Session 7F - DIII-D: Edge and H-Mode.
ORAL session, Thursday morning, November 14
Vail Room, Adam's Mark

Recent \mboxDIII--D fusion science experiments have emphasized control and deployment of transport barriers for fundamental turbulence studies and optimization of MHD stability for tokamak performance improvement. These experiments have been carried out varying the plasma current profile in modest and strongly shaped neutral beam heated plasmas with-and-without divertor pumping, pellet fueling, and fast wave current drive. Systematic disruption studies employing high speed diagnostic capability together with impurity pellet injection have provided new understanding and tools for real time disruption mitigation and amelioration. Divertor experiments have demonstrated the feasibility of the required low divertor temperatures and are now concentrating on developing physics predictive understanding through detailed diagnostic measurements benchmarked with modeling codes. The future DIII--D program will emphasize demonstration of integrated advanced tokamak performance with an improved high-triangularity divertor with baffling of the main plasma region. Participation of collaborators is being facilitated by remote access and Internet communications.

[7F.02] Attainment of High Density in \hboxDIII--D H--mode Plasmas

We report results of experiments designed to identify physical effects that define the tokamak n_e limit. Our results support the postulate that the n_e limit is defined by two generic groups of physical effects, power balance and MHD stability. Power balance limits of the core, boundary, and divertor plasmas result in a n_e limit that increases with heating power P, typically n_e\hbox(max)\propto P^0.5. MHD modes, such as ballooning or neoclassical tearing, limit P from above (through transport) with P_max\propto I_P^2. Thus the n_e limit increases with I_P. With typical H--mode density profiles, divertor and boundary plasmas dominate in the power balance. Using n_e profile control, by divertor pumping and pellet injection, we have increased the \bar n_e relative to the edge values. As a result we have succeeded in increasing the H--mode n_e limit by a factor of two and up to the core radiative limit. Densities up to 1.5 times the Greenwald limit with \tau_E^ \simeq0.9\times\tau_E (JET\slash \hboxDIII--D H--mode) have been obtained.

[7F.03] Enhanced Distributed Radiation in an Extended Divertor in DIII--D

The essential concept of the ITER divertor is to spread the heat load along the side wall of the slot by radiation from an extended zone along the divertor leg. We have produced such an extended radiating zone in \hboxDIII--D. On DIII--D a plasma equilibrium was produced with an outer divertor poloidal length of >50 cm, approximately the same dimension as the main plasma minor radius. This configuration gives five bolometer chords an unambiguous view of the outer divertor to produce an accurate radiation profile from the X--point to the divertor floor. Deuterium gas is puffed into the divertor region until radiation is significantly increased and peak divertor heat flux is reduced by a factor of 3--5. The strong radiation in the divertor is seen to vary by no more than a factor of 2 along the outer let from the X--point to the floor exceeding ITER's requirement of 6:1 uniformity. The distribution of deuterium and intrinsic carbon radiation will be discussed and modeled.

[7F.04] Divertor Plasma Parameters During Radiative Divertor Operation on DIII--D

A large array of divertor diagnostics has been used to characterize the DIII--D divertor conditions during radiative divertor operation. We have used both D_2 and impurities to reduce the divertor heat flux. Several discharge conditions have been obtained, including attached and detached ELMing H-modes. The multi-chord Divertor Thomson Scattering (DTS) system has been used with divertor sweeping to obtain 2-D measurements of n_e and T_e in the divertor. The T_e drops to \leq 2 eV with D_2 puffing, n_e increases, and the electron pressure P_e decreases. The radiation zone, measured by multi-chord bolometry, moves from the inside leg of the divertor to the outside. Comparisons of the 2-D distribution of n_e and T_e and the radiation distribution will be presented.

