1 FTP/3-5

Lithium-Metal Infused Trenches: A New Way to Remove Divertor Heat Flux

W. Xu1, D. Curreli1, D. Andruczyk1, T. Mui1, D. Lee1 and D. N. Ruzic1

1Center for Plasma-Material Interactions, Department of Nuclear, Plasma and Radiological Engineering, University of Illinois at Urbana-Champaign, 104 S. Wright St., Urbana IL 61801, USA

E-mail contact of main author: druzic@illinois.edu Abstract. An experimental characterization of the LiMIT (Lithium/Metal Infused Trenches) device, an innovative system under development at the University of Illinois for heat removal from the divertor of magnetic fusion machines, is presented. The device is designed to continuously present a fresh liquid lithium surface in the divertor, and withstand heat fluxes of up to 10 MW/m2 without significant evaporation. The velocity of liquid lithium flowing within the trenches has been measured by means of surface particle tracing with the help of a fast frame camera. The flow velocity has been characterized for two conditions of low and high magnetic fields, to show the influence of MHD damping at high fields. The results are in agreement with the prediction of a simplified analytical solution of the flow within the trench. The two cases of low and high magnetization are also compared with IR imaging, giving consistent proof of the decrease of the flow velocity at high magnetic fields.

1. Introduction

As we come closer to realizing fusion as a viable power source, many machines which have made significant progress in plasma heating and power find that removing heat from the divertor region has become a challenging problem. Peak heat fluxes up to 20 MW/m2 are typical these days for machines and if there is an ELM event or disruption then it can be even higher. Solid plasma facing components (PFC) materials such as carbon and tungsten suffer from permanent damage due to huge thermal stresses. Flowing liquid metals have been proposed as possible alternatives, since they offer a number of attractive advantages with respect to solid first walls. Traditional solid materials suffer of a number of problems like sputtering, fuzz growth, helium retention, bubble formation, dust, debris release, etc. Especially for high-Z materials like tungsten, the contamination caused to the plasma is critical, being source of high radiative bremsstrahlung losses. Several options have been considered, such as liquid lithium, Sn-Li compounds, gallium and etc [1][2]. Among all the possible candidates, lithium has drawn a lot of interest recently, acting as a beneficial and active surface with regard to the plasma. Significant improvements of the plasma performance have been reported, both on tokamaks (TFTR [3], FTU [4], NSTX, etc.), stellarators (TJ-II), and reversed field pinches (RFX). Lithium offers the possibility to have a low recycling wall [5] [6], with improvements such as higher confinement time, higher plasma temperature and density, ELMs suppression (for tokamaks), and lower Z-effective. At the University of Illinois at Urbana-Champaign (UIUC), a liquid metal divertor system called LiMIT able to withstand heat fluxes up to 10 MW/m2 has been developed. LiMIT is scheduled to be used in the Chinese HT-7 superconducting tokamak. LiMIT, Lithium-Metal Infused Trenches [7] uses thermoelectric magnetohydrodynamics forces [8] to drive liquid lithium inside many narrow parallel trenches made of stainless steel (or molybdenum) with

2 FTP/3-5

millimeter thick gaps between them. The feasibility of the TEMHD drive was proved for the first time at the University of Illinois [9], after inception tests on lithium walls done on the CDX-U spherical tokamak [10]. The trenches of LiMIT are oriented along the poloidal direction of the torus, so that they are transversally crossed by the toroidal magnetic field. Once a temperature gradient is established between the top and bottom of the trench from the plasma hitting the lithium surface, a thermoelectric current is generated in the direction of the temperature gradient. The thermoelectric current drives the liquid inside the trench thanks to the JxB force originated from the interaction of the thermoelectric current with the toroidal field. The flow is self-driven by the instantaneous thermal gradient, hence resulting self-adaptive against the heat flow received from the plasma. In this paper we present the results from an experimental characterization of LiMIT, done on the proof-of-principle experiment "SLIDE" at UIUC’s CPMI labs. The LiMIT tile is heated via a linear electron beam which has a highest heat flux of 10.0+/-1.1 MW/m2. The flow velocity and the heat transfer capability have been characterized for two conditions of high and low magnetic fields. A fast frame camera has been used to trace small impurity particles released on the surface of the flowing lithium as tracers. IR camera has been used to characterize the temperature vs. time during heat shocks. The measurements are in agreement with the predictions of the simplified analytical solution obtained in [7]. The same analytical solution shows that in the presence of a strong magnetic field at tokamak-level, the surface temperature and flow velocity are proportional to the square root of the surface heat flux. This has large heat transfer potential for future applications in fusion devices.

