Characteristics of divertor heat flux distribution with an island divertor configuration on the J-TEXT tokamak

1.   Introduction

In the future, large fusion devices like ITER (International Thermonuclear Experimental Reactor) and CFETR (China Fusion Engineering Test Reactor) will need to operate long-pulse high-power burning plasmas. How to effectively and safely control the power and particle exhaust so it does not exceed the tolerance limits of the target material is one of the great challenges currently restricting the further development of fusion research [1]. In recent years, developing advanced magnetic divertor configurations to address the coupling of heat and particle exhaust with impurity control and access to plasma detachment has attracted increasing interest in fusion research and led to research on the use of a radiative snowflake divertor [2] and super-X divertor [3] on tokamaks. External resonant magnetic perturbation (RMP) has also been considered as an attractive method to mitigate or suppress the large transient heat and particle loads induced by the edge localized mode (ELM) on tokamak, and will be implemented on ITER [4–7]. It has been identified that RMP will induce three-dimensional (3D) boundary magnetic structures in the tokamak plasma [8, 9]. Therefore, the impact of 3D boundary magnetic structures on the divertor heat flux pattern has been studied on various tokamaks [10–12]. The island divertor concept is an innovative and feasible idea to handle heat and particle exhaust in stellarators. On the W7-X stellarator device, the divertor heat flux distribution under a 3D magnetic structure has been studied. The simulation results show that a higher plasma beta can enhance stochastization at the plasma boundary and has profound impacts on access of divertor detachment [13–15].

Reduction of the peak heat flux of the divertor plate by gas fueling has also been achieved on many devices such as ASDEX Upgrade [16], JET [17], DIII-D [18], and EAST [19]. ELM mitigation by supersonic molecular beam injection (SMBI) was also carried out on KSTAR and HL-2A [20]. SMBI is an effective gas fueling method which has a deeper penetration depth compared to normal gas puffing. It was observed that the ELM-induced transient peak ion saturation current and thermal load on the divertor plate decreased significantly with SMBI. These experimental results demonstrate that SMBI can be a very effective tool for heat flux control on the divertor target.

On EAST, the lower hybrid wave (LHW) can induce a change of magnetic topology and form a secondary strike point (SSP) on the lower outer divertor plates [19]. In this experiment, changes in heat flux distributions were observed in LHW heated plasma during deuterium (D) pellet injection or SMBI, i.e., a decrease at the original strike point (OSP) and an increase at the SSP. This result implies that the magnetic topology may impact significantly the mitigation of the divertor heat load induced by gas injection. Therefore, it is of great significance to study the impact on divertor heat load of SMBI under 3D boundary magnetic structures, and this will be important for the further exploration of stable detachment operation in long-pulse high-power burning plasmas.

On the Joint-Texas Experimental Tokamak (J-TEXT), edge magnetic islands can be formed using the island divertor system, which consists of six sets of saddle coils. These coils can induce radial magnetic perturbations with dominant components of m/n = 3/1 and 4/1 [21], and the phase of the coils current can be set as 0° or 180°. Here, m and n are poloidal and toroidal mode numbers. The island divertor configuration can be established by varying plasma edge safety factor, ${q}_{a}$, and moving edge islands opened by the HFS target plate. In order to obtain the temperature and heat flux distribution of the divertor plate, an infrared (IR) camera [22] thermography diagnostic system was recently installed on J-TEXT that allows investigation of heat flux pattern on the HFS target plate. In this work, the dynamic of divertor heat flux distribution during SMBI fueling in an island divertor configuration will be discussed. It has been observed that the peak heat flux on the HFS target plate can be distinctly reduced after each SMBI pulse, and this has a strong correlation with the penetration depth of the gas injection. One purpose of this work is to investigate divertor heat load pattern induced by the formation of island divertor configuration using the IR camera. Another purpose is to analyze the behaviors of divertor heat load during the SMBI fueling process and explore the relationship between the penetration depth and the reduction of peak heat flux.

