerators. A new concept, the ``plasma tornado'', offers the possibility of creating long plasma channels with no nearby structures and at densities lower than can be achieved by capillary discharges. A short plasma channel can focus a laser pulse in much the same manner as a conventional lens or off-axis parabola. When placed in front of the focal point of an intense laser pulse, a plasma channel lens (PCL) can reduce the effective f-number of conventional focusing optics. When placed beyond the focal point, it can act as a collimator. We will present experimental and modeling results for a new plasma tornado design, review experimental methods for generating short PCLs, and discuss potential applications.   

Filtering of higher-order laser modes using plasma structures

Plasma structures based on leaky channels are proposed to filter higher-order laser mode content. The evolution and propagation of non-Gaussian laser pulses in leaky channels is studied, and it is shown that, for appropriate laser-plasma parameters, the higher-order laser mode content may be removed while the fundamental mode remains well-guided. The behavior of the multi-mode laser pulse is described analytically, including the derivation of the leakage coefficients, and compared to numerical calculations. Gaussian laser pulse propagation, without higher-order mode content, improves guiding in parabolic plasma channels, enabling extended interaction lengths for laser-plasma accelerator applications.   

Abstract Withdrawn

Plasma-based accelerators can generate about 1000 times stronger acceleration field compared with RF-based conventional accelerators, which can be done by high power laser and plasma. There are many issues in this research and one of them is development of a good plasma source for higher electron beam energy. For this purpose, we are investigating a special type of plasma source, which is a density-tapered gas cell with a mixed-gas for easy injection. By this type of special gas cell, we expect higher electron beam energies with easy injection in the wakefield. In this poster, some experimental results for electron beam generation with the density-tapered mixed-gas cell are presented. In addition to the experimental results, CFD (Computational-Fluid-Dynamics) and PIC (Particle-In-Cell) simulation results are also presented for comparison studies.   

Wakefield simulation of solid state plasma

Although it is known that the accelerating gradient of wakefield increases when laser frequency increases (i.e. critical density), there was no adequate technology to make intense X-ray laser until recently. With the advent of the invention of Thin Film Compression, we now see the intense X-ray laser technology that fits this need [1]. We have modified the EPOCH PIC code to include the lattice effect [2] of nanomaterials in our simulations. The present results indicate the accelerating gradient \(\sim0.3\frac{TeV}{cm}\) at the plasma density of \(10^{23} cm^{-3}\) which agree well with the wakefield theory. This shows the concept of the solid state plasma wakefield in nanomaterials is validated by computation. This result is also consistent with previous findings [3], in which the lattice effect was neglected. [1] Tajima, T. (2014). Eur. Phys. J. \textbf{223}, 1037. [2] Tajima, T. {\&} Ushioda, S. (1978). Phys. Rev. B \textbf{18}, 1892. [3] Zhang, X. et al., (2016). Phys. Rev. Acc. Beams \textbf{19}, 101004.   

Relativistic Electron Acceleration by Surface Plasma Waves in the High Intensity Regime

High field plasmonics is one of the new research fields which has synergetically benefited from the advances in laser technology. The availability of radiation fields of intensities exceeding 10$^{\mathrm{18\thinspace }}$W/cm$^{\mathrm{2}}$ brought plasmonics into a new regime where relativistic and nonlinear effects start to dominate the dynamics of the surface plasma waves (SPWs). Moreover, surface plasma waves are a very efficient route to transfer the laser energy to the secondary sources including laser driven particle and radiation beams and to control and optimize the physical properties of these sources. We present here experimental evidence of a novel regime of the SPWs excitation by ultra-high intensity laser field (I\textgreater 10$^{\mathrm{20\thinspace }}$W/cm$^{\mathrm{2}})$ on grating targets and its effect on high energy surface electron acceleration. The peak of the electron emission was detected at a laser incidence angle of 45\textdegree . The results indicate new conditions for resonant excitation of SPWs since in the limit of the linear regime (moderate intensities of \textasciitilde 10$^{\mathrm{19\thinspace }}$W/cm$^{\mathrm{2}}$ and step preplasma profile), the resonance angle is predicted at 30\textdegree . 2D PIC simulations and a novel analytical model confirm the experimental data and reveal that, at laser intensities above 10$^{\mathrm{20\thinspace }}$W/cm$^{\mathrm{2}}$, nonlinearities induced by the preplasma condition and relativistic effects change the SPWs resonance.   

An overview of Laser-Produced Relativistic Positrons in the Laboratory

The production of relativistic positrons using ultraintense lasers can facilitate studies of fundamental pair plasma science in the relativistic regime and laboratory studies of scaled energetic astrophysical mechanisms such as gamma ray bursts. The positron densities and spatial scales required for these applications, however, are larger than current capabilities. Here, we present an overview of the experimental laser-produced positron results and their respective modeling for both the direct laser-irradiated process and the indirect process (laser wakefield accelerated electrons irradiating a high-Z converter). Conversion efficiency into positrons and positron beam characteristics are compared, including total pair yield, mean energy, angular divergence, and inferred pair density for various laser and target conditions. Prospects towards increasing positron densities and beam repetition rates will also be discussed.   

Multi-GeV electron-positron pair generation from laser-electron scattering

Positron generation in the laboratory is of great importance, both for fundamental science and potential applications. For laboratory astrophysics, it is particularly important to produce neutral electron-positron plasma, with properties that allow studying their collective behavior. Electron-positron pairs can be generated by first emitting an energetic photon, that later decays into a pair in an intense background field (Breit-Wheeler process). Recently, several experiments demonstrated that high-frequency radiation can be generated in laser-electron beam scattering. Here we propose a new scattering configuration that can both generate electron-positron pairs, and later accelerate them to multi-GeV energies. This configuration allows obtaining an e+e- flow with a higher energy than that of the initial electron beam. We develop an analytical model that predicts the energy cutoff.  We discuss the number of pairs expected, the acceleration and the overall quality of the beam. We also study the role of pulse duration and spatiotemporal synchronization for the overall number of pairs. The work is supported by OSIRIS QED-PIC simulations, and these ideas can be tested with a new generation laser system at ELI Beamlines that will provide 10 PW peak power in a 150 fs pulse duration.   

Plasma diagnosis using the laser-plasma terahertz(THz) radiation

There has been growing interest in laser-plasma terahertz(THz) radiation as a source for plasma diagnostics for its strong radiation energy and broadband characteristics. Yet, it has not been used in actual diagnostics systems as its validity was not thoroughly investigated. In this work, we measured the plasma density in our inductively-coupled plasma(ICP) chamber with the laser-plasma THz radiation. The THz time-domain spectroscopy(THz-TDS) and single-shot detection method were used for the detection of the THz. The diagnosis results from the two different detection methods were consistent with each other, giving a possibility of the realization of the laser-plasma-THz-based plasma diagnostics.   

Directed high-power THz radiation from transverse laser wakefield excited in an electron density filament

A tightly focused femtosecond, weakly relativistic laser pulse partially ionizes the ambient gas, creating a string (a ``filament'') of electron density, locally reducing the nonlinear index and compensating for the self-focusing effect caused by bound electrons. While maintaining the filament over many Rayleigh lengths, the pulse drives inside it a three-dimensional (3D) wave of charge separation - the plasma wake. If the pulse waist size is much smaller than the Langmuir wavelength, electron current in the wake is mostly transverse. Electrons, driven by the wake across the sharp radial boundary of the filament, lose coherence within 2-3 periods of wakefield oscillations, and the wake decays [J.-R. Marques et al., Phys. Plasmas 5, 1162 (1998)]. The laser pulse is thus accompanied by a short-lived, almost aperiodic electron current coupled to the sharp index gradient. The comprehensive 3D hydrodynamic model shows that this structure emits a broad-band THz radiation, with the highest power emitted in the near-forward direction. The THz radiation pattern contains information on wake currents surrounding the laser pulse, thus serving as an all-optical diagnostic tool. The results are tested in cylindrical and full 3D PIC simulations using codes WAKE and EPOCH.   

Terahertz Radiation from Laser Created Plasma by Applying a Transverse Static Electric Field

Terahertz (THz) radiation, which is emitted in narrow cone in the forward direction from femtosecond laser pulse created plasma has been observed by N.Yugami \textit{et al}. [1]. Additionally, T. Loffler \textit{et al}. have observed that a significantly increased THz emission intensity in the forward direction when the transverse static electric field is applied to the plasma [2]. We propose the theoretical model of the THz radiation from laser created plasma by applying the transverse static electric field and conducted both experiments and 2D-PIC simulation to compared with our theory.\\ \\$[1]$ N. Yugami \textit{et al}, Jpn. J. Appl. Phys. \textbf{45,} L1051 (2006). [2] T. Loffler \textit{et al}, Appl. Phys. Lett. \textbf{77}, 453 (2000).   

High-efficiency gamma-ray flash generation from multiple-laser scattering

Gamma-ray flash generation in a near-critical-density target irradiated by four symmetrical colliding laser pulses is numerically investigated. With peak intensities about $10^{23}$ W/cm$^2$, the laser pulses boost electron energy through direct laser acceleration, while pushing them inward with the ponderomotive force. After backscattering with counter-propagating laser, the accelerated electron is trapped in the electromagnetic standing waves of the ponderomotive potential well created by the coherent overlapping of the laser pulses. Electrons emit gamma -ray photons in a multiple-laser-scattering regime, where the electrons act as a medium transferring energy from the laser to gamma-rays in the ponderomotive potential valley [Z. Gong et al, PRE 95, 013210 (2017)].   

Gamma-ray Radiation From Plasma Bubble Hosing

The CEP-dominated few cycle strong ( \begin{figure}[htbp] \centerline{\includegraphics[width=0.40in,height=0.17in]{100720171.eps}} \label{fig1} \end{figure} ) laser pulse could oscillate in a underdense plasma with a period [1], \begin{figure}[htbp] \centerline{\includegraphics[width=1.03in,height=0.19in]{100720172.eps}} \label{fig2} \end{figure} , where \begin{figure}[htbp] \centerline{\includegraphics[width=0.14in,height=0.17in]{100720173.eps}} \label{fig3} \end{figure} is the laser wavelength, \begin{figure}[htbp] \centerline{\includegraphics[width=0.14in,height=0.17in]{100720174.eps}} \label{fig4} \end{figure} is the laser critical density and \begin{figure}[htbp] \centerline{\includegraphics[width=0.15in,height=0.18in]{100720175.eps}} \label{fig5} \end{figure} is the initial plasma density. This oscillation further leads to the hosing-like oscillation of the formed plasma bubble [2] which, in turn, gives a very strong oscillation strength for the electrons trapped inside. With numbers of self-trapped electrons, this scheme is capable server as a strong and bright gamma-ray source. A stretched plasma bubble is achieved by firstly injecting a symmetric, moderately long ( \begin{figure}[htbp] \centerline{\includegraphics[width=0.24in,height=0.18in]{100720176.eps}} \label{fig6} \end{figure} ) and strong ( \begin{figure}[htbp] \centerline{\includegraphics[width=0.40in,height=0.17in]{100720177.eps}} \label{fig7} \end{figure} ) laser in to an underdense plasma. Then, many electrons can be self-trapped along with the bubble breaking due to the nonlinear plasma wave. The head erosion produces the few cycle pulse which enables the oscillation. \textbf{References:} [1] A. A. Silaev et al., Residual-Current Excitation in Plasmas Produced by Few-Cycle Laser Pulses, PhysRevLett.102.115005 (2009); [2] M. C. Kaluza et al., Observation of a Long-Wavelength Hosing Modulation of a High-Intensity Laser Pulse in Underdense Plasma, Phys. Rev. Lett.~\textbf{105}, 095003 (2010).   

Temporal characterization of the wave-breaking flash in a laser plasma accelerator

Wave-breaking injection of electrons into a relativistic plasma wake generated in near-critical density plasma by sub-terawatt laser pulses generates an intense ($\sim$1$\mu$J) and ultra-broadband $(\Delta\lambda\sim300nm)$ radiation flash [1]. In this work we demonstrate the spectral coherence of this radiation and measure its temporal width using single-shot supercontinuum spectral interferometry (SSSI). The measured temporal width is limited by measurement resolution to ~50 fs. Spectral coherence is corroborated by PIC simulations which show that the spatial extent of the acceleration trajectory at the trapping region is small compared to the radiation center wavelength. To our knowledge, this is the first temporal and coherence characterization of wave-breaking radiation.\newline [1] Goers, A. J., et al. "Multi-MeV electron acceleration by subterawatt laser pulses." Physical review letters 115.19 (2015): 194802.   

Impact of a plasma channel on the emission of directed high-energy photons in laser-plasma interaction

\hfill \newline Compact sources of directed high-energy photons are of great interest in current research. Common sources of high-energy photons include synchrotrons and other large and expensive accelerators. Laser-plasma interactions promise sources that are significantly smaller and cheaper than conventional ones. However, they come at the cost of producing either only small number of photons or very undirected ones. A recent study[1] shows, that the use of a plasma channel is able to significantly mitigate these problems while producing a large number of high energy, well collimated photons. We provide an analysis on the physical processes, that lead to the formation of strong magnetic fields responsible for this improvement of emission. Furthermore, we investigate the channel properties in relation to a given laser pulse.\\ \hfill \newline References: [1] Stark \textit{et al.} PRL {\bf 116},185003 (2016).   

X-Ray and electron beam source characterization from Self-Modulated Laser Wakefield Acceleration experiments at Titan

The development of a directional, low-divergence, and short-duration (ps and sub-ps) x-ray probes with energies of tens of keV is desirable for the fields of astrophysics, High Energy Density Science and Inertial Confinement Fusion. In this work we focused the Titan laser beam (1 ps and 150 Joules) into a 4mm helium gas jet to produce an electron beam that in turn generates an x-ray beam. The measured Raman Forward Scattering satellites present in the laser spectrum after the interaction, indicate the generation of a Self-modulated laser wakefield accelerator. This accelerator produced an electron beam with energies up to 250 MeV, a divergence of 16 x 40 mrad and a total charge of 6 nC. Using this high-charge relativistic electron beam we explored the combination of three mechanisms to produce an x-ray beam: Betatron, Compton scattering and Bremsstrahlung. We show the generation of a low divergence (~mrad), small source size (~um) broadband (keV to MeV) x-ray beam that can be used as a backlighter for time-resolved spectroscopy, imaging, and Compton radiography.   