2. Flow characterization

In previous works [7], the thermoelectric-driven liquid lithium flow used to be investigated at a comparably low magnetic field (< 0.059 Tesla). The surface velocity was measured by the particle tracking method with a normal digital camera, which cannot capture the detailed motion of the moving particle because of its low frame rate of 25Hz. Another problem comes from the surface impurity layer. Lithium is highly reactive material and once it is loaded into the tray the impurity layer starts to form on the surface even though it is in vacuum. This impurity layer is not thick but it can cover the whole lithium surface. Even when the lithium moves in the trench the impurity layer does not necessarily move with the flow. Only when the electron beam power is very high some of the impurity that is directly hit can be burned away.

With the fast frame camera the surface velocity of the lithium flow can be measured in detail. A stainless steel brush is suspended above the lithium surface. Although the brush cannot completely remove the impurity layer, it can still break the whole impurity layer into small pieces. During experiment the brush is first used to disturb the surface and then the electron beam is turned on. Once the lithium starts to flow small impurity pieces can move with the flow and their motion recorded by the fast frame camera.

2.1. General description of the fast frame camera imaging

The velocity of the flowing liquid lithium in the trench is measured by the fast frame camera through a glass window which faces the surface of the tray. The optical axis of the camera is actually not perpendicular to the tray surface. The angle between the optical axis and the surface normal is 55 degree. This gives us an inclined top view of the tray surface instead of a vertical top view.

3 FTP/3-5

FIG. 1. A frame image of the lithium tray surface taken by fast frame camera

Figure 1 is a typical frame of the fast frame video, facing the center of the tray. Since the diameter of the window is about 3.5 cm which is comparably smaller than the size of the tray, only part of the tray surface is captured. In figure 1 the width of the captured part is about 6 cm and the height is about 5 cm in spite of the size of the tray is about 9 cm by 9 cm. The distance from the camera to the lithium surface is about 20 cm and a 50mm micro lens is used to focus, which results in a poor focus on the bottom part of the image because there the inclined surface causes different distances from different part of the tray to the lens and we can hardly reach a good focus for the whole image. Still 5 to 6 trenches can be viewed in focus by the system. Another problem is some part of the image is saturated because of the light reflection from the electron beam filament. In Fig.1 it shows in the bottom right reflections from the shiny liquid lithium surface. In this figure the liquid lithium is found to flow from right to left. The image is divided into few regions. Region A is the center of the tray directly hit by the electron beam. The width of this region is about 1 cm. Some dim straight lines can be found in region A and those are the light emission when electrons hit the lithium surface. The right side of region A is the upstream side and the left side of region A is the downstream side of the lithium flow. Some part of the upstream side has light reflection from the electron beam filament while the rest looks dark. The shiny surface of pure lithium can hardly be seen without any light reflection from the surface of liquid lithium. However, when an impurity particle flows across the dark region it looks white because its rough surface can diffusely reflects the light. For example in region B a small particle is flowing along the trench. The left side of region A is the downstream side. Since the thermoelectric current has a positive correlation with the temperature gradient and obviously the direct heating by the electron beam in region A generates a high temperature gradient at the center, after the liquid lithium passes region A it will get accelerated by the Lorentz force and enter the downstream region. From the picture we can see some lithium pile-up in region C and the reason will be discussed later.

4 FTP/3-5

2.2. Velocity measurement through sequence frames

The velocity of the moving particles can be measured through a few sequence images and this velocity is assumed to be the surface velocity of the liquid lithium in the same region. In the experiment the flow on the upstream side seems laminar and the flow on the downstream sides is turbulent. Because of the mass flux conservation we only measure the velocity of the upstream side and use that as an estimate of the average velocity within the whole trench. It is also observed that different trenches have different velocity values due to the non-uniformity of the electron beam heat flux and impurity accumulation. As a result a 12.5 second video is taken for each experimental condition, based on which the velocity values from different trenches are measured and averaged. The standard deviation is used as the error bar.

FIG. 2. Sequence frames (totally 12 frames) of a moving impurity particle on liquid lithium surface

Figure 2 presents an example of measuring the moving particle through sequence frames, which actually includes the frame of figure 1. Due to the size of the image only region B in each frame is shown in figure 2. In figure 2 the moving particle looks obvious with the dark background. Frames 1 to 6 are on the left side and frames 7 to 12 are on the right side. The positions of the particle on each frame look like a straight line which indicates that the velocity of the particle is stable during these 12 frames. The bright horizontal lines on each frame are the light reflection from the edges of the trench, so we can tell that the particle flows almost at the center of the trench. The software ImageJ can easily measure the length of the moving path on the picture with the unit of pixels. The length of the path is scaled to the width of the trench (2 mm) to get the real length during the 12 frames’ time. In the experiment the frame rate is set to be 400 frames per second (fps) so that the time interval between each frame is known. With these values the average velocity of the particle during the 12 frames’ time can be calculated. For each experimental condition about 10 groups of frames are chosen to calculate the velocity.