The structure of this paper is as follows. In section 2, the experimental setup, key diagnostics used in this work, and the heat flux calculation model are introduced. In section 3, the divertor heat flux distribution in island divertor configurations is measured and compared to simulation results. In section 4, the power detachment by SMBI gas fueling is analyzed. In section 5, the effect of injected molecule number and frequency on detachment is determined. Section 6 gives the discussion and conclusion.

2.   Experimental setup

The J-TEXT tokamak [23] is a circular, middle-sized tokamak with a major radius ${R}_{0}=1.05$ m and minor radius $r=0.25-0.29$ m, and the central line-averaged electron density, $\overline{{n}_{\mathrm{e}}}$, is in a range of $\left(0.5-7\right)\times {10}^{19}$ m⁻³. The island divertor configuration has been established by applying island divertor coils [24]. In this experiment, one toroidally and poloidally localized plate at the HFS, so called the HFS target plate, has been used as the main plasma limiter. This means that all of the power from the plasma boundary excluding the radiation loss will be deposited on the HFS target plate, which is fully encompassed within the field of view of the IR camera, as shown in figure 1(b). Hence, it is convenient to study the divertor heat flux distribution and explore the detachment regime on the tokamak with 3D boundary magnetic structure.

Recently, SMBI has been installed on the J-TEXT tokamak and can be used to conveniently fuel the working gas (H₂) or impurity gas such as Ar, Ne, or He [25]. This system can realize vertical injection from bottom of vacuum chamber at Port 10 of the J-TEXT tokamak, as shown in figure 1(a). In this work, H₂ was injected by SMBI at a backing pressure of 0.18 MPa. The pulse width varies between 0.5 and 2 ms, corresponding to the evaluated hydrogen molecular number ranging between $0.275\times {10}^{19}$ and $0.602\times {10}^{19}$ in each pulse.

Figure  1.  (a) Diagram of the location of diagnostic systems mentioned in this paper. (b) Picture captured by IR camera before discharge.

DownLoad: Full-Size Img PowerPoint

Based on these conditions, a series of experiments have been carried out to measure distributions of the heat flux on the HFS target plate and investigate detached operation by SMBI gas fueling with a 3D boundary magnetic structure. All results reported in this paper were obtained in Ohmic hydrogen discharges. The parameters of the plasmas are as follows: ${B}_{\mathrm{T}}=1-1.8\;\mathrm{T},\;\;{I}_{\mathrm{P}}=60-150\;\mathrm{k}\mathrm{A},\;\;\overline{{n}_{\mathrm{e}}}=\left(1-3\right)\times {10}^{19}\;{\mathrm{m}}^{-3}, \; {I}_{\mathrm{c}}=0-6\;\mathrm{k}\mathrm{A};$ the phase of ${I}_{\mathrm{c}}$ is 0° or 180°, where ${B}_{\mathrm{T}}$ is the toroidal magnetic field, ${I}_{\mathrm{P}}$ is the plasma current, $\overline{{n}_{\mathrm{e}}}$ is the central line-averaged electron density, and ${I}_{\mathrm{c}}$ is the current of island divertor coils. More detailed discharge parameters will be introduced in the next section.

2.1   Layout of key diagnostics

The infrared diagnostic system on J-TEXT consists of a TEL-1000MW InSb IR camera of TELOPS, a 150 mm diameter sapphire window, and an external trigger system. The IR camera is located at Port 6, as shown in figure 1(a), which views horizontally the HFS target plate through the sapphire window. The operational bandwidth of the IR camera ranges from 3.7 to 4.95 $\mu \mathrm{m}$, enabling temperature measurement within the range of −20 °C to 180 °C. It offers a sampling rate of 50 Hz at a full frame size of 640×512 and up to 1 kHz at a reduced frame size of 64×320. The picture captured from the IR camera before discharge is shown in figure 1(b), which includes the whole HFS target plate, Langmuir probes arranged in three columns, and four bolts used to hold the target plate. In this picture, the spatial resolution is up to 0.6 mm pixel⁻¹ and the coverage area spans approximately 0.25 m vertically and extends over an angle of 25° in the toroidal direction.