Dynamics of multiple interacting ultra-intense lasers in a plasma

Although light beams do not interact in vacuum, two intense lasers can interact with each other in a nonlinear medium, forming spiraling and braiding patterns when they carry angular momentum. Here, we analyse the interaction between $N$ laser filaments. Using a variational approach based on the Non-Linear Schr\"odinger Equation with the lowest order relativistic mass correction, we obtained a mechanical analog of an exponential $N$-body problem for the beam centroids and spot-sizes. We found that the $N$ beams can spiral in a circular motion around the center of mass analog and determined orbital parameters and trapping criteria for this motion. Furthermore, the lasers could form a “solar system”, with $N-1$ beams orbiting a high-power one. The predictions for both models were tested in Particle-in-Cell simulations, in which other effects, namely non-instantaneous spatiotemporal non-linearities, can be compared to the relativistic mass correction. Finally, the initial configuration of the beams could follow from filamentary processes, with the filamentation of an orbital angular momentum carrying laser originating the spiralling and braiding motions of the filaments.   

Coupling the photon kinetics of soft photons with high energy photons

The description of electromagnetic fields based on the generalized photon kinetic theory, which takes advantage of the Wigner-Moyal description for the corresponding classical field theory, is capable of capturing collective plasma dynamics in the relativistic regime driven by broadband incoherent or partially coherent sources. We explore the possibility to extend this description to include the dynamics of hard photons in the plasma, whose interaction is dominated by single scattering processes. Examples of the modification of classical plasma instabilities due to the presence of hard photons is discussed.   

X-Ray Pulse Compression using Stimulated Brillouin Scattering in Plasma

Stimulated Brillouin scattering may allow cleaning and compression of the output from x-ray free-electron lasers, producing coherent sub-femtosecond pulses with intensities orders-of-magnitude beyond current sources. In contrast to stimulated Raman scattering, which is limited by damping at short wavelengths, particle-in-cell simulations and analytic models suggest that amplification by Brillouin scattering is possible in solid-density plasma at the wavelengths and intensities of free-electron lasers. The nonlinear amplification process is robust to quasi-incoherence in the pump beam, permitting beam cleaning in addition to compression.\footnote{M. R. Edwards, J. M. Mikhailova, and N. J. Fisch, ``X-ray amplification by stimulated Brillouin scattering," arXiv:1705.08599, to appear in Phys. Rev. E (2017).}   

Development of PETAL diagnostics: PETAPhys project

Beginning of autumn 2017, PETAL, a Petawatt laser beam, will be operated for experiments on the LMJ facility at the CEA/ Cesta research center. The PETAPhys project provides a support to the qualification phase of the PETAL laser operation. Within the PETAPhys project, we are developing two simple and robust diagnostics permitting both to characterize the focal spot of the PETAL beam and to measure the hard X-ray spectrum at each shot. The first diagnostic consists in optical imaging of the PETAL beam focal spot in the spectral range of the second and third harmonic radiation emitted from the target. The second diagnostic is a hard X-ray dosimeter consisting in a stack of imaging plates (IP) and filters, either placed inside a re-entrant tube or inserted close to target. Numerical simulations as well as experiments on small scale facilities have been performed to design these diagnostics. If available, preliminary results from PETAL experiments will be discussed.   

Laser-Bioplasma Interaction: Excitation and Suppression of the Brain Waves by the Multi-photon Pulsed-operated Fiber Lasers in the Ultraviolet Range of Frequencies

The novel study of the laser excitation-suppression of the brain waves\footnote{Tae Kim et. al. Cortically projecting basal forebrain parvalbumin neurons regulate cortical gamma band oscillations, Proceedings of the National Academy of Sciences, vol. 112 no. 11, 3535--3540, (2015).} is proposed. It is based on the pulsed-operated multi-photon fiber-laser interaction with the brain parvalbumin (PV) neurons.\footnote{Stefan, APS-March-2017, Abstract: M1.00291; V. Alexander Stefan, Neurophysics\textit{, Stem Cell Physics, and Genomic Physics: Beat-Wave-Driven-Free Electron Laser Beam Interactions with the Living Matter}, (S-U-Press, La Jolla, CA, 2012); Stefan, APS-PPD, 2016, Abstract: JP10.00166; V. Stefan, B. I. Cohen, C. Joshi, \textit{Science}, 243, 4890, (1989).\par } The repetition frequency matches the low frequency brain waves (5-100 Hz); enabling the resonance-scanning of the wide range of the PV neurons (the generators of the brain wave activity). The tunable fiber laser frequencies are in the ultraviolet frequency range, thus enabling the monitoring of the PV neuron-DNA, within the 10s of milliseconds. In medicine, the method can be used as an ``instantaneous-on-off anesthetic.''   

Thin film compression of a high rep rate laser: towards single-cycle pulse generation

Thin Film Compression (TFC) has been proposed [1] as a means of increasing the peak power of ultrashort laser pulses with millijoule and greater energies. As opposed to increasing peak power by increasing the energy of the pulse, TFC instead achieves an amplification in peak power by compressing laser pulse length at a fixed energy. This pulse compression is accomplished by the generation of linearly chirped bandwidth through self-phase modulation, which is then recompressed by dispersive optics. A summary of the results of laser pulse compression experiments towards the production of single-cycle pulses will be presented. Simulations show [1] that laser pulses compressed to the single-cycle regime have the potential to generate single-cycle x-ray pulses which could be used to generate wakefields in solid-density plasma with acceleration gradients of up to TeV/cm [2]. [1] G. Mourou et al., Eur. Phys. J. 223, 1181 (2014) \newline [2] T. Tajima, Eur. Phys. J. 223, 1037 (2014); X.M. Zhang et al., Phys. Rev. Accel Beams. 19, 101004 (2016)   

Narrow bandwidth Laser-Plasma Accelerator driven Thomson photon source development

Compact, high-quality photon sources at MeV energies can be provided by Thomson scattering of a laser from the electron beam of a Laser-Plasma Accelerator (LPA). Recent experiments and simulations demonstrate controllable LPAs in the energy range appropriate to MeV sources. Simulations indicate that high flux with narrow energy spread can be achieved via control of the scattering laser pulse shape and laser guiding, and that undesired background bremsstrahlung can be mitigated by plasma based deceleration of the electron beam after photon production. Construction of experiments and laser capabilities to combine these elements will be presented, along with initial operations, towards a compact photon source system. Work supported by US DOE NNSA DNN R&D and by Sc. HEP under contract DE-AC02-05CH11231.   

Developing DIII-D To Prepare For ITER And The Path To Fusion Energy

DIII-D pursues the advancement of fusion energy through scientific understanding and discovery of solutions. Research targets two key goals. First, to prepare for ITER we must resolve how to use its flexible control tools to rapidly reach Q$=$10, and develop the scientific basis to interpret results from ITER for fusion projection. Second, we must determine how to sustain a high performance fusion core in steady state conditions, with minimal actuators and a plasma exhaust solution. DIII-D will target these missions with: (i) increased electron heating and balanced torque neutral beams to simulate burning plasma conditions (ii) new 3D coil arrays to resolve control of transients (iii) off axis current drive to study physics in steady state regimes (iv) divertors configurations to promote detachment with low upstream density (v) a reactor relevant wall to qualify materials and resolve physics in reactor-like conditions. With new diagnostics and leading edge simulation, this will position the US for success in ITER and a unique knowledge to accelerate the approach to fusion energy.   

Parameter exploration for a Compact Advanced Tokamak DEMO

A new parameter study has explored a range of design points to assess the physics feasibility for a compact 200MWe advanced tokamak DEMO that combines high beta ($\beta_N<4$) and high toroidal field ($B_T=6-7$T). A unique aspect of this study is the use of a FASTRAN modeling suite that combines integrated transport, pedestal, stability, and heating & current drive calculations to predict steady-state solutions with neutral beam and helicon powered current drive. This study has identified a range of design solutions in a compact ($R_0=4$m), high-field ($B_T=6-7$T), strongly-shaped ($\kappa=2$, $\delta=0.6$) device. Unlike previous proposals, C-AT DEMO takes advantage of high-beta operation as well as emerging advances in magnet technology to demonstrate net electric production in a moderately sized machine. We present results showing that the large bootstrap fraction and low recirculating power enabled by high normalized beta can achieve tolerable heat and neutron load with good H-mode access. The prediction of operating points with simultaneously achieved high-confinement ($H_{98}<1.3$), high-density ($f_{GW}<1.3$), and high-beta warrants additional assessment of this approach towards a cost-attractive DEMO device.   

A U.S. Strategy for Timely Fusion Energy Development

Worldwide energy demand is expected to explode in the latter half of this century. In anticipation of this demand, the U.S. DOE recently asked the National Academy of Science to provide guidance on a long-term strategic plan assuming that ``economical fusion energy within the next several decades is a U.S. strategic interest.`` Delivering on such a plan will require an R{\&}D program that delivers key data and understanding on the building blocks of a) burning plasma physics, b) optimization of the coupled core-edge solution, and c) fusion nuclear science to inform the design of a cost-attractive DEMO reactor in this time frame. Such a program should leverage existing facilities in the U.S. program including ITER, provide substantive motivation for an expanding R{\&}D scope (and funding), and enable timely redirection of resources within the program as appropriate (and endorsed by DOE and the fusion community). This paper will outline a potential strategy that provides world-leading opportunities for the research community in a range of areas while delivering on key milestones required for timely fusion energy development.   

Creating Hybrid Plasmas With Off-Axis ECCD for Radiating Divertor Studies in DIII-D

A long duration, high density, high power hybrid scenario has been developed for use in radiative divertor studies in DIII-D. Using 11.2 MW of co-NBI power and 3.4 MW of ECCD, with a total injected energy of up to 56 MJ, high performance hybrid plasmas with $\beta_{\mathrm{N\thinspace }}=$ 3.7 and H$_{\mathrm{98y2\thinspace }}=$ 1.5 were created. The hybrid plasmas were fully non-inductive at densities of n $\approx $ 4.2 \texttimes 10$^{\mathrm{19}}$ m$^{\mathrm{-3}}$ with central ECCD, but the EC deposition needed to be moved to $\rho \quad =$ 0.45 to avoid the right-hand cutoff when the density was raised to n $\approx $ 5.8 \texttimes 10$^{\mathrm{19}}$ m$^{\mathrm{-3}}$ for radiative divertor studies. Although moving the EC deposition to $\rho \quad =$ 0.45 had the effect of dropping $\tau_{\mathrm{E}}$ by 10{\%}, the energy confinement time increased with higher density like $\tau_{\mathrm{E\thinspace }}\propto $ n$^{\mathrm{0.4}}$, allowing high beta to be maintained. While the plasma current profile displays the usual self-organizing properties of hybrids -- an anomalously broad profile with q$_{\mathrm{min\thinspace }}$\textgreater 1 -- local current drive can still have a measurable effect on stability, either positively or negatively. For example, hybrid discharges with radial ECH deposited at $\rho \quad =$ 0.45 proved to be more robustly stable to n $=$ 1 modes (can be either a 1/1 or 2/1 mode) than similar discharges with co-ECCD at the same location. Interestingly, the large 1/1 mode had almost no effect on energy confinement but strongly degraded particle confinement; thus this mode needed to be suppressed to achieve the high pedestal densities required for radiative divertor studies.   

Energy Confinement Improvement with Density in Gas Puff Fueled High Performance Hybrid Plasmas on DIII-D.

In contrast to behavior at moderate NBI heating, an increase in energy confinement is observed in high power, near double null, hybrid discharges in response to D2 gas puff fueling. This difference is tied to how the H-mode pedestal responds to fueling. At low power the pedestal width decreases with gas puff resulting in a strong reduction in the critical pressure gradient for the ballooning branch of the peeling-ballooning mode. At high power the pedestal width remains fixed and access to high pedestal pressure gradient with increased collisionality along the peeling boundary is maintained, allowing the pedestal pressure to increase with density. Access to the high pressure gradient peeling limited regime is also blocked by the ballooning boundary if the shape departs from near double null or q decreases relative to the favorable conditions: DRSEP $\approx $ 2.5mm, q95 $\approx $ 6.   

Magnetic Flux Conversion in the DIII-D Steady-State Hybrid Scenario*

The hybrid is a promising high confinement scenario for ITER. The broader current profile aids discharge sustainment by raising $q_{min} >$ 1 thereby avoiding sawtooth-triggered 2/1 tearing modes. In DIII-D hybrid scenario discharges, the rate of poloidal magnetic energy consumption is more than the rate of energy flow from the poloidal field coils. This is evidence that there is a conversion of toroidal flux to poloidal flux, which may be responsible for the anomalous broadening of the current profile known as flux pumping. The rate of poloidal flux being provided and consumed was tracked with coil and kinetic flux states [1]. During long stationary intervals (~1.5 seconds) with constant stored magnetic energy, a significant flux state deficit rate $>$10 mV was observed. The inequality in the evolution of the flux states was observed in hybrids that were 100$\%$ non-inductive and with successful RMP ELM suppression.\\ $\ast$Work supported by the US DOE under DE-FC02-04ER54698 and DE-AC05-06OR23100. [1] T.C. Luce, Nuclear Fusion {\bf 54}, 093005 (2014)   

Reduced turbulence and H-mode confinement in L-mode negative triangularity discharges on DIII-D

DIII-D has produced inner-wall limited plasmas with an L-mode edge at negative triangularity characterized by confinement and fluctuation levels comparable to those in H-mode plasmas at positive triangularity. On TCV, similar plasmas at low collisionality and with pure electron heating showedimproved energy confinement, as compared to matched discharges at positive triangularity, due to modifications to the toroidal precession drift of trapped electrons. The recent DIII-D experiment used both ECH and NBI heating, thus exploring a more reactor relevant regime where Te\textasciitilde Ti. Compared to matched positive triangularity discharges, the intensity of density and temperaturefluctuations is reduced at negative triangularity both in ECH and in NBI dominated phases. Preliminary TGLF runs indicate the discharges are dominated by TEM modes. More detailed analysis will explore the role of the toroidal precession drift in this new regime.   