3. Flow velocity results

A one dimensional model has been developed [7] to calculate the flow velocity and surface temperature as a function of the magnetic field or the heat flux. For this TEMHD driven trench flow the velocity increases when the heating increases which is a self-adaptive system and this is the advantage of the LiMIT concept. On the other hand the velocity decreases

5 FTP/3-5

when the magnetic field increases which is the disadvantage of this design. Then the feasibility of LiMIT concept for real fusion device lies in if the velocity drops too much when the magnetic field becomes strong because the convection heat transfer is strongly affected by the flow velocity. To find out the performance of LiMIT concept in higher magnetic field eight car batteries are connected to the magnet coil to generate a strong magnetic field that has a 0.19 T magnetic induction on the vertical direction of the trench and this value is three times higher than the field we used before (up to 0.059 T). Figure 3.a is the velocity measured in different magnetic fields compared to the predicted value. The velocity at B=0.059 T was measured by the digital camera and the velocity at B=0.19T is measured by the fast frame camera. The solid curve is the velocity predicted by the 1D model. The heat flux from the electron beam remains the same. Both measurements match the predicted curve which proves that the velocity decreases when the magnetic field increases. However from the trend of the curve we can see that the liquid lithium still has a velocity when B=0.19T which is about 62% (0.22 m/s to 0.13 m/s) of the velocity when B=0.0589T. This also predicts that the velocity will be 0.053 m/s when B=1T and 0.035m/s when B=2T, which means that lithium can flow across half of the trench length within 1 second even in the magnetic field as strong as the fusion reactor magnetic field.

FIG. 3. (a) Liquid lithium flow velocity in different magnetic fields. (The left figure) FIG.3. (b) Liquid lithium flow velocity changes with the incoming heat flux (The right figure)

Figure 3.b presents how the velocity changes with the heat flux when the magnetic field keeps constant. In this figure the measured values match the predicted curve. When the heat flux is increased the temperature gradient will also increase which leads to a higher flow velocity to offset the heat flux increase. This leads to a higher convection heat transfer rate and proves the existence of the self-adaptive effect. Such flow effectively transfers the heat from a narrow area to the whole volume of the fluid and even when some transient event happens the explosively growing heat flux can accelerate the flow velocity and the extra deposited heat will eventually be removed.

3.1.Other observation with the fast frame camera

In the experiment the electron beam is turned on first and then the magnetic field is turned on. When the high magnetic field starts it is discovered that lithium starts to accumulate on the

6 FTP/3-5

downstream side and the velocity of the flow starts to increase. Until about 0.1 second later the velocity becomes stable and the lithium pile-up stops growing. This process is shown by figure 5.

When the lithium starts to flow the velocity is still low and the center temperature gradient builds up quickly. Due to the self-adaptive effect the driven force keeps rising and accelerates the flow to the balance point. The 0.1 second acceleration time actually reveals the response time of the flow to the variation of the heat flux.

FIG.4. Change of the lithium surface when the flow starts. Pile-up and speed-up are observed in this 0.1 second period.

Once the liquid lithium starts to flow the mass keeps moving from the upstream side to the downstream side. This process lowers the fluid height on the upstream side and raises the fluid height on the upstream side, which forms the pile-up. The height difference creates a pressure difference between both sides and drives the back flow under the trench from left to right. The pile-up may become a more significant problem in tesla-level fields because the MHD damping becomes large in the backflow channel and requires a higher pile-up to provide enough pressure difference across the backflow channel. To combat this an active heater has been added to the sides and bottom of the newer versions of LiMIT providing a TEMHD driven force inside the backflow channel and help to speed up the backflow. One of these newer LIMIT trays is installed in the HT-7 tokamak in Hefei, China. Another interesting phenomenon is that the flow velocities on the upstream side and the downstream side are different. After the heating region the flow velocity is much higher than the upstream flow. The lithium is accelerated after the heating region. From mass flow conservation we expect a strong turbulence below the surface on the downstream side, as observed in the experiment. Althought the flow on the upstream side is laminar it immediately becomes turbulent passing the heating region. This turbulence actually benefits the convection heat transfer within the fluid.

4. Heat transfer characterization

The surface temperature of the liquid lithium is measured by IR camera in both low and high magnetic fields as an indication of its heat transfer ability. The IR camera is installed at the same position as the fast frame camera with a ZnSe window instead. The measuring and calibration method for the IR camera has been discussed in previous work [7] [11]. The

7 FTP/3-5

increase of absolute temperature of the liquid lithium surface heated by a 4 MW/m2 heat flux for about 10 seconds is presented in figure 5. The temperature values of the lithium surface before and after the heating are both measured and the subtraction is the absolute temperature increase. The black region is caused by the saturation during the measurement and the value there cannot be read. Fig.5.a reveals the temperature increase in the low field. In low field the flow velocity is high which brings the heat away immediately. The liquid lithium flows from left to right and the downstream side has a higher temperature than the upstream side. The temperature at the direct heating region is slightly higher than the downstream side which indicates good heat convection ability.