Compared with the common visible light window, the sapphire window has a weaker light absorption in the infrared band and higher mechanical hardness. It is necessary to determine the transmittance of the sapphire window at Port 6 for more accurate temperature measurement. Therefore, the IR camera is calibrated through the sapphire window using a blackbody source before experiments.

On the HFS target plate, there are 21 rows of Langmuir probes, evenly distributed across the vertical range of $-0.08\;\mathrm{t}\mathrm{o}\;+0.08\;\mathrm{m}$. Each row of three Langmuir probes forms a set of triple probes, measuring plasma floating potential, ion saturation current, electron temperature and density, and so on. The absolute extreme ultraviolet (AXUV) arrays were also updated and optimized on J-TEXT. The AXUV arrays aim to provide more accurate and reliable information about the total radiated power [26]. The view sight of AXUV arrays can cover the entire poloidal cross-section in $\phi\ =68\text{°}$ on J-TEXT, as shown in figure 1(a).

2.2   Heat flux calculation of the HFS target plate on J-TEXT

A graphite target plate, 28 cm×22 cm in length and width with a thickness of 4 cm, was installed at the HFS of J-TEXT. A two-dimensional heat flux calculation model has been developed on J-TEXT, as shown in figure 2. The model assumes that thermal conduction occurs solely in the depth and axial directions, rather than in the toroidal direction.

Figure  2.  2D heat flux calculation model of the HFS target plate.

DownLoad: Full-Size Img PowerPoint

The temperature inside the target plate model during discharge can be derived by solving sequentially the two-dimensional heat diffusion equation,

where $\rho$ is the density, $c$ is the specific heat capacity, and ${k}_{z}$ and ${k}_{y}$ represent heat conduction coefficients of the target material in the axial and depth directions, respectively. For the HFS target plate on J-TEXT, ${k}_{z}={k}_{y}$ has been considered in this model. In the calculation process, it is assumed that there is no heat transfer between the upper and lower boundaries (Side 1 and Side 3 in figure 2) of the target plate and the external environment, which can be expressed as

The HFS target plate is not actively cooled, and the supporting structure of the HFS target plate is connected directly to the vessel wall, which cools naturally in the air (Side 2 in the figure 2). This boundary condition can be expressed as

where ${k}_{\mathrm{b}\mathrm{o}\mathrm{t}\mathrm{t}\mathrm{o}\mathrm{m}}$ is the convective heat transfer coefficient, ${T}_{\mathrm{b}\mathrm{o}\mathrm{t}\mathrm{t}\mathrm{o}\mathrm{m}}$ is the temperature of the bottom layer in this model, and ${T}_{\mathrm{E}\mathrm{N}\mathrm{V}}$ is the environment temperature. Furthermore, the surface temperature of the target plate measured by the infrared camera at every moment is input to the upper layer of the model. This boundary condition can be expressed as

where $\phi_0$ is a certain toroidal position, and then the heat flux distribution on the surface of the HFS target plate can be calculated through the difference of temperature between the upper layer and the second layer, as expressed by the equation

