Experimental observation of ion heating by mode- converted ion Bernstein waves in tokamak plasmas

We report the experimental observation of ion heating by the mode-converted ion Bernstein waves (MC IBWs) in tokamak plasmas. The MC IBW is created from the fast waves launched from the high-field-side antenna in the HT-7 tokamak in plasmas consisting of deuterium majority, hydrogen minority and 7Li ions. Experimental evidence and numerical simulation show that the interaction between the MC IBW and 7Li ions at the first ion-cyclotron harmonic resonance of 7Li (i.e. ω = 2 Li-7) is the main mechanism for radio-frequency power deposition. By comparing with previous experiments of direct-launch IBW flow drive on tokamaks and existing theories, we hypothesize that this MC IBW and 7Li interaction also leads to the observed flow drive effect. (Some figures may appear in colour only in the online journal) Radio-frequency (rf) power in the ion-cyclotron range of frequencies (ICRFs) has been used to heat plasmas and also drive plasma flows in tokamaks. The ICRF system on ITER can help heat ions [1], but the direct ion heating will diminish at high electron temperature [2], thus finding a way to directly heat fusion ions is of great interest. An ion heating scenario based on cyclotron damping of the mode-converted ion Bernstein waves (MC IBWs) by tritium minority at the first ion-cyclotron harmonic of tritium ions (ω = 2 T) has been proposed for JET and IGNITOR [3, 30], but such a scenario has not been experimentally verified. Second harmonic minority heating by ICRF power has been demonstrated in previous theoretical [4] and experimental studies [5, 6]. As pointed out in [7], on the high-field side (HFS) of the MC layer, the FW can be converted to an IBW. On the low-field side (LFS) of this layer, the FW can be converted to an electromagnetic ion-cyclotron wave (ICW). Ion heating in the MC regime has 5 Author to whom any correspondence should be addressed. been demonstrated in previous experiments [8–11]. MC waves have been shown to drive toroidal and poloidal flows in D (3He) plasmas on Alcator C-Mod [8] and JET [10, 11] in a narrow parameter space, where the interaction between the MC ICWs and 3He ions is thought to be the key for flow drive. The nonlinear interaction of slow waves with ions has been proposed as a mechanism for the momentum transfer in these ICRF flow drive experiments [12–14]. Here we report the experimental demonstration of direct ion heating and flow drive by the MC IBW in tokamak plasmas. HT-7 is a circular superconducting tokamak with major radius R = 1.22 m and minor radius a = 0.27 m [15]. Fast magnetosonic waves (fast waves, or FW) at 27 MHz are launched externally from a two-strap antenna on the HFS. The Faraday screen of the antenna is approximately 0.5 cm to the plasma edge. The antenna current straps are 112 cm long and 15 cm wide, and the gap between the straps is 70 cm. At [0, π ] dipole phasing, the launched fast waves 0029-5515/12/082003+04$33.00 1 © 2012 IAEA, Vienna Printed in the UK & the USA Nucl. Fusion 52 (2012) 082003 Letter Figure 1. Experimental set-up on the HT-7 tokamak (R = 1.22 m and a = 0.27 m). The ICRF antenna is on the HFS. Bt0 = 2 T, frf = 27 MHz, nH/ne = 15% and nLi-7/ne = 1%. The grey region indicates the location where IBWs are excited and damped by 7Li ions and electrons. have peak power at toroidal numbers nφ = ±6, corresponding to k‖ ∼ ±5 m−1. Lithium powder (mainly 7Li) has been used extensively to condition the walls of the tokamak to control impurities and also reduce the hydrogen (H) level [16]. Spectroscopic measurements indicated that the typical level of 7Li ions in the plasmas was nLi/ne = 0.5%–1.0%. As a result, the typical plasma composition in the experiment was a mixture of deuterium (D), H and 7Li. Figure 1 shows the cross section of HT-7 together with the location of the ICRF antenna, H cyclotron layers, MC layer (D-H hybrid layer) and the first harmonic 7Li IC layer in the experiment for Bt0 = 2.0 T and nH/ne = 15%. Plasma rotation correlated with the launched rf power has been observed in these D-H-7Li plasmas. The flow (rotation) velocity Vθ is inferred from the Doppler shift in the turbulence spectra measured by a collective Thomson scattering system using a CW CO2 laser [17]. As shown in figure 2, we plot the traces of the rotation with that from the ohmic period subtracted, Vθ , in two regions, 0 < r/a < 0.4 and 0.3 < r/a < 0.7. A significant change in Vθ is clearly shown following the application of the ICRF power. The opposite directions of Vθ in the two spatial regions suggest that the application of ICRF power creates a dipolar structure in the poloidal plasma rotation. The magnitude of the rotation is also larger at a higher rf power. The two-dimensional full wave code TORIC [18, 19] has been used to model the ICRF physics in these plasmas and to determine the power absorption via different mechanisms. In TORIC simulations, the impurity lithium ion was taken into account. Energy exchange between plasma particles was not taken into account. The TORIC simulation results [17] show that the interaction between the MC IBW and 7Li ions at the first ion-cyclotron harmonic resonance of 7Li (i.e. ω = 2 Li-7) is the main mechanism for rf power deposition. To illustrate the detailed wave–particle physics, in figure 3 we plot the 2D contours of the electric field from the TORIC simulation with nφ = +6. The E field is decomposed to E+, E− and Ez, i.e. left-handed polarization, right-handed Figure 2. Flow observed corresponding to the ICRF power. (a) ICRF power (source power) was applied in HT7.The red line is 600 kW, the blue line is 800 kW; (b) Vθ in 0 < r/a < 0.4 versus time; (c) Vθ in 0.3 < r/a < 0.7 versus time. polarization, and parallel to B. In figure 3(a), over the large structure of fast waves, the fine structure of the shorter wavelength MC waves is shown in E− contours. Whereas the fast wave has near to zero Ez, the MC IBW is clearly shown in figure 3(b) on the HFS of the MC layer and the MC ICW on the LFS of the MC layer. While propagating towards the HFS, the amplitude of the MC IBW has a steep drop at the ω = 2 Li-7 layer, indicating a strong local absorption. The power deposition is shown in figure 4. The total power to the electrons is dominated by the MC IBW and ICW as shown in figure 4(a). In figure 4(b), the power to 7Li via the interaction of the MC IBW and 7Li ions is shown to be strongly peaked and localized near the resonance layer. The total power to 7Li ions is similar to the total power to electrons. To experimentally verify the MC IBW and 7Li ions interaction is not straightforward. In the following, we show this from the response of the neutron rate versus RF power and versus a magnetic field scan. In figure 5, traces of two discharges at two different ICRF power levels are compared. The plasmas have Bt0 = 2.0 T, XMC = 2 cm and the 7Li first harmonic layer at XLi-7 = −3 cm, where X is defined as the distance to the magnetic axis, X = R − R0. The fusion neutron rate on both plasmas increases immediately (in a time scale of less than 20 ms) after the ICRF power is applied. Because the plasma density actually decreases slightly due to the ICRF density pump-out effect, the increase in neutron rate indicates either an increase in deuterium ion temperature (TD) or the generation of energetic D ions. In these plasmas at Te0 ∼ 1.5 keV, the electron–ion heat exchange time is about 300 ms [20], an order of magnitude longer than the observed neutron rate rising time; therefore, the direct electron heating, e.g. from the MC IBW and ICW Landau damping, cannot explain the fast rise of the neutron rate. Harmonic D heating and the collisional exchange between the hot H ions (due to H minority absorption) and bulk D ions could cause fast neutron rise, but this is shown not to be the case. In figure 6, we show discharges at three