[7F.05] Modeling Detached Plasmas in DIII-D

The ITER divertor design relies on operation of the machine with a detached divertor plasma as a means of reducing the divertor heat load to manageable levels. This operating mode has been seen on all of the world's diverted tokamaks, and is characterized by very low plate temperatures and ion currents. Experimental results on DIII-D have shown the plate electron temperature is between 1 and 2 eV. We describe the results of modeling these detached plasmas with the UEDGE code in this paper. Plasma detachment can be achieved in a variety of ways in the code as well as in experiment. Simulations indicate the detachment process occurs in two steps: a thermal collapse in which the plate temperature drops to 1 to 2 eV, followed by a decrease in the plate ion current. When the low temperature region extends off the plate, parallel momentum of the plasma is reduced by ion/neutral interactions. The plate ion current decreases when the parallel momentum is reduced sufficiently to permit volume recombination processes to compete with ion flow to the plate.

[7F.06] Measurements of Divertor Impurity Emissions on DIII--D

On DIII--D, visible and ultraviolet spectroscopy are used to characterize divertor emissions. Carbon emissions during ELMing H--mode and D_2 radiative divertor operations account for 50%--80% of the total radiated power in the divertor with the remainder from deuterium.(R.D. Wood, et al.), 22nd EPS Conference on Controlled Fusion and Plasma Physics, Kiev, Ukraine, 1996.^,(R.C. Isler, et al.), submitted to Physics of Plasmas, 1996. In the case of neon injection with D_2 puffing and pumping, the strong radiating zone near the X-point consists of 80% carbon and less than 20% neon. With nitrogen injection, about half of the total radiated power comes from nitrogen. In all these cases, the total radiated power derived from spectroscopy is in good agreement with bolometer measurememts. Details of the spectroscopic analysis and a comparision of these data with two-dimensional radiated power density distribuiton and visible TV data will be presented.

[7F.07] Neon Enrichment and Particle Balance During ``Puff and Pump'' Experiments on DIII--D

We report the results of experiments conducted on DIII--D with simultaneous external D_2 gas injection (\sim 100--200 torr-\ell/s) and divertor exhaust to assess the dependence of scrape-off-layer flow on divertor retention of neon. During these experiments, direct measurements have been made of the neon concentration in the core (CER spectrscopy) and divertor (SPRED UV spectroscopy) plasmas and in the exhaust gas (modified Penning gauge). In these experiments, exhaust enrichment (ratio of neon fraction in exhaust gas to the neon fraction in the core plasma) is in the range 1--2, suggesting that significant compression of neon in the divertor plasma is difficult. Systematic scans of various parameters expected to determine divertor retention (e.g., external D_2 gas flow magnitude and location, divertor recycling, ELM frequency, gas fueling) have been carried out. Particle balance analysis suggests that wall retention is likely a major player in the overall neon particle balance in these discharges.

[7F.08] Absorption of Fast Wave Power By Energetic Ions During FWCD in DIII--D

Fast wave current drive can be adversely affected by parasitic absorption processes which compete with direct electron heating for the RF power. One such competing absorption process is high harmonic resonant ion absorption by energetic deuterium ions which are used to heat the plasma during neutral beam injection. Harmonic resonance numbers for these ions vary between about 3 and 9 in DIII--D depending on the toroidal magnetic field and RF frequency. In this study, the toroidal magnetic field is varied between about 1.4 T and 2.1 T at the machine center to study the effect of energetic ion absorption and plasma beta on electron heating and current drive. Results are compared to theoretical calculations with the PICES full-wave ICRF code. The energetic ion component is treated as a Maxwellian tail with mean energy and density derived from fast ion slowing down calculations from the ONETWO and TRANSP codes.