3.   Heat flux distributions under the island divertor configuration

In the J-TEXT discharge #1089444, a stable plasma was operated with ${B}_{\mathrm{T}}=1.5\;\mathrm{T}$ and ${I}_{\mathrm{P}}=85\;\mathrm{k}\mathrm{A}$ (${q}_{a}\sim 4.3$) before the island divertor coil current was applied. The m/n = 4/1 and 3/1 magnetic islands were formed at the plasma edge when ${I}_{\mathrm{c}}$ ramps up to +5 kA at $t=0.33\;\mathrm{s}$, as indicated by the black dotted line in figure 3. A beta-induced Alfvén eigenmode (BAE) was observed at $t=0.33\;\mathrm{s}$ in the spectrogram of the poloidal magnetic fluctuation, $\delta {B}_{\theta }$, measured by Mirnov coils as shown in figure 3(e). In previous experiments, it was found that the emergence of BAE mode is often related to the formation of m/n = 3/1 islands on J-TEXT [27]. The contours of the ion saturation current, ${I}_{\mathrm{s}}$, and the floating potential, ${V}_{\mathrm{f}}$, measured by the Langmuir probes, are shown in figures 3(c) and (d). The divertor footprints pattern changed significantly when the plasma configuration was modified from the limiter configuration to the island divertor configuration at $t=0.33\mathrm{\ s}$. Then a scan of the edge ${q}_{a}$ from 4.3 to 3.3 was performed by ramping up the ${I}_{\mathrm{P}}$ from 85 to 112 kA during $t=0.37-0.52\mathrm{\ s}$, while ${I}_{\mathrm{c}}$ keeps constant during this process. With ${q}_{a}$ deceasing, the footprint pattern on the HFS target plate changes continuously due to a change of boundary magnetic structure.

Figure  3.  Time evolution of (a) ${I}_{\mathrm{P}}$ (black line) and ${q}_{a}$ (red line), (b) ${I}_{\mathrm{c}}$ (black lines) and central line-averaged density $\overline{{n}_{\mathrm{e}}}$ (red lines), (c) contour of the ion saturation current ${I}_{\mathrm{s}}$, (d) contour of the floating potential measured ${V}_{\mathrm{f}}$, (e) spectrogram of $\delta {B}_{\theta }$ from one toroidal Mirnov probe.

DownLoad: Full-Size Img PowerPoint

Figure 4(a) shows the temporal evolution of heat flux on the HFS target plate at the toroidal angel $\phi =140 {\text{°}}$. Before the application of the ${I}_{\mathrm{c}}$, the plasma is operated in a limiter configuration and the heat flux is mainly distributed between $Z\sim0.1\mathrm{\ m}$ and $Z\sim-0.07\mathrm{\ m}$ on the target plate. When the m/n = 4/1 island forms and interacts with the HFS target plate ($t=0.33\mathrm{\ s}$), a strike point splitting occurs. There is a dynamic evolution in the position of strike points throughout the ${I}_{\mathrm{P}}$ ramp-up stage. Therefore, the heat flux distribution on the HFS target plate also changes. Poincare plots and field-line connection length distributions calculated by the 3D non-linear MHD equilibrium code HINT [28] under experimental conditions at different moments of discharge #1089444 are shown in figure 4(b). As $q_a\ $decreases the m/n = 3/1 island moves outward and is opened by the HFS target plate, as shown in figure 4(b). This results in dynamic evolution of the heat flux distribution on the HFS target plate.

Figure  4.  Temporal evolutions of (a) heat flux at the HFS target plate, ${I}_{\mathrm{P}}$ (blue line), and ${q}_{a}$ (red line). (b) Poincare plot and length of magnetic field line.

DownLoad: Full-Size Img PowerPoint

The plasma was operated with ${B}_{\mathrm{T}}=1.6\;\mathrm{T}$ and ${I}_{\mathrm{P}}=100\;\mathrm{k}\mathrm{A}$ before the application of the ${I}_{\mathrm{c}}$ in the discharge #1089415 as shown in figure 5. The ${I}_{\mathrm{c}}$ was applied at $t=0.32\;\mathrm{s}$ and reaches the flat top $+5\;\mathrm{k}\mathrm{A}$ at $t=0.4\;\mathrm{s}$. The heat flux distribution pattern on the HFS target plate changes obviously at $t=0.39\;\mathrm{s}$, which means that the island divertor configuration has formed at this moment. After $t\sim0.4\mathrm{\ s}$, two strike points were observed, located at $Z\sim 0.06\;\mathrm{m}$ and $Z\sim 0.01\;\mathrm{m}$ on the HFS target plate.