Turbulent Ion Fluctuation Measurements in Negative Triangularity Plasmas

A new detector array on the UF-CHERS (Ultra Fast CHarge Exchange Recombination Spectroscopy) diagnostic at DIII-D has resulted in significantly improved signal to noise ratio and sensitivity to ion thermal fluctuations. UF-CHERS measures local, long-wavelength Carbon density, ion temperature, and toroidal velocity fluctuations at turbulence-relevant spatiotemporal scales (1 $\mu $s time resolution, \textasciitilde 1 cm spatial resolution which is approximately the turbulence correlation length) from emission of the CVI n$=$8$\to $7 transition. UF-CHERS and BES fluctuation measurements were obtained in equivalent positive and negative triangularity ($\delta )$ discharges with an L-mode edge to compare with theoretical models of turbulence-driven transport and elucidate the mechanisms for improved confinement with negative-$\delta $. Finite coherence is observed between UF-CHERS and co-located BES channels, demonstrating that critical multifield fluctuations such as \textless n*T$_{\mathrm{i}}$\textgreater and \textless n*v$_{\mathrm{tor}}$\textgreater can be measured. Initial analysis shows positive-$\delta $ has radially decreasing, low coherency between \textasciitilde 20-200 kHz for main ion density (BES) and Carbon density, ion temperature, and toroidal rotation (UF-CHERS) fluctuations.   

A Conformal Conducting Wall for Robust Stability of High $\beta_N$, Fully Noninductive Discharges in DIII-D

A conducting surface inside the DIII-D vacuum vessel, closer to the plasma, can increase the ideal-wall MHD stability limit above the high normalized beta ($\beta_N$) needed for 100\% noninductively-driven current at power plant relevant $q_{95}$. In discharges modeled with the planned heating/current drive upgrades, the required $\beta_N$ is as high as 5. This is roughly the calculated limit for n = 1 ideal-wall stability, even with a broad current density profile designed to couple well to the present conducting wall. Tearing and resistive wall modes will very likely limit $\beta_N$ to a value that is lower, but which is expected to scale with the ideal-wall limit. Conceptual designs for an axisymmetric wall that better matches the plasma shape raise the ideal-wall stability limited $\beta_N$ above 7. Analysis with VALEN of a 3-D wall model predicts $\beta_N\sim 6.4$. Increased stability margins are also expected for a wide range of DIII-D discharge scenarios even without a broad current density profile.   

Developing on DIII-D a QH-mode edge solution for Q$=$10 in ITER.

Experiments on DIII-D are advancing toward a demonstration of access to and control of Q(uiescent)H-mode at normalized performance equivalent to Q$=$10 in plasmas with ITER shape and collisionality, and low NBI torque throughout the pulse length. Earlier experiments had shown that ELMs return at NBI torque magnitude below \textasciitilde 3.5 Nm, which is above the expected level of normalized torque input in ITER. Recent experiments have been able to extend QH-mode operation to lower torque \textasciitilde 2 Nm and higher energy confinement by increasing the plasma-wall gap at the outboard mid-plane, and by reducing the input energy of the beam-injected ions. Two hypotheses are being tested: (a) the outer gap affects the scrape-off layer flows and thus the rotation gradient at the edge of the plasma, known to affect QH-mode access; (b) fast ion losses bombarding the wall may release impurities that affect the edge collisionality and stability making QH-mode operation more difficult.   

Mutually Exclusive Relation between High Pedestal and Large-Radius Internal Transport Barrier in High Betap Scenario on DIII-D

Statistical analysis of experimental data from DIII-D high betap plasmas implies that a natural boundary exists hindering the plasma to simultaneously achieve high pedestal (electron temperature and density) and strong large-radius internal transport barrier (ITB) . In the previous study, we revealed a betap threshold about 1.9 for the formation of large-radius ITB in both Te and ne channels. With strong gas puffing, we observed higher betap threshold (about 2.2) for the formation of ne-ITB that may be due to (1) the higher edge density and pedestal height and therefore high local bootstrap current; (2) the penetration of edge inductive current and turbulence. Meanwhile, the betap threshold for the formation of Te-ITB doesn't change. The observed mutually exclusive relation in experiments is important because sustaining a large-radius ITB is favorable for developing high betap scenario with optimized confinement and stability.   

Stability analysis of the high poloidal bet scenario on DIII-Dtowards operation athigher plasma current

The high poloidal beta scenario with plasma current I$^{\mathrm{P}}$\textasciitilde 600 kA and large-radius internal transport barrier (ITB) on DIII-D is subject to n$=$1 MHD kink modes when the current profile becomes very broad at internal inductance values li\textasciitilde 0.5-0.6. It is desirable to extend this scenario to higer plasma current (\textasciitilde 1 MA) for highernormalized fusionperformance. However, higher current at constant normalized beta, ?$^{\mathrm{N}}$\textasciitilde 3, would reducethe poloidal bet, ?$^{\mathrm{P}}$, below the threshold for ITB sustainment, observed at ?$^{\mathrm{P}}$\textasciitilde 1.9. Thus, to avoid loss ofthe IT,?$^{\mathrm{N??}}$ must be increased together with I$^{\mathrm{P}}$ while avoiding the kink instability. MHD analysis is presented that explains possible paths to high?$^{\mathrm{N}}$stability limit for the kink mode in tis scenario. *Work supported by National Magnetic Confinement Fusion Program of Chin under 2015GB110001 and 2015GB102000£¬National Natural Science Foundation of China under Grant No. 1147521 and by US DOE under DE-FC02-04ER54698.   

Extension of high poloidal beta scenario in DIII-D to lower q95 for steady state fusion reactor

DIII-D/EAST joint experiments have improved the high poloidal beta scenario with sustained large-radius internal transport barrier (ITB) extended to high plasma current I$_{\mathrm{p}}$\textasciitilde 1MA with q95\textasciitilde 6.0. Slight off-axis NBCD is applied to obtain broader current density profile, ITBs can now be sustained below the previously observed $\beta_{\mathrm{p}}$ threshold with excellent confinement (H$_{\mathrm{98y2}}$\textasciitilde 1.8). The scenario also exhibits a local negative shear appearing with q increased at rho\textasciitilde 0.4, which helps ITB formation and sustainment. This confirms TGLF prediction that negative magnetic shear can help recover ITB and achieve high confinement with reduced q95. Detailed analysis shows that the Shafranov shift and q profile is critical in the ITB formation at high $\beta _{\mathrm{p\thinspace }}$regime.   

Central Safety Factor and Normalized Beta Control Under Near-Zero Input Torque Constraints in DIII-D

DIII-D experiments have assessed the capability of combined central safety factor (q$_{\mathrm{0}})$ and normalized beta ($\beta_{\mathrm{N}})$ control under near-zero net torque to facilitate access to QH-mode with reverse I$_{\mathrm{p}}$ and normal B$_{\mathrm{t}}$. Regulation of q$_{\mathrm{0}}$ and $\beta_{\mathrm{N}}$ can prevent magneto-hydrodynamic instabilities that deteriorate plasma performance in discharges with a monotonically increasing safety-factor profile. Zero-input-torque scenarios are of special interest because future burning plasma tokamaks such as ITER will most likely operate with very low input torque, which makes these scenarios more susceptible to locked modes. To support studies of such scenarios, a controller for simultaneous regulation of q$_{\mathrm{0}}$ and $\beta_{\mathrm{N}}$ has been developed using near-zero net input torque actuators including balanced neutral beam injection (NBI) and electron-cyclotron heating {\&} current drive (ECH/ECCD). Experimental results show that in spite of the presence of locked modes the use of feedback control resulted in good tracking of the commanded q$_{\mathrm{0}}$ and $\beta_{\mathrm{N}}$ when both ECCD/ECH and NBI were available.   

Predictive Rotation Profile Control for the DIII-D Tokamak

Control-oriented modeling and model-based control of the rotation profile are employed to build a suitable control capability for aiding rotation-related physics studies at DIII-D. To obtain a control-oriented model, a simplified version of the momentum balance equation is combined with empirical representations of the momentum sources. The control approach is rooted in a Model Predictive Control (MPC) framework to regulate the rotation profile while satisfying constraints associated with the desired plasma stored energy and/or $\beta_{\mathrm{N\thinspace }}$limit. Simple modifications allow for alternative control objectives, such as maximizing the plasma rotation while maintaining a specified input torque. Because the MPC approach can explicitly incorporate various types of constraints, this approach is well suited to a variety of control objectives, and therefore serves as a valuable tool for experimental physics studies. Closed-loop TRANSP simulations are presented to demonstrate the effectiveness of the control approach.   

Effects of Plasma Shaping on Intrinsic Rotation in DIII-D

Intrinsic rotation in an axisymmetric tokamak must have its source in a momentum flux that passes through the boundary of the plasma. Since it is well-known that shaping in diverted discharges modifies the pedestal in H-mode discharges [1], we have performed experiments on DIII-D in which the shaping is changed during a discharge and the accompanying change in the intrinsic rotation profile is measured. We see that the change in intrinsic rotation magnitude in the outer plasma region, rho $=$ 0.7, is correlated with the plasma stored energy to a large extent. At the next level of response, there are changes in the rotation profiles related to the pedestal pressure, and in the interior possibly to the q profile. An additional focus of these experiments was to make measurements of the shape-induced changes in turbulence strength and spectra with Beam Emission Spectroscopy and Doppler Back Scattering. Both show clear changes in frequency with, for example, a change in the major radius of the X-point in a single null diverted plasma. [1] R.J. Groebner et al, Nucl. Fusion \textbf{49}, 085037 (2009).   

Scaling of Intrinsic Rotation with Normalized Gyroradius in DIII-D and Comparison to Intrinsic Torque Scaling

New experiments at DIII-D have investigated the scaling of intrinsic rotation with the normalized gyroradius, $\rho $*, by performing a dimensionless parameter scan in electron cyclotron heated H-mode plasmas with no external torque injection. Intrinsic rotation was measured for both the dominant impurity and the main-ion species. The main experimental result is that the Mach no. (toroidal velocity normalized to either the sound speed or the Alfven velocity) was nearly constant or slightly increasing with decreasing $\rho $*. These intrinsic rotation results corroborate the previous measurements$^{\mathrm{1}}$ of the intrinsic torque and momentum confinement time scaling with $\rho $*, which indicates that the fast-ion content from significant neutral beam injection in the previous experiment did not influence those results. The potential effect of neutral particle transport in the pedestal is also investigated. Predictions of the intrinsic rotation in ITER are reviewed.   

Reynolds Stress-Driven Edge Momentum Transport in DIII-D

Tokamak plasma rotate toroidally due to an intrinsic edge source [1]. Reynolds Stress has been proposed $\langle$nv$_\phi$v$_\rho$$\rangle$=$\langle$n$\rangle$$\langle$\tilde{v}$_\phi$ \tilde{v}$_\rho$$\rangle$+$\hspace{0mm}$$\langle$v$_\phi$$\rangle$$\langle$\~n$_\phi$\tilde{v}$_\rho$$\rangle$+$\hspace{0mm}$$\langle$\tilde{n}$\tilde{v}$_\phi$\tilde{v}$_\rho$$\rangle$ as the transport mechanism. The term $\langle$n$\rangle$$\langle$\tilde{v}$_\phi$\tilde{v}$_\rho$$\rangle$ peaking $\sim$-1e26m$^\wedge$-1s$^\wedge$2 just inside the separatrix, causes a significant inward pinch due to cross-phase effects while the outward convection term, $\langle$v$_\phi$$\rangle$$\langle$$\tilde{n}$\tilde{v}$_\rho$$\rangle$, peaking at $\sim$-1E26m$^\wedge$-1s$^\wedge$2 roughly balances it. Surprisingly, the triple correlation term, $\langle$\tilde{n}$\tilde{v}$_\rho$\tilde{v}$_\phi$$\rangle$ peaking at t$\sim$-1E25m$^\wedge$-1s$^\wedge$2 becomes important as other terms almost null out. A rough momentum balance finds that the momentum flux from the RS term can explain the observed momentum balance.   

Collisionality and temperature dependence of the edge main-ion co-current rotation profile feature on DIII-D

A new edge main-ion (D$^{\mathrm{+}})$ CER system and upgraded edge impurity system are revealing clear differences between the main-ion and dominant impurity (C$^{\mathrm{6+}})$ toroidal rotation from the pedestal top to the scrape off layer on DIII-D with implications for intrinsic rotation studies. A peaked co-current edge toroidal rotation is observed for the main ion species near the outboard midplane separatrix with values up to 140km/s for low collisionality QH modes. In lower power (P$_{\mathrm{NBI}}=$0.8MW) H-modes the edge rotation is still present but reduced to \textasciitilde 50km/s. D$^{\mathrm{+}}$ and C$^{\mathrm{6+}}$ toroidal rotation differences are presented for a variety of scenarios covering a significant range of edge collisionality and T$_{\mathrm{i}}$. Observations are compared with predictions from several models including collisionless ion orbit loss calculations and more complete modeling using the XGC0 code, which also predicts 140km/s edge rotation for low collisionality QH mode cases.   

Isotope Mass Scaling of Turbulence and Transport*

The dependence of turbulence characteristics and transport scaling on the fuel ion mass has been investigated in a set of hydrogen $(A=1)$ and deuterium $(A=2)$ plasmas on DIII-D. Normalized energy confinement time $(B*\tau_{E})$ is two times lower in hydrogen (H) plasmas compare to similar deuterium (D) plasmas. Dimensionless parameters other than ion mass $(A)$, including $\rho^{*}$, $q_{95}$, $T_{e}/T_{i}$, $\beta_{N}$, $\nu^{*}$, and Mach number were maintained nearly fixed. Matched profiles of electron density, electron and ion temperature, and toroidal rotation were well matched. The normalized turbulence amplitude $(\tilde{n}/n)$ is approximately twice as large in H as in D, which may partially explain the increased transport and reduced energy confinement time. Radial correlation lengths of low-wavenumber density turbulence in hydrogen are similar to or slightly larger than correlation lengths in the deuterium plasmas and generally scale with the ion gyroradius, which were maintained nearly fixed in this dimensionless scan. Predicting energy confinement in D-T burning plasmas requires an understanding of the large and beneficial isotope scaling of transport.   

