A new method of gas fueling——pulsed supersonic molecular beam injection （MBI） has been introduced in the HL-1M tokamak. Its performances are something between the gas puffing （GP） and the pellet injection. The mean velocity of well collimated hydrogen beam in the vacuum chamber is above 500 m/s. With

Improved confinement is one of the main topics in the tokamak experiment research. Several improved confinement regimes have been explored in the plasma biasing, LHCD and pellet injection experiments on the HL-1M tokamak. Here we concentrate the enhanced confinement regimes observed in HL-1M Ohmic discharges.

This paper describes a Mach probe array with four pins which can measure not only parallel flows but the flow perpendicular to the magnetic field as well. Experimental measurements of the fluctuations and velocities of the toroidal and the poloidal flow have been carried out on both of SOL and the boundary region of HL-1M for ohmic, biased H-mode and LHCD discharges. The results show that the suppressions of the fluctuations are related to poloidal rotations produced by electrode biasing and LH-waves injecting in the improved particle confinement modes.

A multi-functional measurement probe which consists of a Langmuir probe array, a group of directional probes, a Mach number probe and two magnetic probes, is installed in the HL-1 tokamak to study the edge plasma behaviors. This is the first time the Mach number of plasma flow is measured in the HL-1 tokamak using a Mach probe. The local Mach number range is from - 0.3 to 0.4. An experimental model of the potential profile for the two sides of the Mach probe has been suggested. By using the directional probe, we have studied anisotropies of floating potential and ion saturation current in taroidal, poloidal and radial directions. During ECRH, the floating potentials decrease with increasing ion saturation current but their radial profiles remain unchanged, indicating that the diffusion processes do not change significantly. Meanwhile the outward particle flux increases, suggesting a degradation in particle confinement and the increase of the Mach number seems to be contrary to the results of confinement improvements.

A computer code SAWMOD was developed to simulate the sawtooth oscillation by making use of a cylindrical plasma transport code TRANPY. Two models, the reconnection model and the turbulent model, were employed to describe the sawtooth oscillations. The former is mainly used to analyze the sawtooth of high qa discharges, while the latter is applicable to the low qa discharges (qa<0.9). With the SAWMOD code, analyzed in detail were the temperature, density, particle and energy confinement time, sawtooth period and inverse radius for a typical discharge with high density in the HL-1 tokamak. The calculated results from the two models are in good agreement with those from experiment.

A wall surface carbonization experiment was carried out with an ac discharge on the HL-1 tokamak and homogeneous carbon films were deposited. An empirical criterion for the carbon layer formation has been presented, taking into account the ratio of the partial pressures of carbon oxide and O-H species. The wall condition discharge just after carbonization with Kr is better than that with He for depleting hydrogen and oxygen within the carbon layer. An analysis of the layers deposited on INCONEL GH39 samples, which were exposed to the carbonization process, regarding chemical composition on the surface and in-depth profiles, interface formation and thickness, has been made with SEM, AES and SIMS. Impurity contents in the plasma during the tokamak discharge are measured by the analysis of resonance line intensities and QMA, metal contents are reduced to 20%. The hydrogen particle confinement time tau(p) and recycling coefficient R increased by about 20% compared with those before carbonization.

Experiments with a biased pump limiter of positive potential have been performed on the HL-1 tokamak. The line averaged density n(e)BAR increases by 50%. The D(alpha)(or H(alpha)) light from the plasma edge decreases a little (10%). The particle flux signal taken at the plasma edge decreases by 50%. The OVI signal at the edge increases by 60% and the loop voltage increases by 20%. The particle flux signal in the throat decreases by 60%. The particle confinement time tau(rho)* increases by 50-100%, the energy confinement time tau(E) increases by 20-40%. But the CIII, CrI lines near the neutralizer plate increase for both positive and negative biasing. Generally, effects with negative biasing are much less pronounced than with positive biasing. For example, the line averaged density n(e) increases by 25% only with a negative biasing of 200 V.

Density limit disruptions have been observed in the HL-1 device. The characteristics are that the displacement of the magnetic surface toward the vacuum vessel is very small and sawteeth always exist before disruption. After disruption, the current drops partly or wholly, and the dropping time is about 20ms.In this paper, we carefully analyze the characteristics of density limit disruptions during current ramp-up, flattop and ramp down phases.One of these limits is the Murakami limit which has no precursor and the plasma energy quenches in-0.1ms. The disruption first takes place at the side of strong magnetic field. Another is the Hugill limit which has precusors relaxed for several milliseconds. The reason is that the radiation changes the profile of temperature and current density, and then excites MHD instability. Most discharges can recover from the disruption. The maximum Murakami parameter is 0.35×1020m-2·T-1 in the HL-1 device.