Figure  5.  Temporal evolutions of the contour of (a) floating potential measured ${V}_{\mathrm{f}}$ and radial displacement $Dx$ (white line), (b) contour of ion saturation current ${I}_{\mathrm{s}}$ and ${I}_{\mathrm{c}}$ (white line), and (c) heat flux on the HFS target plate.

DownLoad: Full-Size Img PowerPoint

Figure 6(a) shows the temporal evolution of heat flux profile at a given angel of Φ = 141°, which is calculated from IR camera. Figure 6(b) shows the simulation result ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ on the HFS target plate calculated by HINT. The ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ represents the minimal normalized poloidal flux which can connect to the HFS target plate along magnetic field lines. Additionally, the smaller value of ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ means that this field line connected to the target plate originates from a deeper core plasma, typically with higher temperature and density. Therefore, a larger amount of heat fluxes may direct flow along the field line towards the location with a smaller ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ on the HFS target plate. It can be predicted that the blue region (small ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ region) in figure 6(b) corresponds to the position with larger heat flux on the HFS target plate. The chosen toroidal position corresponds to the black dashed line on the HFS target plate in figure 6(b). There are two strike points are respectively located at $Z\sim0.06\;\mathrm{m}$ and $Z\sim 0.01\;\mathrm{m}$, the location of strike points is in good agreement with the location of the minimum ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$. Figure 6(c) shows the comparison of simulation results with the distribution of heat flux results at $t=0.44\;\mathrm{s}$ and $t=0.54\;\mathrm{s}$. Overall, a decrease in the value of ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ (black line) corresponds to an increase in heat flux (red lines).

Figure  6.  (a) Time evolution of heat flux. Black dotted lines indicate $t=0.44\;\mathrm{s}$ and $t=0.54\;\mathrm{s}$. (b) ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ simulation result of the HFS target plate. (c) Comparison of heat load curve at $t=0.44\;\mathrm{s}$ and $t=0.54\;\mathrm{s}$ (red lines) and the ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ curve at $\phi =141 {\text{°}}$ (black line).

DownLoad: Full-Size Img PowerPoint

However, it is difficult to predict the exact location and accurate value of heat flux from ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ alone. Based on this simulation and the heat flux calculated from the IR camera, the heat flux distribution on the HFS target plate in the J-TEXT tokamak can be preliminarily determined under island divertor configurations.

4.   Effect of SMBI gas fueling on the divertor heat load

In the previous section, the experimental results showed that the heat flux distribution on the HFS target plate with an island divertor configuration has the characteristics of strike point splitting, which reflects the complexity of the boundary magnetic structure compared to the conventional poloidal divertor configuration. In this case, it can be expected that localized gas fueling will have a more complex influence on edge plasma transport and divertor plate heat load. Thus, it is necessary to investigate the impact of SMBI gas fueling on the distribution of heat flux and detachment under a 3D edge magnetic structure.