Impact of Magnetic Island - Turbulence Multi-Scale Interaction on Plasma Confinement and Magnetic Island Stability

Recent measurements$^{1}$ and gyrokinetic simulations$^{2}$ reported the reduction of turbulent density fluctuations (ñ) inside magnetic islands, and ñ increase outside magnetic islands, when the island width (W) exceeds a threshold (W$_{T}$). As the cross-field transport is dominantly driven by ñ, this calls into question the conventional understanding of confinement ($\tau$$_{e}$) degradation and Neoclassical Tearing Mode (NTM) stability physics. We report that the increase in ion-scale ñ outside the island correlates with higher heat and particle fluxes, i.e., ñ increases temporarily when $\tau$$_{e}$ is decreasing, while in the following stationary state ñ is comparable to before NTM onset. This indicates that the decrease of the plasma stored energy results from ñ-NTM interaction. On the other hand, simultaneous ñ reduction at the O-point has a destabilizing effect on NTMs. These observations suggest that driving ñ at the O-point could prevent small islands from growing large, allowing better plasma confinement and safer tokamak operation. [1] Bardóczi, L. et al. Phys. Plasmas 24, 056106 (2017) [2] Navarro, A. B. et al. Plasma Phys. Control. Fusion 59, 034004 (2017)   

Particle transport in DIII-D plasmas

By analyzing the plasma opacity and density evolution during the ELM cycle in DIII-D H-mode plasmas in which the amount of gas fueling was altered, we find evidence for an inward particle pinch at the plasma edge which seems to become more pronounced at higher density. Furthermore, at the plasma edge we find a correlation between the pedestal density and opacity, which measures neutral penetration depth. The changes in edge opacity during an ELM cycle were calculated by using a detailed time history of measured plasma profiles. At the same time, the density evolution during an ELM cycle was investigated. We find that if the edge density increases through an increase in gas fueling, then opacity increases and neutral fueling penetration depth decreases. We also find that density at the top of the pedestal recovers faster following an ELM when the overall density level is higher, leading to a hollow profile inside of the pedestal top. All these results indicate that there must be an inward particle pinch in the pedestal which will be crucial in the fueling of future burning plasma devices.   

Microtearing instabilities and resulting electron thermal transport in DIII-D discharges

A reduced transport model for microtearing modes (MTMs), has been developed for use in integrated predictive modeling studies. A unified fluid/kinetic approach is employed in the derivation of the nonlinear MTM dispersion relation. The dependence of the MTMs real frequency and growth rate in DIII-D like L-mode and H-mode plasma discharges is examined for a range of plasma parameters. The saturated amplitude of the magnetic fluctuations is calculated utilizing numerically determined MTM eigenvalues in the nonlinear MTM envelope equation. It is found that the electron temperature gradient in the presence of moderate collision frequency is required for MTMs to become unstable. The effects of small and large collisionality and small and large wavenumbers on MTMs are found to be stabilizing, while the effects of density gradient, plasma beta, low current density, and large magnetic shear are found to be destabilizing. The MTM growth rate, magnetic fluctuation strength, as well as electron thermal diffusivity is found to be larger in the H-mode plasma than in the L-mode plasma.   

Extracting 3D Information from 1D and 2D Diagnostic Systems on the DIII-D Tokamak

The interpretation of tokamak data often hinges on assumptions of axisymetry and flux surface equilibria, neglecting 3D effects. This work discusses examples on the DIII-D tokamak where this assumption is an insufficient approximation, and explores the diagnostic information available to resolve 3D effects while preserving 1D profiles. Methods for extracting 3D data from the electron cyclotron emission radiometers, density profile reflectometer, and Thomson scattering system are discussed. Coordinating diagnostics around the tokamak shows the significance of 3D features, such as sawteeth[1] and resonant magnetic perturbations. A consequence of imposed 3D perturbations is a shift in major radius of measured profiles between diagnostics at different toroidal locations. Integrating different diagnostics requires a database containing information about their toroidal, poloidal, and radial locations. Through the data analysis framework OMFIT, it is possible to measure the magnitude of the apparent shifts from 3D effects and enforce consistency between diagnostics. Using the existing 1D and 2D diagnostic systems on DIII-D, this process allows the effects of the 3D perturbations on 1D profiles to be addressed.   

Measurement of internal equilibrium and fluctuating magnetic field on DIII-D by using Radial Polarimeter-Interferometer

A Faraday-effect based Radial Interferometer-Polarimeter (RIP) has been implemented on DIII-D for measurements of the magnetic equilibrium and the magnetic axis vertical position with fast time response that can be exploited for position and instability control. Furthermore, RIP can also measure internal radial magnetic fluctuations, e.g. internal kink - precursor to sawtooth crash, NTM seeding and plasma disruptions. By utilizing counter-rotating circular polarization technique, the diagnostic measures Faraday rotation and line-integrated electron density simultaneously with time response at microsecond scale and phase noise \textasciitilde 0.01 degree/sqrt(kHz), corresponding to 1 Gs/sqrt(kHz) for typical DIII-D plasma conditions. Measurement errors has been assessed and minimized. Systematic comparison between RIP measurement and MSE-constrained EFIT using a synthetic diagnostic shows good agreement, manifesting consistency of internal magnetic field measurement between RIP and other magnetic diagnostics. \textit{*Supported by USDOE grants DE-FG03-01ER54615 and DE-FC02-04ER54698.}   

Synthetic Diagnostic for Doppler Backscattering (DBS) Turbulence Measurements based on Full Wave Simulations

Plasma full-wave simulations of the DIII-D DBS system including its lenses and mirrors are developed using the GPU-based FDTD2D code [1], verified against the GENRAY ray-tracing code and TORBEAM paraxial beam code. Our semi-analytic description of the effective spot size for a synthetic diagnostic reveals new focusing and defocusing effects arising from the combined effects of the curvature of the reflecting surface and that of the Gaussian beam wavefront. We compute the DBS transfer function from full-wave simulations to verify these effects. Using the synthetic diagnostic, nonlinear GYRO simulations closely match DBS fluctuation spectra with and without strong electron heating, without adjustment or change in normalization, while both GYRO and GENE also match fluxes in all transport channels [2]. Density gradient driven TEMs that are observed by the DBS diagnostic on DIII-D are reproduced by simulations as a band of discrete toroidal mode numbers which intensify during strong electron heating.\newline [1] B. C. Rose, S. Kubota, and W. A. Peebles, in Proc. 18th Topical HTPD Conference (Wildwood, N.J., 2010). Also J. Hillesheim et al., Rev. Sci. Instrum. 83, 10E331 (2012).\newline [2] D. R. Ernst et al., Phys. Plasmas 23, 056112 (2016). Also 2014 IAEA FEC paper CN-221/EX/2-3.   

DIII-D Neutron Measurement: Status and Plan for Simplification and Upgrade

Neutron diagnostics play key essential roles on DIII-D. Historically an 18-channel 2.45MeV D-D neutron measurement system based on $^{\mathrm{3}}$He and BF$_{\mathrm{3}}$ proportional counters was inherited from Doublet-III including associated electronics and CAMAC data acquisition. Three fission chambers and two neutron scintillators were added in the 1980s and middle 1990s respectively. For Tritium burn-up studies, two 14MeV D-T neutron measurement systems were installed in 2009 and 2010. Operation and maintenance experience have led to a plan to simplify and upgrade these aging systems to provide a more economical and reliable solution for future DIII-D experiments. On simplification, most conventional expensive NIM and CAMAC modules will be removed. Advanced technologies like ultra-fast data acquisition and software-based pulse identification have been successfully tested. Significant data reduction and efficiency improvement will be achieved by real-time digital pulse identification with a field-programmable gate array. The partly renewed system will consist of 4 neutron counters for absolute calibration and 4 relatively calibrated neutron scintillators covering a wide measurement range.   

Observations of highly sheared turbulence in the H-mode pedestal using Phase Contrast Imaging on DIII-D

Highly sheared turbulence with short radial correlation lengths has been measured near the top of the H-mode pedestal, in addition to the previously measured highly-sheared turbulence measured in the $E_r$ well. Turbulence in regions of large velocity shear is characterized by radial correlation lengths shorter than the poloidal wavelength ($L < \lambda\sim 2$ cm) and large Doppler-shifted frequencies ($f > 200$ kHz). The phase contrast imaging (PCI) diagnostic on DIII-D is ideally suited to measuring this density turbulence due to the measurement geometry and high frequency bandwidth. Radial localization is achieved by optical filtering, varying the ExB profile, and shifting the plasma position. Reconfiguration of the $E_r$ well, such as at the L-H transition or the transition to wide pedestal QH-mode, shows a near-instantaneous change ($t < 1$ ms) to the sheared turbulence in the $E_r$ well ($\sim 1$ cm inside the separatrix). In contrast, the sheared turbulence near the top of the pedestal ($\sim 2$ cm inside the separatrix) varies over times scales of tens of ms, consistent with pedestal evolution.   

Increase in turbulent transport at DIII-D pedestal top due to RMP-induced reduction of electric field shear

It has been demonstrated that resonant magnetic perturbations (RMPs), applied with the right conditions, suppress or mitigate edge localized modes (ELMs) in DIII-D at low, ITER-like, collisionality. Along with the RMP ELM suppression, observations in DIII-D, via BES, DBS, and other fluctuation diagnostics, show an increase in electrostatic turbulence near the top of the pedestal, where the mean radial electric field (Er) shearing rate decreases. The Gyrokinetic Toroidal Code (GTC) simulations show that there is a strong correlation between the reduction of the Er shearing rate and an increase in turbulence and transport near the pedestal top of the DIII-D shot 158103 during RMP ELM suppression at 03050 ms. A nonlinear outward spreading of the turbulence is observed, which allows a stronger microturbulence on the pedestal top during ELM suppression by RMP. For comparison, the turbulence and transport near the pedestal top remain at low levels when the plasma is ELMing, i.e. when the Er shearing rate is not decreased in the same shot at 03750 ms (ELMing w/ RMP on) and in another shot 158104 at 1350 ms (ELMing without RMP). Furthermore, GTC simulations show that the kink responses to the 3D RMP has little effects on the growth rate of electromagnetic kinetic-ballooning mode and on the turbulent transport and zonal flow damping in electrostatic turbulence$^{\mathrm{1}}$.   

Installation and Testing of ITER Integrated Modeling and Analysis Suite (IMAS) on DIII-D

A critical objective of the ITER Integrated Modeling Program is the development of IMAS to support ITER plasma operation and research activities. An IMAS framework has been established based on the earlier work carried out within the EU. It consists of a physics data model and a workflow engine. The data model is capable of representing both simulation and experimental data and is applicable to ITER and other devices. IMAS has been successfully installed on a local DIII-D server using a flexible installer capable of managing the core data access tools (Access Layer and Data Dictionary) and optionally the Kepler workflow engine and coupling tools. A general adaptor for OMFIT (a workflow engine) is being built for adaptation of any analysis code to IMAS using a new IMAS universal access layer (UAL) interface developed from an existing OMFIT EU Integrated Tokamak Modeling UAL. Ongoing work includes development of a general adaptor for EFIT and TGLF based on this new UAL that can be readily extended for other physics codes within OMFIT.   

Neural Network Based Models for Fusion Applications*

Whole device modeling, engineering design, experimental planning and control applications demand models that are simultaneously physically accurate and fast. This poster reports on the ongoing effort towards the development and validation of a series of models that leverage neural-­network (NN) multidimensional regression techniques to accelerate some of the most mission critical first principle models for the fusion community, such as: the EPED workflow for prediction of the H­-Mode and Super H­-Mode pedestal structure the TGLF and NEO models for the prediction of the turbulent and neoclassical particle, energy and momentum fluxes; and the NEO model for the drift­-kinetic solution of the bootstrap current. We also applied NNs on DIII-D experimental data for disruption prediction and quantifying the effect of RMPs on the pedestal and ELMs. All of these projects were supported by the infrastructure provided by the OMFIT integrated modeling framework.   

Coupled core-pedestal simulations with self-consistent transport of impurities

This poster reports on the ongoing development of a production tool that robustly predicts density, temperature and rotation profiles from on-axis to the separatrix. The number of free parameters and assumptions in the simulations are reduced by using physics based models that are self-consistently coupled to one another. Previous coupled core-pedestal simulations were shown to be able to reproduce experimental profiles [Meneghini PoP 2016, NF 2017], but relied on prior knowledge of the plasma Z$_{\text{eff}}$ and pedestal rotation boundary condition. Z$_{\text{eff}}$ is an important parameter since it influences both core performance, through transport and line radiation, and pedestal stability via its effect on the bootstrap current. To self-consistently account for the effects of impurities the previously mentioned core-pedestal workflow is iteratively coupled to the 1D impurity transport code STRAHL [Dux 2006] within the OMFIT framework. In this scheme NEO and TGLF will provide the transport fluxes used to calculate the diffusion and pinch profiles used in STRAHL. The new workflow also implements a boundary condition for the plasma rotation based on first principles [Stoltzfus-Dueck PRL 2015].   

Uncertainty Propagation in OMFIT

A rigorous comparison of power balance fluxes and turbulent model fluxes requires the propagation of uncertainties in the kinetic profiles and their derivatives. Making extensive use of the python uncertainties package, the OMFIT framework has been used to propagate covariant uncertainties to provide an uncertainty in the power balance calculation from the ONETWO code, as well as through the turbulent fluxes calculated by the TGLF code. The covariant uncertainties arise from fitting 1D (constant on flux surface) density and temperature profiles and associated random errors with parameterized functions such as a modified tanh. The power balance and model fluxes can then be compared with quantification of the uncertainties. No effort is made at propagating systematic errors. A case study will be shown for the effects of resonant magnetic perturbations on the kinetic profiles and fluxes at the top of the pedestal. A separate attempt at modeling the random errors with Monte Carlo sampling will be compared to the method of propagating the fitting function parameter covariant uncertainties.   

Turbulence and sheared flow dynamics during q95 and density scans across the L-H transition on DIII-D

Measurements of long wavelength density fluctuation characteristics have been obtained in the edge of Deuterium (D) plasmas across the L-H transition on DIII-D during density and q95 scans. The relative density fluctuation amplitude measured by Beam Emission Spectroscopy (BES) increases with higher q95. The power threshold is found to increase with plasma current (i.e., lower q95) but with complex density dependence: the largest increase of P$_{\mathrm{LH}}$ is seen at n$_{\mathrm{e}}$\textasciitilde 3.2e19 m$^{\mathrm{-3}}$. Interestingly, a dual counter-propagating mode is observed for cases when P$_{\mathrm{LH}}$ is low. The existence of the dual mode is correlated with increasing flow shear. Estimation of the turbulence kinetic energy transfer from turbulence to the flow increases prior to the transition. The complex behaviors of the turbulence characteristics and dual frequency modes interactions impact the flow shear generation, the transition process and the power threshold scaling.   