H-mode like discharges have been achieved by electrode biasing on the HL-1 tokamak. During the biasing period, the energy confinement time tau(E) increased and density fluctuations were suppressed. It is shown that the fluctuation spectrum with biasing is very different from the unbiased case. The auto-power spectrum value of the density fluctuation decreased, and the fluctuation power had a different value as a function of the biased voltage variation.

The transition of a low density plasma to high density in ohmic discharge was demonstrated in the HL-1 tokamak with metal wall and limiter configuration. The line averaged density n(e)BAR jumps up from 3-4 X 10(19) m-3 to 7-10 X 10(19) m-3 within 50-80 ms by switching off additional He pulse gas puffing into the D+ plasma and from 2.5-3 X 10(19) m-3 to 5-7 X 10(19) m-3 within 50-70 ms by turning off additional D2 pulse gas puffing. The density transition characterized by the improvement of confinement is associated with a change of the edge plasma parameters in terms of temperature, density and their profiles in SOL as well as neutral particle pressure and D(alpha) emission. Our investigations concentrated on the edge plasma phenomena in pure deuterium plasma and mixed deuterium and helium plasma before and after the density transition. Central impurity (OVI, CIII) line emission intensities and total radiation losses were also observed. The comparison of the experimental results of the edge plasma in low density with high density ones can be very informative for understanding the whole mechanism of confinement improvement.

This paper describes a pump limiter experiment in the HL-1 Tokamak. The pump limiter consists of a single-throat head covered by graphite tiles, a stainless steel neutralizer plate and a limiter chamber and it is pumped by a 1500 1 s-1 turbomolecular pump.The experimental results showed that the exhaust efficiency epsilon of the pump limiter for particles of the scrape-off-layer (SOL) was about 0.07 and the pump limiter was capable of exhausting helium ash from the enhanced H-alpha line and HeI line emissions in the vicinity of the neutralizer plate in H2 (60%) + He(40%) mixture gas discharges. Compared with the discharges without a pump limiter, the recycling coefficient R decreased about 20% (when nBAR(e) approximately 3-4 x 10(13) cm-3), and the particle confinement time tau-p increased about 18%. The average heat load measured over the graphite limiter head was about 250 W cm-2 in normal discharges. The non-symmetric distributions of the heat load in toroidal and poloidal directions were also observed.

In this paper we describe experiments undertaken to evaluate the retention of target produced impurities in the ASDEX divertor. The titanium divertor plates have recently been replaced by water-cooled copper plates. Since copper is a material not previously present inside the ASDEX torus, a unique possibility existed to study the transport of the target plate material from the divertor chamber into the main plasma. By spectroscopic measurement of the Cu flux emerging from the target plate and of the Cu density in the main plasma, the retention is assessed. We find a poor retention (R = 1) for low ne (ne < 3.0 × 1013 cm−3) but a substantially improved one (R > 30) at high densities. During additional heating (N1, ICRH) at the same density as in the OH case, the retention is lower, but approaches similar values at slightly higher densities. The retention seems to be independent of the heating power.

Investigations of impurity accumulation phenomena in ASDEX are reviewed. There are four different operating regimes where pronounced accumulation is observed and these regimes are also characterized by improved energy confinement. In particular, medium-Z metallic ions are involved in accumulation processes whereas low-Z ions appear almost unaffected.

Experiments on the auxiliary heating, fueling of plasma and wall conditioning were carried out on HL-1M. ECRH experiments were conducted successfully with Te increase more than 50%. The double sawtooth in soft X-ray radiation were observed, which imply that the reversed magnetic shear could be formed during ECRH. An eight-shot pellet injector (PI) was used for experiments. After the pellet injection, a hollow electron temperature and peaked density profile were obtained, accompanied with the increase of energy confinement time. The pellet ablation process was investigated with a CCD camera and an H¶ ¡ emission detector array. Obviously asymmetry in the pellet cloud was observed in both the toroidal and poloidal direction. It is found that the velocity of pellet is slowed down obviously after the pellet enters into the plasma. The safety-factor q-profile was estimated with the inclination angle of ablation cloud with respect to the torus. Density limit investigations have been performed at different wall condition with three kinds of fuelling methods. It is found that higher density limit can be achieved in following conditions: one is the strong reduction of the impurity content after siliconization, another is the peaked density profile with pellet injection and/or SMBI. With a NBI system of 1MW, preliminary results of NBI experiments were obtained with increase of ion temperature from 600eV to 800eV.