FIG. 5. The temperature increase of the liquid lithium surface after it has

been heated by the 4 MW/m2 heat flux. The magnetic field in the left figure (Fig.5.a) is 0.033 T and that in the right figure (Fig.5.b) is 0.19 T

In Fig.5.b the temperature distribution in high field is different from that in the left figure. The uneven temperature distribution still exists which means the flowing liquid lithium can still transfer the heat to the downstream side. But the temperature increase is higher than the low field result and this is in agreement with the lower velocity measurement in the high field. The direct heating region can be clearly identified in this figure because the damped velocity lowers the heat transfer ability.

5. Conclusion

The key to assessing the feasibility of using the LiMIT concept in fusion reactors is evaluating its performance in the strong magnetic field. Although temporarily a large scale tokamak level field cannot be built in our system, the LiMIT concept has been tested in a comparably strong field. Both fast frame camera and IR camera are used to characterize the flow and heat transfer ability of this type of design. The experiment results show lowered velocities and heat transfer abilities at high field, but these changes of the flow are still acceptable and the experiment results match the theoretical prediction well. As predicted from a simplified 1D model, the flow velocity and the heat transfer ability are still enough in a tokamak-level magnetic field, which indicates that the LiMIT concept could be a potential choice as the plasma facing component in the divertor of magnetic fusion reactors.

6. Acknowledgements

The authors would like to thank Dr. Dennis Youchison from Sandia National Lab, Guizhong Zuo from Institute of Plasma Physics, Chinese Academy of Science and Dr. Mike Jaworski

8 FTP/3-5

from Princeton Plasma Physics Lab for their generous help. This work is supported under DOE contracts DE-FG02-99ER54515.

References

[1] M. Greenwald, R. Callis, D. Gates, B. Dorland, J. Harris, R. Linford, M. Mauel, K. McCarthy, D. Meade, F. Najmabadi, B. Nevins, J. Sarff, m. Ulrickson, M. Zarnstorff and S. Zinkle, Tech. Rep., Fusion Energy Science Advisory Committee (2006)

[2] M.A. Abdou, The APEX TEAM, A. Ying, N. Morley, K. Gule, S. Smolentsev, M. Kotschenreuther, S. malang, S. Zinkle, T. Rognlien, P. Fogarty, B. Nelson, R. Nygren, K. McCarthy, M.Z. Youssef, N. Ghoniem, D. Sze, C. Wong, M. Sawan, H. Khater, R. Woolley, R. Mattas, R. Moir, S. Sharafat, J. Brooks, A. Hassanein, D. Petti, M. Tillack, M. Ulrickson, T. Uchimoto, Fusion Eng. Des. 54 (2001) 181

[3] D.K. Mansfield, D.W. Johnson, B. Grek, H.W. Kugel, M.G. Bell, R.E. Bell, R.V. Budny, C.E. Bush, E.D. Fredrickson, K.W. Hill, D.L. Jassby, R.J. Maqueda, H.K. Park, A.T. Ramsey, E.J. Synakowski, G. Taylor, G.A. Wurden, Nuclear Fusion 41 (2001) 1823

[4] M. L. Apicella, G. Apruzzese, G. Mazzitelli, V. Pericoli Ridolfini, A. G. Alekseyev, V. B. Lazarev, S. V. Mirnov and R. Zagorski, Plasma Phys. Control. Fusion 54 (2012) 035001

[5] M. Nieto, D.N. Ruzic, W. Olczak, R. Stubbers, J. Nucl. Mater., 350 (2006) 101-112.

[6] D.N. Ruzic, M. Nieto, J.P. Allain, M.D. Coventry, J. Nucl. Mater. 313-316 (2003) 646-650

[7] D.N. Ruzic, W. Xu, D. Andruczyk and M.A. Jaworski, Nuclear Fusion, 51 (2011) 102002

[8] J.A. Shercliff, J.Fluid Mech, 91 (1979) 231 [9] M.A. Jaworski, T.K. Gray, M. Antonelli, J.J. Kim, C.Y. Lau, M.B. Lee, M.J. Neumann,

W. Xu and D.N. Ruzic, Physics Rev. Lets. 104 (2010) 094503 [10] R. Kaita, R. Majeski, T. Gray, H. Kugel, D. Mansfield, J. Spaleta, J. Timberlake, L.

Zakharov, R. Doerner, T. Lynch, R. Maingi and V. Soukhanovskii, Phys. Plasma 14 (2007) 056111

[11] W. Xu, V. Surla, M.A. Jaworski, M. Lee, T. Mui, M.J. Neumann and D.N. Ruzic, Journal of Nuclear Materials, 415 (2011) 981