L-H Transition Dynamics in ITER-Similar D, He, and H Plasmas

Recent work at DIII-D has revealed important differences in L-H transition trigger dynamics between deuterium (D), helium (He) and hydrogen (H) plasmas. The ion flux/polarization current induced by the Reynolds stress is shown to be decisive for the fast time evolution of the edge electric field across the L-H transition at intermediate and low plasma density in the plateau collisionality regime, in D and He plasmas. As the corresponding $j\times B_{\mathrm{\thinspace }}$torque increases, concomitant turbulence suppression occurs within 100-200 $\mu $s of the peak Reynolds stress gradient. H-plasmas show lower Reynolds stress and \begin{figure}[htbp] \centerline{\includegraphics[width=0.38in,height=0.22in]{130720171.eps}} \label{fig1} \end{figure} torque, and reduced toroidal correlation, of the self-organized \begin{figure}[htbp] \centerline{\includegraphics[width=0.42in,height=0.17in]{130720172.eps}} \label{fig2} \end{figure} edge flow layer, and longer transition times concomitantly with substantially higher required L-H transition threshold power. In H-plasmas, the Reynolds force is comparable in magnitude to the neoclassical bulk ion viscosity and the force due to thermal ion orbit loss, potentially explaining the increased power threshold. Supported by the U.S. DOE under DE-FC02-04ER54698, DE-FG02-08ER54984, DE-AC02-09CH11466 and DE-FG02-08ER54999.   

Influence of 3D magnetic fields on turbulence at the L-H transition and across magnetic islands in DIII-D

Measurements of long-wavelength density fluctuations using beam emission spectroscopy (BES) reveal the significant impact of 3D magnetic fields on turbulence amplitudes and characteristics. Fluctuations are measured across an applied static m,n $=$ 2,1 magnetic island that is rotated toroidally through the BES sightlines, providing quasi-3D measurement capability. A single unstable broadband turbulence mode is observed near the O-point, but near the X-point, this mode is accompanied by a second mode propagating in the opposite direction; fluctuation amplitudes are also much higher near the X-point than the O-point. 2D fluctuations are also measured in the L-mode edge leading up to L-H transitions with applied resonant and non-resonant magnetic perturbations. Normalized fluctuation amplitudes are \textasciitilde 4 times larger with resonant fields than with non-resonant fields. Additionally, dual counter-propagating modes are observed with resonant fields, while only a single mode is observed with non-resonant fields. These measurements may reveal how magnetic perturbations raise the L-H power threshold by altering turbulence-flow dynamics leading up to the transition.   

Phase-Space Dependence of Fast-Ion Transport by Neoclassical Tearing Modes

The fast-ion transport caused by neoclassical tearing modes (NTM) in H-mode plasmas is investigated in different parts of fast-ion phase space using the newly developed beam modulation technique and a variety of fast-ion diagnostics that are sensitive to different parts of the distribution function. As measured by electron cyclotron emission, the $(m,n)=(2,1)$ tearing modes have an island width of $\sim10$~cm and change phase $180^\circ$ at the $q=2$ surface. (Here, $m$ is the poloidal mode number and $n$ is the toroidal mode number.) The fast ions are produced by deuterium neutral beam injection at 75-81~keV. To measure fast-ion transport in different parts of phase space, one neutral-beam source is modulated at 20~Hz. Flows in phase space are obtained through comparisons of measured neutron, solid-state neutral particle analyzer, and fast-ion D-alpha signals with the expected signals in the absence of wave-induced transport. In order to populate different parts of phase space, beams with six different injection geometries are modulated on successive discharges. Initial analysis indicates that the largest transport occurs for on-axis, tangentially-injected ions, while smaller transport occurs for off-axis or perpendicular injection. Simulations show similar trends.   

Orbit Tomography: A Method for Determining the Population of Individual Fast-ion Orbits from Experimental Measurements

Due to the complicated nature of the fast-ion distribution function, diagnostic velocity-space weight functions are used to analyze experimental data. In a technique known as Velocity-space Tomography (VST), velocity-space weight functions are combined with experimental measurements to create a system of linear equations that can be solved. However, VST (which by definition ignores spatial dependencies) is restricted, both by the accuracy of its forward model and also by the availability of spatially overlapping diagnostics. In this work we extend velocity-space weight functions to a full 6D generalized coordinate system and then show how to reduce them to a 3D orbit-space without loss of generality using an action-angle formulation. Furthermore, we show how diagnostic orbit-weight functions can be used to infer the full fast-ion distribution function, i.e. Orbit Tomography. Examples of orbit weights functions for different diagnostics and reconstructions of fast-ion distributions are shown for DIII-D experiments. This work was supported by the U.S. Department of Energy under DE-AC02-09CH11466 and DE-FC02-04ER54698.   

Using Electron Cyclotron Emission Images to localize the drive and damping of Alfv\'en eigenmodes

Alfv\'en Eigenmodes (AE) are routinely imaged in DIII-D with the Electron Cyclotron Emission Imaging system (ECE-I). From the ECE-I images it was found that the AE wave fronts show a clear radial phase variation, which reflects a radial plasma displacement that is induced by the AEs. We use the measured plasma displacement to extract the fluctuating electric and magnetic fields and use these fields to calculate the Poynting flux to determine radial wave-induced energy flows for saturated AEs. These energy flows arise when the drive of the mode does not coincide with the location of the damping of the mode. We will use the measured curved AE wave fronts to determine the radial energy flow that is induced by the AEs and show that the location of fast-ion drive of the AEs does not coincide with the location of the strongest damping of the mode.   

Alfven Eigenmode Control in DIII-D.

Alfven eigenmodes (AE) driven by fast ions from neutral beam and ion cyclotron heating are common in present day tokamak plasmas and are expected to be destabilized by alpha particles in future burning plasma experiments. Because these waves have been shown to cause loss and redistribution of fast ions which can impact plasma performance and potentially device integrity, developing control techniques for AEs is of paramount importance. In the DIII-D plasma control system, spectral analysis of real-time ECE data is used as a monitor of AE amplitude, frequency, and location. These values are then used for feedback control of the neutral beam power to control Alfven waves and reduce fast ion loss. This work describes tests of AE control experiments in the current ramp up phase, during which multiple Alfven eigenmodes are typically unstable and fast ion confinement is degraded significantly. Comparisons of neutron emission and confined fast ion profiles with and without active AE control will be made.   

Optimization of DIII-D discharges to avoid AE destabilization

The aim of the study is to analyze the stability of Alfven Eigenmodes (AE) perturbed by energetic particles (EP) during DIII-D operation. We identify the optimal NBI operational regimes that avoid or minimize the negative effects of AE on the device performance. We use the reduced MHD equations to describe the linear evolution of the poloidal flux and the toroidal component of the vorticity in a full 3D system, coupled with equations of density and parallel velocity moments for the energetic particles, including the effect of the acoustic modes. We add the Landau damping and resonant destabilization effects using a closure relation. We perform parametric studies of the MHD and AE stability, taking into account the experimental profiles of the thermal plasma and EP, also using a range of values of the energetic particles $\beta$, density and velocity as well the effect of the toroidal couplings. We reproduce the AE activity observed in high poloidal $\beta$ discharge at the pedestal [J. Huang, 58th APS DP/2016] and reverse shear discharges [W. W. Heidbrink, Nucl. Fusion, 53, 093006, (2013)].   

Experimental investigation of stability, frequency and toroidal mode number of compressional Alfv\'{e}n eigenmodes in DIII-D

An experimental investigation of the stability of Doppler-shifted cyclotron resonant compressional Alfv\'{e}n eigenmodes (CAE) using the flexible DIII-D neutral beams has begun to validate a theoretical understanding and realize the CAE's diagnostic potential. CAEs are excited by energetic ions from neutral beams [Heidbrink, NF 2006], with frequencies and toroidal mode numbers sensitive to the fast-ion phase space distribution, making them a potentially powerful passive diagnostic. The experiment also contributes to a predictive capability for spherical tokamak temperature profiles, where CAEs may play a role in energy transport [Crocker, NF 2013]. CAE activity was observed using the recently developed Ion Cyclotron Emission diagnostic---high bandwidth edge magnetic sensors sampled at 200 MS/s. Preliminary results show CAEs become unstable in BT ramp discharges below a critical threshold in the range 1.7 -- 1.9 T, with the exact value increasing as density increases. The experiment will be used to validate simulations from relevant codes such as the Hybrid MHD code [Belova, PRL 2015].   

Characteristics of Ion Cyclotron Emission on the DIII-D Tokamak

Understanding the relationship between Ion Cyclotron Emission (ICE) and the energetic particle distribution is important in modern-day tokamaks, since passive measurements of ICE in a reactor environment, such as ITER, could provide information on the alpha particle population and fast-ion losses. ICE is readily excited in DIII-D plasmas by kinetic instabilities resulting from neutral beam injection across a wide operational space, including in both helium and deuterium plasmas. A large set of ICE measurements has been collected over the past two years with multiple receiving antennas digitized at 200 Msamples/sec. The fundamental ICE frequency observed in DIII-D plasmas is in the 5-20 MHz range with typical toroidal magnetic fields of 1--2 T; harmonics are observed up to the Nyquist limit at 100 MHz. These frequencies correspond to both core and edge locations; however, ICE is more often observed at frequencies correlated with ion cyclotron harmonics at the outboard edge. ICE dependencies on plasma and beam parameters such as beam geometry, injection voltage, beam power, plasma density, toroidal field, neutron rate, and ion species are presented. Rapid changes of ICE during ELMs and sawteeth may provide insight into the fast evolution of the beam ion distribution due to these instabilities.~Correlation of the ICE signals with the results of other fast-ion diagnostics is essential to compare with modelling of underlying kinetic instabilities.   

Transport of shattered pellet material during fast shutdown of DIII-D plasmas

A second shattered pellet injection (SPI) system has been installed on DIII-D, allowing new measurements of radial and toroidal transport of injected impurities during fast shutdown of the plasma. Toroidally separated injections from the two systems vary the impurity profiles and resulting radiative heat loads, relative to available disruption diagnostics. Infrared imaging and radiometry are used to compare heat loads near the injection port with those located toroidally away, in order to quantify radiation asymmetries. Radial transport mechanisms are studied by directing the SPI trajectory away from the magnetic axis in order to reduce the ballistic transport of solid pellet fragments to the core, which more closely matches the ITER injection geometry. The assimilation of this edge deposited SPI is compared to SPI directed towards the core, and also with massive gas injection whose assimilation is strongly dependent on thermal quench MHD mixing.   

Variation of Argon Impurity Assimilation with Runaway Electron Current in DIII-D

Measurements of the effect of runaway electron (RE) pressure upon argon impurity assimilation in DIII-D are reported. Intentionally created post-disruption RE beams are ramped to different plasma currents to vary the RE pressure, while impurity levels are varied by injecting argon gas (in addition to Ar already present from the small pellet used to create the disruption). Based on comparisons of current decay rates and hard x-ray, spectroscopic, interferometer, and Thomson scattering data, it is found that argon is not mixed uniformly through the plasma radially but appears to be preferentially moved out of the center of the plasma toward the walls, relative to the main species (deuterium). This exclusion appears to be stronger at higher plasma current, indicating that this force originates from the runaway electrons.   

Investigation of runaway electron dissipation in DIII-D using a gamma ray imager.

We report the findings of a novel gamma ray imager (GRI) to study runaway electron (RE) dissipation in the quiescent regime on the DIII-D tokamak. The GRI measures the bremsstrahlung emission by RE providing information on RE energy spectrum and distribution across a poloidal cross-section. It consists of a lead pinhole camera illuminating a matrix of BGO detectors placed in the DIII-D mid-plane. The number of detectors was recently doubled to provide better spatial resolution and additional detector shielding was implemented to reduce un-collimated gamma flux and increase single-to-noise ratio. Under varying loop voltage, toroidal magnetic field and plasma density, a non-monotonic RE distribution function has been revealed as a result of the interplay between electric field, synchrotron radiation and collisional damping. A fraction of the high-energy RE population grows forming a bump at the RE distribution function while synchrotron radiation decreases. A possible destabilizing effect of Parail-Pogutse instability on the RE population will be also discussed.   

Simulation of excitation of whistler waves and momentum diffusion of runaway electrons in DIII-D tokamak

In recent quiescent runaway electron experiments (QRE) experiments in DIII-D, whistler waves excited by runaway electrons have been observed and are found to be associated with the fluctuation of electron cyclotron emission (ECE) signals. To understand this connection and how the whistler instabilities affect the runaway electron distribution in momentum space, a self-consistent kinetic simulation of runaway electrons, including both the secondary generation and the quasilinear diffusion effects from the excited modes, is conducted. The results show that three different branches of waves can be excited. The low frequency whistler waves and the high frequency magnetized plasma waves are excited by runaway electrons in high energy and low energy regimes respectively, through anomalous Doppler resonance. Due to the close phase velocities of these two branches, the Landau damping of them happens at the same energy regime. These two branches of waves are not observed directly in experiments due to their high frequencies. In addition, we find a third branch of waves with wavevector almost oblique to the magnetic field direction, excited by the bump-on-tail distribution of the runaway electrons. These waves are in the 100-200 MHz frequency range, which agrees with the experimental observations. The cyclic behavior of excitation and damping of whistler waves associated with the fluctuations of ECE signals are also reproduced.   

Asymmetric SOL Current in Vertically Displaced Plasma

Experiments at the DIII-D tokamak demonstrate a non-monotonic relationship between measured scrape-off layer (SOL) currents and vertical displacement event (VDE) rates with SOL currents becoming largely n=1 dominant as plasma is displaced by the plasma control system (PCS) at faster rates. The DIII-D PCS is used to displace the magnetic axis $\sim$10x slower than the intrinsic growth time of similar instabilities in lower single-null plasmas. Low order (n$\leq$2) mode decomposition is done on toroidally spaced current monitors to attain measures of asymmetry in SOL current. Normalized to peak n=0 response, a 2-4x increase is seen in peak n=1 response in plasmas displaced by the PCS versus previous VDE instabilities observed when vertical control is disabled. Previous inquiry shows VDE asymmetry characterized by SOL current fraction and geometric parameters of tokamak plasmas[1]. We note that, of plasmas displaced by the PCS, short displacement time scales near the limit of the PCS temporal control appear to result in larger n=1/n=2 asymmetries. [1] Fitzpatrick 2011 Nucl. Fus. 51 053007.   

Instability Prediction and Disruption Avoidance.