Figure 7 shows the time evolutions of plasma parameters of discharge #1093618, which is one pure Ohmic discharge with multiple pulses of SMBI on J-TEXT. In this discharge, ${B}_{\mathrm{T}}=1.6\;\mathrm{T}$ and $I_{\mathrm{P}}=80\; \mathrm{k}\mathrm{A}$, which corresponds to ${q}_{a}\sim 4.3$. The central line-averaged density $\overline{{n}_{\mathrm{e}}}$ was maintained around $0.6\times {10}^{19}\;{\mathrm{m}}^{-3}$ before SMBI pulses. The island divertor coil current was applied at $t=0.25\;\mathrm{s}$, and the ${I}_{\mathrm{c}}$ ramped up to maximum value ($-4.5\;\mathrm{k}\mathrm{A}$) at $t=0.27\;\mathrm{s}$. The m/n = 4/1 magnetic island occurred at the boundary of plasma and was opened by the HFS target plate, which leads to the sudden change of the heat flux pattern (figure 7(d)) and ion saturation current ${I}_{\mathrm{s}}$ (figure 7(e)). Four SMBI pulses of hydrogen gas were injected into the plasma with the same backing pressure of 0.18 MPa, and each pulse width is 2 ms, corresponding to an estimated $0.61\times {10}^{19}$ molecules per pulse. The experimental results show that the heat flux on the HFS target plate experiences a significant reduction in delay following each SMBI injection. The precise delay time cannot be determined due to constraints in the time resolution of the IR camera. Meanwhile, there will be a movement of the strike points on the target plate due to horizontal displacement, as shown in figure 7(d). The total radiated power measured by AXUV arrays (figure 7(c)) started to rise after each SMBI pulse and then decreased after reaching saturation. The peak heat flux reduced significantly in the process of AXUV increase and then slowly returned to the initial level before the next SMBI pulse. As shown in figures 7(e) and (f), although the ion saturation current and electron temperature decreased to some extent after the SMBI pulses, they did not decrease to the threshold of detachment. It should be noted that the experiment did not achieve particle detachment, but the power detachment was realized by SMBI gas fueling.

Figure  7.  Time evolutions of (a) $I\mathrm{_P}$ and ${q}_{a}$, (b) $I\mathrm{_c}$ and central line-averaged density $\overline{{n}_{\mathrm{e}}}$, (c) total radiation loss and horizontal displacement of plasma, (d) HFS target heat flux distribution and SMBI injection time, (e) contour of ion saturation current ${I}_{\mathrm{s}}$, and (f) contour of electron temperature ${T}_{\mathrm{e}}$ for discharge #1093618.

DownLoad: Full-Size Img PowerPoint

Under 3D edge magnetic structures, the alteration in local radiation caused by SMBI pulses may be the primary factor contributing to the reduction of peak heat flux. More details about the evolutions of the divertor peak heat flux and delay time of peak radiation in different AXUV channels after each SMBI pulse in this discharge were shown in figure 8(a). The position of the view sight of AXUV arrays is shown in figure 8(b). The moment of the peak value for the most bounded channel of AXUV arrays in one period is defined as $t=0$. The delay time ($\Delta t$) represents the duration between $t=0$ and the moment of peak value for the other chosen channels in this period. There are four periods in discharge #1093618.

Figure  8.  (a) Time evolutions of peak heat flux (dashed lines), the moment of reaching peak value for different AXUV channels (solid lines) during each SMBI pulse period. The black, magenta, blue, and green lines correspond, in turn, to the first period through the fourth period, c–c represents the chord length between the sight line of AXUV and the center of the device. (b) Poincare plot and AXUV arrays (red lines).

DownLoad: Full-Size Img PowerPoint

It can be clearly seen that in all four periods, the ${q}_{\mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}}$ decreased to the lowest value when the AXUV signal at $\sim 171\;\mathrm{m}\mathrm{m}$ reached its peak value. The increase of radiation loss in this region results in a decrease in the power deposited on the HFS target plate. The ${q}_{\mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}}$ decreased sharply from about $0.4\;\mathrm{M}\mathrm{W}/{\mathrm{m}}^{2}$ to $0.23\;\mathrm{M}\mathrm{W}/{\mathrm{m}}^{2}$ within 19 ms in the first period and then started to recover. Although the local radiation growth caused by the SMBI pulse continued to penetrate deeper into the plasma core, the ${q}_{\mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}}$ will not continue to decrease. Obviously, in the later three periods, the minimum value of ${q}_{\mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}}$ is getting lower and lower, from the second period to the fourth period, and the reduction of ${q}_{\mathrm{p}\mathrm{e}\mathrm{a}\mathrm{k}}$ caused by SMBI is successively $~0.26$, $~0.3,$ and $\sim 0.33\;\mathrm{M}\mathrm{W}/{\mathrm{m}}^{2}$. In particular, the penetrated process of radiation front in the first period is a little different with the later three periods; the possible reason for this phenomenon is that the first injection pulse triggered the m/n = 3/1 island which makes the boundary magnetic field more stochastic. However, the exact reason is not clear now. In the region with a chord–center distance of 65 mm towards the plasma core, the AXUV signals were still growing and at a different rate than before, this was caused by the increase of the plasma density.