[7F.09] ELM Studies on \hboxDIII--D

Three classes of Edge Localized Modes, or ELMs, simply labeled Types I, II, and III, in the \hboxDIII--D work, are widely observed. Type I ELMs are distinguished by the fact that their frequency increases with increasing input power. A scaling for the Type I ELM energy loss predicts 3% for ITER. The frequency of Type III ELMs decreases with increasing input power. The energy loss per Type III ELM is a factor of 2 to 4 below that for Type I at the same input power. Experiments on \hboxDIII--D suggest that proximity to the H--mode threshold power is the critical parameter for Type III ELMs. In contrast to Type I ELMs, the pressure gradient near the separatrix at a Type III ELM is often well below the ideal ballooning mode limit. Medium n magnetic precursor oscillations are observed with Type III ELMs, while no magnetic precursors are observe with Type I ELMs. Type II ELMs have very high frequency and low energy loss compared to Type I ELMs. Type II ELMs do not require low input power in contrast to Type III ELMs. Type II ELMs are associated with edge second stability in combination with either high q or high \beta_P.

[7F.10] Recent Physics Results from the DIII--D Disruption Program

Recent disruption experiments on DIII--D have provided a variety of new characterization and mitigation data with which to better understand the physics of disruptive instabilities. Peak halo current amplitudes are reduced by up to 50% in triggered VDEs with both neon and argon ``killer'' pellets. Halo current toroidal peaking factors are also reduced from \sim 3 to 1.1 for these discharges. Impurity radiation accounts for at least 90% of the thermal quench during ``killer'' pellet injection. Runaway electrons are generated on some neon ``killer'' pellet injection discharges but not others. Argon ``killer'' pellets typically generate more runaway electrons than neon. Runaway electrons are also seen during some negative central shear (NCS) disruptions. Results on the successful implementation of a real-time neural network used to predict the high beta disruption boundary will also be discussed.

[7F.11] In-Vessel Component Materials Dynamic Compatibility under DT-Plasma Bombardment

If in-vessel components in a magnetic fusion device are made of more than one materials, materials mixing will be unavoidable due to plasma-induced erosion and redeposition. This materials mixing alters surface characteristics, which can then affect the wall lifetime as well as edge plasma interactions. Related to such materials mixing, we have observed an interesting phenomenon in the recent PISCES-B Mod experiments that under deuterium plasma bombardment beryllium tends to be deposited with carbon impurities, a phenomenon referred to as "carbon poisoning"[1]. Importantly, carbon poisoning reduces beryllium erosion significantly. Recently, a similar phenomenon has been observed in JET where beryllium and carbon are used for in-vessel components[2]. First-order modeling on carbon poisoning has been carried out successfully and the mechanism is explained as carbon impurity trapping followed by thermally activated mixing with beryllium via inter-diffusion. Consistent with this model, experimental data indicate that carbon poisoning occurs only at elevated temperatures. In the present work, this model has been used to analyze materials dynamic compatibilities under DT-plasma bombardment for a variety of impurity-surface combinations that are relevant to existing fusion experiments and also future reactor designs. These combinations include: lithium, beryllium, boron, and carbon as the impurities and carbon, vanadium, molybdenum, and tungsten as the surface materials. [1] Y. Hirooka et al. J.Nucl.Mater. 230(1996)173. [2] H. Guo et al. Presented at the 12th PSI-Conf. St.Raphael (1996).

[7F.12] L--H Power Threshold and Transition Studies in \mboxDIII--D

Neutrals have been suspected of playing a role in the L--H transition and may be a hidden variable in the determination of the power threshold scaling. Generally, there is a strong correlation between electron density and neutral pressure which makes it difficult to separate the role of neutrals from that of the density in the L--H transition. We have performed density ramping experiments using gas puffing and cyropumping to break this correlation in order to study the effects of neutrals on the L--H transition and the power threshold. Preliminary analysis indicates that for a factor of two increase in the neutral pressure, the power threshold increases by 50 percent. Further analysis of the edge conditions, including neutral particle modeling, will be presented and compared with theoretical models of E_r formation and poloidal rotation. The power hysteresis of the H--mode and the conditions leading to a back transition have also been investigated. The H--L transition occurs on about a 100 \mus time scale and a 30%--60% power hysteresis is observed.

Part 7 of program listing