Disruption avoidance requires both a prediction of the instability proximity and an estimate of the 'disruptability' - the likelihood that the instability will ultimately result in a disruptive event. MHD spectroscopy is a promising option for obtaining information on the proximity of instabilities. Both the direct response and the antenna impedance provide valuable information on the low frequency global normal modes corresponding to stabilized kink modes. Data from DIII-D experiments and available nonlinear simulations are used to define quantitative criteria that signify when instabilities ultimately disrupt and when they saturate or dissipate. The key distinction in this approach is the use of physical characteristics of the modes rather than more accessible operation parameters. Simple characteristics of the linear instability for example include the linear growth or damping rate and the mode spatial extent. Criteria can also involve identifying sources of free energy in the nonlinear evolution.   

Real-time plasma response control for disruption avoidance

DIII-D experiments demonstrate the viability of using the plasma response to applied non-axisymmetric perturbations as a real-time control variable for disruption avoidance in low-torque ITER baseline demonstration discharges. The response to long-wavelength, low-frequency perturbations like the those used here correlates with plasma stability. Consequently, it is sensitive to key parameters that influence stability, increasing with $\beta_N$ and decreasing with plasma rotation. In the experiments, the plasma response amplitude was used to feedback modulate the neutral beam (NB) power, and thereby the plasma stored energy and stability. While the response was being controlled, the NB torque was slowly ramped down, resulting in a decrease in stored energy as the feedback acted to keep the response constant under the changing conditions. A case where the torque was ramped through zero to $-$0.5 Nm while maintaining stability was demonstrated, indicating that control can be maintained in the challenging ITER-like parameter regime.   

Disruption avoidance and fast ramp-down techniques for the DIII-D experimental scenarios

Plasma current ramp-down in ITER will continue in H-mode from 15 MA to 10 MA, and will keep a diverted shape until termination. This is in contrast to the limited ramp-down scenarios typically used in DIII-D operations. Additionally, fast emergency ramp-down scenarios for ITER and future reactors are a priority for disruption avoidance. New experiments in DIII-D use the ramp-down phase of a variety of experiments including in the ITER baseline scenario to survey and identify optimized ramp-down scenarios for both scheduled terminations and terminations triggered by off-normal event detection. Systematic scans in current ramp-rate (1-5 MA/s), neutral beam power (including $\beta_{\mathrm{N}}$ feedback) and ramp-down shaping (limited versus continued diverted) have identified fast ramp-down scenarios for Lower Single Null (LSN) and Double Null (DN) plasmas. Scenario-specific methods and their rates of successful termination will be presented and compared relative to a historical data-set of ramp-down programming in the limiter configuration. Locked modes are found to be the most significant challenge to disruption avoidance in diverted ramp-downs. Results for LSN diverted discharges that begin the rampdown with large locked-modes will also be presented. If available, results of similar experiments on EAST will be presented.   

Applications of “3D” Magnetic Diagnostics in DIII-D

Measurements of non-axisymmetric “3D” magnetic fields have been successfully employed in DIII-D to validate models of the plasma response to external magnetic perturbations, to validate predictions of the detailed spatial structure of unstable plasma modes, to measure damping rates of stable MHD modes, and to provide input for feedback control of resistive wall modes and of intrinsic error fields. Other possible applications that are ripe for development will be discussed, including the following. External magnetic data allows a direct measurement of the electromagnetic torque exchanged between the plasma and external coils, potentially an indicator of magnetic island onset. “3D” magnetic data in the absence of plasma may be used for direct measurement of error fields caused by coil asymmetries. Spatially resolved real-time measurements of non-axisymmetric fields can enable early detection of disruption precursors, independent of whether the instability is rotating or stationary.   

Optimization of 3D Field Design

Recent progress in 3D tokamak modeling is now leveraged to create a conceptual design of new external 3D field coils for the DIII-D tokamak. Using the IPEC dominant mode as a target spectrum, the Finding Optimized Coils Using Space-curves (FOCUS) code optimizes the currents and 3D geometry of multiple coils to maximize the total set's resonant coupling. The optimized coils are individually distorted in space, creating toroidal ``arrays'' containing a variety of shapes that often wrap around a significant poloidal extent of the machine. The generalized perturbed equilibrium code (GPEC) is used to determine optimally efficient spectra for driving total, core, and edge neoclassical toroidal viscosity (NTV) torque and these too provide targets for the optimization of 3D coil designs. These conceptual designs represent a fundamentally new approach to 3D coil design for tokamaks targeting desired plasma physics phenomena. Optimized coil sets based on plasma response theory will be relevant to designs for future reactors or on any active machine. External coils, in particular, must be optimized for reliable and efficient fusion reactor designs.   

Unlocking locked tearing-mode by applied rotating 3D field

Tokamak reactors require control of locked tearing modes. Pre-emptive applications of a rotating 3D field controlled with (M. Okabayashi: IAEA2016) or without (D. Shiraki: APS/ DPP13/PO4.15) feedback have demonstrated promising paths for recovering H-mode operation even in n$=$1 3D perturbed equilibria. Once a tearing mode becomes deeply locked with near-zero rotation across the radial profile, it is challenging to unlock before disruption. Preliminary observations suggest that the deeply locked state is a configuration with multiple instances of torque bifurcation and internal locking between multiple rational surfaces. Full rotation recovery was found in a narrow range of applied 3D field frequency or after one event of forced reconnection, reflecting the complex transient process of replacing the uncorrected error field with another 3D field. Initial comparison with a non-linear reduced MHD code (AEOLUS-IT) shows qualitative agreement. This work is supported in part by the US Department of Energy under DE-AC02-09CH11466, DE-FG02-99ER54531, DE-SC0003913, and DE-FC02-04ER54698.   

Poloidal structure of the plasma response to n=2 perturbations

A study of the plasma response to n=2 resonant magnetic perturbations (RMP) in DIII-D plasmas highlights the presence of two dominant modes, in good agreement with predictions from the MHD code MARS-Q. The use of RMPs offers potential benefits for nuclear fusion, for example ELM suppression or the correction of error fields, although their effect on the plasma needs to be better understood to predict how best to apply these fields. RMPs with n=2 and variable poloidal spectra are applied in plasma discharges with $q_{95}\sim~4.1$ and $\beta_N \sim 2.2$. Singular value decomposition (SVD) analysis was found to decouple the poloidal structure of the plasma response from the dependence on the spectrum of the perturbation applied. This analysis highlighted the presence of two modes, with the dominant mode peaking at the low-field-side midplane and the secondary one off-midplane, with indications of the latter being correlated with ELM suppression. The experimental observations are in good agreement with predictions of a dual-mode response from MARS-Q, improving prospects for projecting optimization of ELM control without triggering deleterious instabilities in future reactors.   

Demonstration of ECCD Stabilization of m/n$=$\textbf{2/1 NTMs in the Equivalent Low-Torque ITER Baseline Scenario in DIII-D}

Experiments in DIII-D are studying how best to minimize the average Electron Cyclotron Current Drive power directed at q$=$2 for stabilization of neoclassical tearing modes in discharges with the ITER shape and equivalent low-torque, low q95\textasciitilde 3.1 and low betaN\textasciitilde 1.8. ITER relies on localized ECCD to stabilize NTMs that would otherwise wall-lock and lead to disruption. The work contrasts the control strategies of pre-emption by continuous ECCD at the rational surface (``Active Tracking'') vs. suppression by a pulse of ECCD whenever a growing mode is detected (``Catch {\&} Subdue''). The large rho\textasciitilde 0.75 for q$=$2 and concomitant low Te make the EC current drive relatively weak per MW so that the EC power from 4\textasciitilde 5 well-aligned gyrotrons of 2.5\textasciitilde 2.8 MW, is just marginal for stabilization at about 70{\%} of the neutral beam injection power. The low-torque makes early mode detection and good initial alignment imperative for prompt suppression before wall-locking. Requirements for stabilization will be presented.   

Ideal kink and neoclassical tearing mode identification in DIII-D with ECE

Detection of neoclassical tearing modes (NTMs), which can degrade plasma confinement or cause disruptions, is important in tokamaks. We have developed a code to cross-correlate ECE/magnetics data to get the amplitude and phase profiles of the electron temperature (Te) oscillation caused by the rotating magnetic island and/or a kink. It has been observed that the $\Delta $Te amplitude on the two sides of the island center can be very different in some discharges. Also, a discrepancy often exists between the location of the rational q surface according to MSE-constrained EFIT and the location of island center according to ECE; this can be an issue for ECCD suppression of NTMs. We explore the possible causes of these two phenomena in terms of ECE location and calibration accuracy. By analyzing the Te fluctuation phase evolution after a large sawtooth crash which triggers an NTM, the presence of a kink-like mode before the onset of NTM can be discerned. Work supported by the US DOE under DE-FG02-97ER54415 and DE-FC02-04ER54698.   

Plasma stability analysis using Consistent Automatic Kinetic Equilibrium reconstruction (CAKE)

Presented here is the Consistent Automatic Kinetic Equilibrium (CAKE) code. CAKE is being developed to perform real-time kinetic equilibrium reconstruction, aiming to do a reconstruction in less than 100ms. This is achieved by taking, next to real-time Motional Stark Effect (MSE) and magnetics data, real-time Thomson Scattering (TS) and real-time Charge Exchange Recombination (CER, still in development) data in to account. Electron densities and temperature are determined by TS, while ion density and pressures are determined using CER. These form, together with the temperature and density of neutrals, the additional pressure constraints. Extra current constraints are imposed in the core by the MSE diagnostics. The pedestal current density is estimated using Sauters equation for the bootstrap current density. By comparing the behaviour of the ideal MHD perturbed potential energy ($\delta $W) and the linear stability index ($\Delta $') of CAKE to magnetics-only reconstruction, it can be seen that the use of diagnostics to reconstruct the pedestal have a large effect on stability.   

Detection of plasma stability on DIII-D, using the experimentally extracted plasma transfer function based on 3D MHD spectroscopy

Three-dimensional (3D) magnetohydrodynamic (MHD) spectroscopy is successfully applied to extract the plasma transfer function from DIII-D experiments. The method uses upper and lower internal coils to perform scans of frequency and poloidal mode spectrum, and measure the corresponding n$=$1 plasma response on 3D magnetic sensors. The transfer function is extracted, based on Pad\'{e} approximation, by fitting the measured signals on different sensors simultaneously. The experimental transfer function not only points out the multi-mode plasma response but also shows the number of dominant modes and the contribution of each mode to the plasma response. The extracted damping rate of the least stable mode can be a new index indicating plasma stability quantitatively. This method has the potential to optimize ELM suppression and monitor the plasma stability in future fusion reactors. Results and analysis of 3D MHD spectroscopy experiments will be presented.   

Development of the Pushered Single Shell Experimental Platform on NIF

The goal of the Pushered Single Shell (PSS) experimental campaign is to study mix of partially ionized ablator material into the hotspot. To do this we use a uniformly Si doped plastic capsule, the inner few microns of which can be doped with a few percent Ge. To diagnose mix, we use separated reactants [1]; deuterating the inner Ge-doped layer, CD/Ge, while putting Tritium into the Hydrogen capsule fill gas. Mix is then inferred by measuring the neutron yields from DD, DT, and TT reactions. In order to accentuate the cooling of the hot-spot due to Bremsstrahlung radiation when Ge is present, we required high hot-spot ion temperatures: \textasciitilde 3 keV. This, in turn, requires a fast, symmetric implosion. Using the Two-Shock campaign [2] as a starting point, we increased the capsule radius by \textasciitilde 25{\%} to 844 $\mu $m and the peak laser power by over 10{\%} to 475 TW. We also used a low, 0.3 mg/cc, He fill in the hohlraum to maintain control over implosion symmetry. This paper will describe the sequence of keyhole, 1DConA, 2DConA, and Symcap experiments we performed over the last year to tune the PSS implosions. We were successful in achieving our design goals; the PSS is the fastest CH capsule implosion in the laboratory, with peak velocity \textasciitilde 400 $\mu $m, a round hot-spot, with hotspot P2 $=$ 0 within errors, and a hot-spot ion temperature \textasciitilde 3.5 keV. This work was performed under the auspices of the Lawrence Livermore National Security, LLC, (LLNS) under Contract No. DE-AC52-07NA27344 [1] D.C. Wilson et al. Physics of Plasmas~\textbf{18}, 112707 (2011) [2] S.F. Khan et al. Physics of Plasmas \textbf{23}, 042708 (2016)   

Rugby and elliptical-shaped hohlraums experiments on the OMEGA laser facility.

We are pursuing on the OMEGA laser facility indirect drive implosions experiments in gas-filled rugby-shaped hohlraums in preparation for implosion plateforms on LMJ. The question of the precise wall shape of rugby hohlraum has been addressed as part of future megajoule-scale ignition designs [1]. Calculations show that elliptical-shaped holhraum is more efficient than spherical-shaped hohlraum. There is less wall hydrodynamics and less absorption for the inner cone, provided a better control of time-dependent symmetry swings. In this context, we have conducted a series of experiments on the OMEGA laser facility. The goal of these experiments was therefore to characterize energetics with a complete set of laser-plasma interaction measurements and capsule implosion in gas-filled elliptical-shaped hohlraum with comparison with spherical-shaped hohlraum. Experiments results are discussed and compared to FCI2 radiation hydrodynamics simulations. [1] S. Laffite, Physics of Plasmas 17, 102704 (2010).   

3D integrated HYDRA simulations of hohlraums including fill tubes*

Measurements of fill tube perturbations from hydro growth radiography (HGR) experiments on the National Ignition Facility show spoke perturbations in the ablator radiating from the base of the tube.$^{\mathrm{1}}$ These correspond to the shadow of the 10 $\mu $m diameter glass fill tube cast by hot spots at early time. We present 3D integrated HYDRA simulations of these experiments which include the fill tube. Meshing techniques are described which were employed to resolve the fill tube structure and associated perturbations in the simulations. We examine the extent to which the specific illumination geometry necessary to accommodate a backlighter in the HGR experiment contributes to the spoke pattern. Simulations presented include high resolution calculations run on the Trinity machine operated by the Alliance for Computing at Extreme Scale (ACES) partnership. 1. A. G. MacPhee, et al., Phys. Rev. E~\textbf{95}, 031204(R) (2017) *This work was performed under the auspices of the Lawrence Livermore National Security, LLC, (LLNS) under Contract No. DE-AC52-07NA27344   

Abstract Withdrawn

We transform the state-of-the art of plasma modeling by taking advantage of novel computational techniques for fast and robust integration of multiscale hybrid (full particle ions, fluid electrons, no displacement current) and full-PIC models. These models are implemented in 3D HYPERS [1] and axisymmetric full-PIC CONPIC [2] codes. HYPERS is a massively parallel, asynchronous code. The HYPERS solver does not step fields and particles synchronously in time but instead executes local variable updates (events) at their self-adaptive rates while preserving fundamental conservation laws. The charge-conserving CONPIC code has a matrix-free explicit finite-element (FE) solver based on a sparse-approximate inverse (SPAI) algorithm. This explicit solver approximates the inverse FE system matrix (``mass'' matrix) using successive sparsity pattern orders of the original matrix. It does not reduce the set of Maxwell's equations to a vector-wave (curl-curl) equation of second order but instead utilizes the standard coupled first-order Maxwell's system. We discuss the ability of our codes to accurately and efficiently account for multiscale physical phenomena in 3D magnetized space and laboratory plasmas and axisymmetric vacuum electronic devices. 1. Y.A. Omelchenko and H. Karimabadi, \textit{J. Comp. Phys}. 231,1766-1780, 2012. 2. D-Y. Na, Y.A. Omelchenko, H. Moon, B-H. Borges, F.L. Teixeira, \textit{J. Comp. Phys}. 346, 295-317, 2017.   