The m/n = 4/1 magnetic island is generated and opened by the HFS target plate after applying the island divertor coil current, so that the LCFS shifts inward rather than the limiter configuration. The sight of the AXUV signal at $171\;\mathrm{m}\mathrm{m}$ is marked by the red solid line in figure 8(b), almost crossing the LCFS. Each SMBI pulse will induce an increase in radiated power. When the front of radiation growth occurs in the SOL region, the heat flux on the target decreases rapidly, reaching its minimum value when there is maximal radiation near the LCFS.

5.   Effect of amount of SMBI injected molecules on divertor heat load

According to the above experiment, it has been known that when a significant enough radiation occurs near the LCFS, power detachment can be achieved within a short time frame. However, the level of radiated power is related to the quantity of injected molecules by SMBI, which is also a key factor in achieving a significant reduction of power deposition on the HFS target plate. Therefore, the injected hydrogen molecule number of the SMBI was modulated in this experiment by scanning the SMBI pulse width, and the behaviors of heat flux on the HFS target plate were studied.

Figures 9(a)–(d) show four discharges with different SMBI pulse widths on J-TEXT using an island divertor configuration. The main parameters of the target plasma are ${B}_{\mathrm{T}}=1.6\;\mathrm{T},\;{I}_{\mathrm{P}}=95\;\mathrm{k}\mathrm{A},\;{I}_{\mathrm{c}}=+6\;\mathrm{k}\mathrm{A},\;\overline{{n}_{\mathrm{e}}}=\left(1-1.4\right)\times {10}^{19}\;{\mathrm{m}}^{-3}$, corresponding to $q_a\sim3.84$. The backing pressure of the SMBI system was maintained at 0.18 MPa in this experiment. The pulse widths for these four discharges are 2, 1.5, 1, and 0.5 ms, respectively, corresponding to the injected hydrogen molecule numbers of $0.602\times {10}^{19},\;0.494\times {10}^{19}, 0.386\times {10}^{19},\; \mathrm{a}\mathrm{n}\mathrm{d}\;0.278\times {10}^{19}.$

Figure  9.  Time evolutions of heat flux on the HFS target plate with different SMBI pulse widths.

DownLoad: Full-Size Img PowerPoint

It can be clearly seen that there are two distinct splitting strike points on the divertor target plate, $Z\sim0.02\; \mathrm{m} \mathrm{a}\mathrm{n}\mathrm{d}\ Z\sim-0.02\; \mathrm{m}.$ After the formation of the island divertor configuration at $t=0.29\;\mathrm{s}$, the strike points wobble slightly due to the displacement of plasma.

When SMBI pulse width is 2 ms, the heat flux on the two strike points decreases significantly after injection for a delay time, and then begins to recover to the initial state. In particular, the behavior of heat flux on the target plate was highly reproducible during the last three injection periods. However, the reduction of the peak heat flux on the target plate after the SMBI pulse is not significant with the gradual decrease of the pulse width. When the width of the SMBI pulse is 1 ms, the peak heat flux at the secondary strike point ($Z\sim -0.02\;\mathrm{m}$) can still be significantly reduced, but is only slightly decreased at the primary strike point ($Z\sim0.02\mathrm{\ m}$). Furthermore, when the width of the SMBI pulse is 0.5 ms, SMBI gas fueling has a rather weak effect on the peak heat flux. The slight changes in the position and peak heat flux values of the strike points are caused by the displacement of the plasma.