Numerical and experimental analysis of plasma generated on a solid target by a multi-MeV electron beam

We quantitatively investigate the interaction of a high energy, high flux electron beam with a solid target based on an approach that combines numerical and experimental characterization. The experimental study is carried out diagnosing the interaction of a multi-MeV electron beam, delivered by the CEA ASTERIX high-pulsed power driver, with an aluminum-tantalum target. The numerical analysis builds upon results from Particle-In-Cell simulations, to reproduce the electron beam dynamics and characteristics, and Monte-Carlo simulations, to simulate the interaction of the electron beam with the solid target. The main plasma features emerging from this analysis are analyzed using a 1D radiative transfer model which enabled the experimental spectra to be reproduced numerically with a good consistency. From the numerical integration of the 1D radiative transfer equation, plasma characteristics such as electron temperature and density profiles, as well as ion densities of constitutive species, are determined.   

On the use of compatible discretizations with the multi-fluid plasma model

In this presentation, we discuss the advantages and disadvantages of using compatible discretizations in continuous and discontinuous finite element methods for solving the multi-fluid plasma model. Maxwell’s equations, core components to the multi-fluid plasma model, are difficult to accurately represent due to the divergence involutions governed by Gauss’ laws. Many methods have been developed to deal with these ‘divergence errors’, notably the generalized Lagrange multiplier methods discussed in Munz et al 2000 and the vector basis discretization approach in Nedelec 1980. While the Lagrange multiplier cleaning schemes are effective and simple to implement, over long time scales the residual errors can have a large influence on the plasma. Compatible discretizations, such as those that represent the electric and magnetic fields on mixed HCurl ‘edge’ elements and HDiv ‘face’ elements, have been shown to be especially useful in the PIC community, however, their implementation can be complex. The goal of this research is to understand the benefits of using some of these divergence handling schemes and compare them in application to finite element methods with both implicit and explicit time integration.   

Efficient Implicit Plasma Simulation Using Quadrature Moment Inverson

Quadrature moment inversion algorithms [1-3] are one route to reducing the computational effort required for fully implicit PIC plasma simulation. These algorithms compute a sparse quadrature representation of the velocity distribution from a set of velocity moments. A Jacobian-free Newton Krylov (JFNK) solver can then be used to concurrently solve Maxwell’s equations and the quadrature node equations of motion implicitly differenced in time using the midpoint rule [4], retaining the fully kinetic character of the overall system. The results of several test problems will be presented along with an exploration of routes to achieving convergence of the complete set of PIC particles and field equations. [1] R.O. Fox, J. Comp. Phys. 227 (2008) [2] O. Desjardins, et. al., J. Comp. Phys. 227 (2008) [3] Yuan, C., and R. O. Fox, J. Comp. Phys. 230 (2011) [4] S. Markidis and G. Lapenta, J. Comp. Phys. 230.18 (2011)   

Reduction of collisional-radiative models for transient, atomic plasmas

Interactions between plasmas and any radiation field, whether by lasers or plasma emissions, introduce many computational challenges. One of these computational challenges involves resolving the atomic physics, which can influence other physical phenomena in the radiated system. In this work, a collisional-radiative (CR) model with reduction capabilities is developed to capture the atomic physics at a reduced computational cost. Although the model is made with any element in mind, the model is currently supplemented by LANL's argon database\footnote{Argon Atomic Data Sets. https://www-amdis.iaea.org/LANL/argon/}, which includes the relevant collisional and radiative processes for all of the ionic stages. Using the detailed data set as the true solution, reduction mechanisms in the form of Boltzmann grouping\footnote{Le et al. \textit{Phys. Plasmas} 20, 1-19 (2013).}, uniform grouping, and quasi-steady-state (QSS), are implemented to compare against the true solution. Effects on the transient plasma stemming from the grouping methods are compared. \\ \\ Distribution A: Approved for public release; unlimited distribution, PA (Public Affairs) Clearance Number 17449   

A fully implicit, conservative, hybrid kinetic-ion fluid-electron algorithm

The hybrid model with full-orbit kinetic ions and fluid electrons is a promising approach to describe a wide range of space and laboratory plasmas [e.g. 1]. Explicit hybrid algorithms typically use a predictor-corrector method with sub-cycling or a semi-implicit field solve to deal with the strict Whistler-wave CFL condition. However, these do not conserve momentum or energy, and are susceptible to numerical instability. While fully implicit methods have been recently explored [2,3] to step over such timescales in a stable manner, these studies have not considered conservation properties. Here, we present a novel particle-based non-linear hybrid algorithm that features discrete conservation of mass, momentum, energy and the solenoidal condition of the magnetic field. The scheme combines fully implicit time advance with orbit-averaging of the ion particles and the flexibility of conservative smoothing to reduce numerical noise. We verify the algorithm for a number of test problems and demonstrate the unique conservation properties. \\\\ 1. A. Stanier, et. al., {\it Phys. Plasmas} {\bf 24}, 022124 (2017).\\ 2. B. Sturdevant, et. al., {\it J. Comput. Phys.} {\bf 316}, 519 (2016). \\ 3. J. Cheng, et. al., {\it J. Comput. Phys.} {\bf 245}, 364 (2013).   

Metriplectic Gyrokinetics and Discretization Methods for the Landau Collision Integral

We present two important results for the kinetic theory and numerical simulation of warm plasmas: 1) We provide a metriplectic formulation of collisional electrostatic gyrokinetics that is fully consistent with the First and Second Laws of Thermodynamics. 2) We provide a metriplectic temporal and velocity-space discretization for the particle phase-space Landau collision integral that satisfies the conservation of energy, momentum, and particle densities to machine precision, as well as guarantees the existence of numerical H-theorem. The properties are demonstrated algebraically. These two result have important implications: 1) Numerical methods addressing the Vlasov-Maxwell-Landau system of equations, or its reduced gyrokinetic versions, should start from a metriplectic formulation to preserve the fundamental physical principles also at the discrete level. 2) The plasma physics community should search for a metriplectic reduction theory that would serve a similar purpose as the existing Lagrangian and Hamiltonian reduction theories do in gyrokinetics. The discovery of metriplectic formulation of collisional electrostatic gyrokinetics is strong evidence in favor of such theory and, if uncovered, the theory would be invaluable in constructing reduced plasma models.   

Variational principle for the parallel-symplectic representation of electromagnetic gyrokinetic theory

The nonlinear (full-$f$) electromagnetic gyrokinetic Vlasov-Maxwell equations are derived in the parallel-symplectic representation from an Eulerian gyrokinetic variational principle. The gyrokinetic Vlasov-Maxwell equations are shown to possess an exact energy conservation law, which is derived by Noether method from the gyrokinetic variational principle. Here, the gyrocenter Poisson bracket and the gyrocenter Jacobian contain contributions from the perturbed magnetic field. In the full-$f$ formulation of the gyrokinetic Vlasov-Maxwell theory presented here, the gyrocenter parallel-Amp\`{e}re equation contains a second-order contribution to the gyrocenter current density that is derived from the second-order gyrocenter ponderomotive Hamiltonian.   

A domain-decomposed multi-model plasma simulation of collisionless magnetic reconnection

Collisionless magnetic reconnection is a process relevant to many areas of plasma physics in which energy stored in magnetic fields within highly conductive plasmas is rapidly converted into kinetic and thermal energy. Both in natural phenomena such as solar flares and terrestrial aurora as well as in magnetic confinement fusion experiments, the reconnection process is observed on timescales much shorter than those predicted by a resistive MHD model. As a result, this topic is an active area of research in which plasma models with varying fidelity have been tested in order to understand the proper physics explaining the reconnection process. In this research, a hybrid multi-model simulation employing the Hall-MHD and two-fluid plasma models on a decomposed domain is used to study this problem. The simulation is set up using the WARPXM code developed at the University of Washington, which uses a discontinuous Galerkin Runge-Kutta finite element algorithm and implements boundary conditions between models in the domain to couple their variable sets. The goal of the current work is to determine the parameter regimes most appropriate for each model to maintain sufficient physical fidelity over the whole domain while minimizing computational expense.   

Abstract Withdrawn

Numerical studies of the validity of the guiding-center approximation are performed with the code {\sf ASCOT} [1] through a careful analysis of the full-orbit and guiding-center-orbit trajectories of nonrelativistic and relativistic charged particles moving in axisymmetric and nonaxisymmetric magnetic-field geometries. The validity of the guiding-center approximation is also investigated through the exact analytical solution for the motion of a charged particle in a straight nonuniform magnetic field with a constant field gradient [2]. \newline \noindent [1] E.~Hirvijoki, {\it et al.}, Comput.~Phys.~Commun.~{\bf 185}, 1310 (2014). \newline \noindent [2] A.~J.~Brizard, Phys.~Plasmas {\bf 24}, 042115 (2017).   

The UPSF code: a metaprogramming-based high-performance automatically parallelized plasma simulation framework

UPSF (Universal Plasma Simulation Framework) is a new plasma simulation code designed for maximum flexibility by using edge-cutting techniques supported by C++17 standard. Through use of metaprogramming technique, UPSF provides arbitrary dimensional data structures and methods to support various kinds of plasma simulation models, like, Vlasov, particle in cell (PIC), fluid, Fokker-Planck, and their variants and hybrid methods. Through C++ metaprogramming technique, a single code can be used to arbitrary dimensional systems with no loss of performance. UPSF can also automatically parallelize the distributed data structure and accelerate matrix and tensor operations by BLAS. A three-dimensional particle in cell code is developed based on UPSF. Two test cases, Landau damping and Weibel instability for electrostatic and electromagnetic situation respectively, are presented to show the validation and performance of the UPSF code.   

Exact collisional moments for plasma fluid theories

The velocity-space moments of the often troublesome nonlinear Landau collision operator are expressed exactly in terms of multi-index Hermite-polynomial moments of the distribution functions. The collisional moments are shown to be generated by derivatives of two well-known functions, namely the Rosenbluth-MacDonald-Judd-Trubnikov potentials for a Gaussian distribution. The resulting formula has a nonlinear dependency on the relative mean flow of the colliding species normalised to the root-mean-square of the corresponding thermal velocities, and a bilinear dependency on densities and higher-order velocity moments of the distribution functions, with no restriction on temperature, flow or mass ratio of the species. The result can be applied to both the classic transport theory of plasmas, that relies on the Chapman-Enskog method, as well as to deriving collisional fluid equations that follow Grad's moment approach. As an illustrative example, we provide the collisional ten-moment equations with exact conservation laws for momentum- and energy-transfer rate.   

Finding the crossover from phase mixing to collisions with an integral transform.

The one-dimensional linearized Vlasov-Poisson system can be exactly solved using the ``G transform'' [1], an integral transform based on the Hilbert transform. This transform removes the electric field, leaving a simple advection equation. We investigate how this integral transform interacts with the Fokker-Planck collision operator. The commutator of this collision operator with the G transform (the ``shielding term'') is shown to be negligible. We exactly solve the advection-diffusion equation without the shielding term. This solution determines when collisions dominate and when advection (i.e. Landau damping) dominates. Introducing an energy source term that balances the energy lost to dissipation allows for the creation of a steady state distribution function that transfers energy from larger to smaller velocity scales via Landau damping. Unlike the Kolmogorov cascade, this is an energy cascade that occurs completely in velocity space: there is no nonlinearity and the spatial dependence is trivial. We hope that the G transform will be used to simplify gyrokinetic codes or other kinetic models. [1] P. Morrison and D. Pfirsch, Phys Fluids B 4, 3038 (1992); P. Morrison Trans. Theo. Stat. Phys. 29, 397 (2000).   

Implementation of parallel moment equations in NIMROD

As collisionality is low (the Knudsen number is large) in many plasma applications, kinetic effects become important, particularly in parallel dynamics for magnetized plasmas. Fluid models can capture some kinetic effects when integral parallel closures are adopted. The adiabatic and linear approximations are used in solving general moment equations\footnote{J.-Y. Ji and E. D. Held, Phys. Plasmas 15, 102101 (2008).} to obtain the integral closures. In this work, we present an effort to incorporate non-adiabatic (time-dependent) and nonlinear effects into parallel closures. Instead of analytically solving the approximate moment system, we implement exact parallel moment equations in the NIMROD fluid code. The moment code is expected to provide a natural convergence scheme by increasing the number of moments.   

Generalized approach to variational and Hamiltonian kinetic-Maxwell plasma theory.

Hamiltonian and variational techniques have proved particularly helpful for constructing new kinetic and hybrid plasma models. Here, we present a geometric construction that relates these approaches for a generalized kinetic-Maxwell theory, producing Lagrangian and Eulerian variants. We then present a particular specialization, in which the well-known Maxwell-Vlasov theory is shown to emerge under appropriate choices of symplectic form and energy function.   

Hybrid drift-kinetic pressure coupling scheme from a variational principle.

A hybrid-kinetic pressure coupling scheme (PCS) model is presented in the low-gyrofrequency approximation. This fully energy-conserving model is derived from variations of a Lagrangian over a semidirect product manifold, and produces important terms in the Vlasov equation that correct models previously appearing in the literature.   