The normalized total AXUV signals of these four discharges were compared. From the experimental results and the previous discussion, it is shown that the radiation will increase to the saturation value after each SMBI and then start to decline. However, when the amount of injected gas is insufficient, it cannot cause sufficient radiation energy dissipation. As shown in figure 10, when the pulse width of SMBI in the experiment is 0.5 ms, the radiation growth in the SOL region is too small to significantly reduce heat load on the HFS target plate.

Figure  10.  Temporal evolutions of total radiation with different widths of SMBI pulses.

DownLoad: Full-Size Img PowerPoint

6.   Discussion and conclusion

An IR camera was first applied on J-TEXT to obtain heat flux distribution on the HFS target plate. In this work, the characteristics of the divertor heat flux distribution with the island divertor configuration have been determined. The heat flux pattern will be divided into multiple strike points when the plasma transfers from the limiter to the island divertor configuration, and as the ${q}_{a}$ changes, the heat flux pattern will change owing to the change of the boundary magnetic structure. The dynamic characteristics of the heat flux distribution on the HFS target plate were studied in this process. Additionally, the measured heat flux pattern is in good agreement with the distribution of ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$, the simulated magnetic footprint on the target plate. However, it should be noted that ${\psi }_{\mathrm{m}\mathrm{i}\mathrm{n}}$ cannot explain the actual transport process at the plasma boundary. Therefore, a more complex transport code, EMC3-ERIENE, will be applied for J-TEXT island divertor case in the next step.

Furthermore, power detachment in a short time by SMBI gas fueling has been realized in the island divertor configuration on J-TEXT, and the effect of penetration depth and injected hydrogen molecule number on the power-detachment phenomenon has been also studied. Firstly, results show that SMBI gas fueling can effectively reduce the peak heat flux on the HFS target plate. When the radiation front induced by the injected fuel particles is located outside the LCFS, the peak heat flux decreases insignificantly. When the main radiation area is near the LCFS, where the magnetic field lines are not cut by the HFS target plate, the peak heat flux reaches its lowest value. After that, the peak heat flux began to recover, even though the radiation in the core was still growing. Secondly, the injected hydrogen molecule number was modulated by changing the SMBI pulse width. It has been observed experimentally that when injected particles cannot induce sufficient growth of radiated power near the LCFS, the peak heat flux on the target does not decrease effectively. The preliminary results from J-TEXT suggest that the radiation near LCFS plays an important role in cooling the SOL plasma and decreasing the divertor heat load. However, experimental evidence was limited by the temporal resolution of the IR camera and the ability to take radiation measurements. Especially during energy detachment, the exact location of radiation belts near LCFS, both toroidally and poloidally, is uncertain.

In fact, the asymmetry radiation distribution and its influence on the edge plasma are complicated for the island divertor configuration. Experiments [29] and simulations [14] from W7-X with an island divertor configuration have suggested that the radiation bands near LCFS can form a cooling layer at the edge with a large surface coverage, and the radiation bands will facilitate the development of a fully detached divertor plasma, even if the divertor plasma remains in an excitation or partial recombination state [30]. It can only trigger partial detachment when the plasma bands are insufficiently wide enough, as evidenced by experiments conducted on W7-AS [31].

Finally, the particle flux on the HFS target plate did not decrease significantly following SMBI pulse as illustrated by ${I}_{\mathrm{s}}$ in figure 7(e), although the peak heat flux shows considerable reduction. One possible reason is that the HFS target plate is the only local target on the high-field side of J-TEXT. Therefore, the divertor region has poor closure and cannot form a high neutral pressure. A complete, toroidally closed target plate may facilitate further research on J-TEXT.

To summarize, heat flux distributions on the HFS target plate in island divertor configurations have been measured by IR camera on J-TEXT. Significant reduction in peak heat load by SMBI gas fueling has been achieved with the island divertor configuration when enough radiation occurs near the LCFS. Applying this mechanism to achieve power detachment in the RMP ELM suppression H-mode plasma or the island divertor plasma is still a great challenge. In addition, it is worth investigating the difference in the effect of injecting impurities and the position of the injection in the future.