Beatification: Flattening Poisson brackets for plasma theory and computation

A perturbative method called beatification$^{\dagger}$ is presented for producing nonlinear Hamiltonian fluid and plasma theories. Plasma Hamiltonian theories, fluid and kinetic, are naturally described in terms of noncanonical variables. The beatification procedure amounts to finding a transformation that removes the explicit variable dependence from a noncanonical Poisson bracket and replaces it with a fixed dependence on a chosen state in the phase space. As such, beatification is a major step toward casting the Hamiltonian system in its canonical form, thus enabling or facilitating the use of analytical and numerical techniques that require or favor a representation in terms of canonical, or beatified, Hamiltonian variables. Examples will be given. \\ $\dagger$ P.~J.~Morrison and J.~Vanneste, Ann.~Phys.~{\bf 368}, 117 (2016); T.~F.~Viscondi, I.~L.~Caldas, and P.~J.~Morrison, Phys.~Plasmas {\bf24}, 032102 (2017); J.\ Phys.~A {\bf49}, 165501 (2016).   

Nonlinear saturation of the ITG instability with fully kinetic ions

We study the growth and saturation of the ion-temperature-gradient (ITG) instability in simulations with fully kinetic ions and adiabatic electrons. The ion trajectories are integrated using the full Lorentz force, fully resolving the cyclotron motion and capturing the corresponding finite Larmor radius effects. In slab geometry, the linear growth and nonlinear saturation characteristics show good agreement with analogous gyrokinetic simulations across a wide range of parameters, and the fully kinetic simulation correctly reproduces the nonlinearly generated zonal flow. We discuss our progress on the extension of this model to toroidal geometry to study more realistic turbulence. This work represents an important step towards the extension of kinetic modeling of plasma turbulence to regimes where the gyrokinetic ordering is violated or the gyrokinetic equations are questioned.   

Stabilization approach for an explicit hybrid particle-in-cell method for bridging multiple time-scales

Fully-kinetic Particle-in-Cell (PiC) simulations of magnetized plasma sheaths including electrons, plasma ions, and heavy material impurities remain a big challenge because of the large discrepancy between the mass of the light species and that of the heavy species. The time-scales required for such simulations span over multiple orders of magnitude, from picoseconds for the electron dynamics, to microseconds for heavy-ion transport across the sheath. In this work, we analyze a numerical approach applicable to explicit PiC schemes that iterates between a fully-kinetic representation of the electrons and a reduced electron model; the method allows to capture fully-kinetic effects at time scales relevant to heavy-ion transport. A frequency analysis of the approach reveals the strategy required for stable operation of the method and prevent unstable drift-like behavior in the phase space. The method has been applied to plasma sheath simulations with oblique magnetic fields. In order to highlight the benefits of the method, the moments of the distribution function are compared to both long-time fully-kinetic PiC simulations, and to PiC simulations with only a reduced electron model. Finally, comparisons with a continuum Boltzmann-Poisson code solving the same problem are reported.   

Open-source Framework for Storing and Manipulation of Plasma Chemical Reaction Data

We present a new open-source framework for storage and manipulation of plasma chemical reaction data that has emerged from our in-house project MUNCHKIN. This framework consists of python scripts and C$++$ programs. It stores data in an SQL data base for fast retrieval and manipulation. For example, it is possible to fit cross-section data into most widely used analytical expressions, calculate reaction rates for Maxwellian distribution functions of colliding particles, and fit them into different analytical expressions. Another important feature of this framework is the ability to calculate transport properties based on the cross-section data and supplied distribution functions. In addition, this framework allows the export of chemical reaction descriptions in LaTeX format for ease of inclusion in scientific papers. With the help of this framework it is possible to generate corresponding VSim (Particle-In-Cell simulation code) and USim (unstructured multi-fluid code) input blocks with appropriate cross-sections.   

A One-Step Variational Guiding Center Integrator using Toroidal Regularization

Guiding center and gyro-center particle advances --- central to test particle, drift-kinetic, and gyro-kinetic simulations --- stand to benefit from symplectic integration techniques, which have had a profound impact in other physics disciplines. The non-canonical Hamiltonian formulation of these systems has kept such symplectic integration thus far elusive, except for in restricted magnetic geometries or by using computationally expensive transformations to canonical coordinates. In this work, we perform a near-identity Lie transformation to the guiding center coordinates to obtain a ``toroidally regularized'' Lagrangian for which symplectic integration can be more readily achieved. This transformation also eliminates the effective magnetic field appearing in the denominator of the guiding center equations and correspondingly eliminates the large parallel velocity singularities from the equations. The recently developed technique of degenerate variational integration is then applied to the regularized Lagrangian to obtain a one-step variational integrator valid for any magnetic geometry with non-zero toroidal magnetic field.   

A PICKSC Science Gateway for enabling the common plasma physicist to run kinetic software

Computer simulations offer tremendous opportunities for studying plasmas, ranging from simulations for students that illuminate fundamental educational concepts to research-level simulations that advance scientific knowledge. Nevertheless, there is a significant hurdle to using simulation tools. Users must navigate codes and software libraries, determine how to wrangle output into meaningful plots, and oftentimes confront a significant cyberinfrastructure with powerful computational resources. Science gateways offer a Web-based environment to run simulations without needing to learn or manage the underlying software and computing cyberinfrastructure. We discuss our progress on creating a Science Gateway for the Particle-in-Cell and Kinetic Simulation Software Center that enables users to easily run and analyze kinetic simulations with our software. We envision that this technology could benefit a wide range of plasma physicists, both in the use of our simulation tools as well as in its adaptation for running other plasma simulation software.   

The Particle-in-Cell and Kinetic Simulation Software Center

The UCLA Particle-in-Cell and Kinetic Simulation Software Center (PICKSC) aims to support an international community of PIC and plasma kinetic software developers, users, and educators; to increase the use of this software for accelerating the rate of scientific discovery; and to be a repository of knowledge and history for PIC.  We discuss progress towards making available and documenting illustrative open-source software programs and distinct production programs; developing and comparing different PIC algorithms; coordinating the development of resources for the educational use of kinetic software; and the outcomes of our first sponsored OSIRIS users workshop. We also welcome input and discussion from anyone interested in using or developing kinetic software, in obtaining access to our codes, in collaborating, in sharing their own software, or in commenting on how PICKSC can better serve the DPP community.   

PlasmaPy: initial development of a Python package for plasma physics

We report on initial development of PlasmaPy: an open source community-driven Python package for plasma physics [1]. PlasmaPy seeks to provide core functionality that is needed for the formation of a fully open source Python ecosystem for plasma physics. PlasmaPy prioritizes code readability, consistency, and maintainability while using best practices for scientific computing such as version control, continuous integration testing, embedding documentation in code, and code review. We discuss our current and planned capabilities, including features presently under development. The development roadmap includes features such as fluid and particle simulation capabilities, a Grad-Shafranov solver, a dispersion relation solver, atomic data retrieval methods, and tools to analyze simulations and experiments. We describe several ways to contribute to PlasmaPy. PlasmaPy has a code of conduct and is being developed under a BSD license, with a version 0.1 release planned for 2018. The success of PlasmaPy depends on active community involvement, so anyone interested in contributing to this project should contact the authors. [1] The code repository is at https://github.com/PlasmaPy/PlasmaPy   

Exact Energy and Momentum Conservation in Variational Macro-Particle Plasma Models

We consider a class of variational macro-particle plasma models that exhibit simultaneous conservation of energy and momentum. These models retain translation invariance by using a Fourier representation of the electromagnetic fields in place of a spatial grid. That is, the Fourier amplitudes of the fields are the fundamental quantities. From the discrete Lagrangian, a canonical Hamiltonian system is obtained in the usual way, for which we introduce a symplectic integrator. We present a general formulation of the method with examples drawn from 1-1/2D studies of intense laser-plasma interactions. We comment on the relative merits of the Lagrangian vs. Hamiltonian formulations and discuss efficiency and practicality of using this technique in three dimensions.   

A Computational Model for Predicting Gas Breakdown

Pulsed-inductive discharges are a common method of producing a plasma. They provide a mechanism for quickly and efficiently generating a large volume of plasma for rapid use and are seen in applications including propulsion, fusion power, and high-power lasers. However, some common designs see a delayed response time due to the plasma forming when the magnitude of the magnetic field in the thruster is at a minimum. New designs are difficult to evaluate due to the amount of time needed to construct a new geometry and the high monetary cost of changing the power generation circuit. To more quickly evaluate new designs and better understand the shortcomings of existing designs, a computational model is developed. This model uses a modified single-electron model as the basis for a Mathematica code to determine how the energy distribution in a system changes with regards to time and location. By analyzing this energy distribution, the approximate time and location of initial plasma breakdown can be predicted. The results from this code are then compared to existing data to show its validity and shortcomings.   

GPU Acceleration of Particle-In-Cell Methods

Graphics processing units (GPUs) have become key components in many supercomputing systems, as they can provide more computations relative to their cost and power consumption than conventional processors. However, to take full advantage of this capability, they require a strict programming model which involves single-instruction multiple-data execution as well as significant constraints on memory access. To bring the full power of GPUs to bear on plasma physics problems, we must adapt the computational methods to this new programming model. We have developed a GPU implementation of the particle-in-cell (PIC) method, one of the mainstays of plasma physics simulation. This framework is highly general and enables advanced PIC features such as high order particles and absorbing boundary conditions. The main elements of the PIC loop, including field interpolation and particle deposition, are designed to optimize memory access. We describe recent progress in these algorithms, including arbitrary grid types and multiple GPUs per node.   

Recent Performance Results of VPIC on Trinity

Trinity is a new DOE compute resource now in production at Los Alamos National Laboratory. Trinity has several new and unique features including two compute partitions, one with dual socket Intel Haswell Xeon compute nodes and one with Intel Knights Landing (KNL) Xeon Phi compute nodes, use of on package high bandwidth memory (HBM) for KNL nodes, ability to configure KNL nodes with respect to HBM model and on die network topology in a variety of operational modes at run time, and use of solid state storage via burst buffer technology to reduce time required to perform I/O. An effort is in progress to optimize VPIC\footnote{K. J. Bowers, B. J. Albright, L. Yin, B. Bergen, and T. J. T. Kwan, Phys. Plasmas 15, 055703 (2008)} on Trinity by taking advantage of these new architectural features. Results of work will be presented on performance of VPIC on Haswell and KNL partitions for single node runs and runs at scale. Results include use of burst buffers at scale to optimize I/O, comparison of strategies for using MPI and threads, performance benefits using HBM and effectiveness of using intrinsics for vectorization.   

A three-dimensional particle-in-cell simulation of the diocotron instability for cylindrical geometry

In a non-neutral plasma like an electron beam under a magnetic field, the diocotron instability can occur with a shear in the flow velocity of surface waves, which is a type of Kelvin-Helmholtz instability in principle. Recently, there has been advanced theories that explain the evolution of the diocotron instability using non-modal analysis considering shearing modes. In a previous study, a two-dimensional particle-in-cell simulation was performed for verification of the theory with an initially loaded cylindrical annular plasma column surrounded by a conducting boundary. The growth rates of the diocotron instability measured in the simulation agree well with the theory. As an extension of the previous work, we have extended the model to a three-dimensional cylindrical particle-in-cell simulation and compared the results with those of the two-dimensional simulation. In addition, the effect of the particle flows in the axial direction has been investigated.   

An Approach to Radiation Hydrodynamics Within a Generalized Continuum Mixture Theory

Mixed material cells have long posed challenges for radiation hydrodynamic treatments in multi-physics codes. Baer and Nunziato [1] developed a two-phase mixture formulation which is recently expanded by Baer and Schumacher to a generalized model. We present an extension to this generalized continuum mixture theory by this model incorporating energy-based\sout{ }gray\sout{ }radiation diffusion in a multi-component fluid. The extended model is presented along with a numerical approach including verification examples. [1] Baer, M.R. and Nunziato, J.W., Int. J. Multiphase Flow, Volume 12, No. 6 1986   

KEEN Wave Simulations: Comparing various PIC to various fixed grid Vlasov to Phase-Space Adaptive Sparse Tiling {\&} Effective Lagrangian (PASTEL) Techniques

We compare various ways of solving the Vlasov-Poisson and Vlasov-Maxwell equations on rather demanding nonlinear kinetic phenomena associated with KEEN and KEEPN waves. KEEN stands for Kinetic, Electrostatic, Electron Nonlinear, and KEEPN, for electron-positron or pair plasmas analogs. Because these self-organized phase space structures are not steady-state, or single mode, or fluid or low order moment equation limited, typical techniques with low resolution or too much noise will distort the answer too much, too soon, and fail. This will be shown via Penrose criteria triggers for instability at the formation stage as well as particle orbit statistics in fully formed KEEN waves and KEEN-KEEN and KEEN-EPW interacting states. We will argue that PASTEL is a viable alternative to traditional methods with reasonable chances of success in higher dimensions.   

Code Modernization of VPIC

The ability of scientific simulations to effectively deliver performant computation is increasingly being challenged by successive generations of high-performance computing architectures. Code development to support efficient computation on these modern architectures is both expensive, and highly complex; if it is approached without due care, it may also not be directly transferable between subsequent hardware generations. Previous works have discussed techniques to support the process of adapting a legacy code for modern hardware generations, but despite the breakthroughs in the areas of mini-app development, portable-performance, and cache oblivious algorithms the problem still remains largely unsolved. In this work we demonstrate how a focus on platform agnostic modern code-development can be applied to Particle-in-Cell (PIC) simulations to facilitate effective scientific delivery. This work builds directly on our previous work optimizing VPIC, in which we replaced intrinsic based vectorisation with compile generated auto-vectorization to improve the performance and portability of VPIC. In this work we present the use of a specialized SIMD queue for processing some particle operations, and also preview a GPU capable OpenMP variant of VPIC. Finally we include a lessons learnt s   

Evaluation of tungsten effect on CFETR phase I performance

An integrated modeling workflow using OMFIT/TGYRO is constructed to evaluate Tungsten (W) impurity effects on China Fusion Engineering Test Reactor (CFETR) performance. Self-consistent modeling of W core density profile, accounting for both turbulence and neoclassical transport contribution, is performed based on the CFETR steady-state scenario developed by D.Zhao (ZhaoDeng, APS, 2016). It's found that the fusion performance degraded in a limited level with increasing W concentration. The main challenge arises in sustainment of H-mode with significant W radiation. Assuming the power threshold of H-L back transition is approximately the same as that of L-H transition; it is found that the W concentration is not allowed to exceed 3e-5 to stay in H-mode for CFETR phase I according to the scaling law found by Takizuka (Takizuka etc, Plasma Phys. Control Fusion, 2004). In addition, it's found that the tolerance of W concentration decreases with increasing pedestal density by trade-off study of pedestal density and temperature. A future step is to connect this requirement to W wall erosion modeling.