Stable Deuterium-Tritium plasmas with improved confinement in the presence of energetic-ion instabilities

Abstract

Providing stable and clean energy sources is a necessity for the increasing demands of humanity. Energy produced by Deuterium (D) and Tritium (T) fusion reactions, in particular in tokamaks, is a promising path towards that goal. However, there is little experience with plasmas formed by D-T mixtures, since most of the experiments are currently performed in pure D. After more than 20 years, the Joint European Torus (JET) has carried out new D-T experiments with the aim of exploring some of the unique characteristics expected in future fusion reactors, such as the presence of highly energetic ions in low plasma rotation conditions. A new stable, high confinement and impurity-free D-T regime, with reduction of energy losses with respect to D, has been found. Multiscale physics mechanisms critically determine the thermal confinement. These crucial achievements importantly contribute to the establishment of fusion energy generation as an alternative to fossil fuels.

Fig. 1. Main characteristics of the D-T discharge #99896.

A Time evolution of discharge #99896, with toroidal current I_p = 1.9 MA, magnetic field B_T = 2.75 T, and q₉₅ = 4.5, heated mainly with ICRF power, P_ICRF = 4.5 MW. The NBI power, P_NBI ∼ 3.5 MW, was also injected with deuterium beams, before 9 s, and tritium beams, after 9 s. The radiated power, P_rad represents 60% of the total input power. The power produced by D-T fusion reactions obtained reaches a maximum of P_fus ∼ 0.5 MW. B Time evolution of edge fluctuations as obtained from the B_eII line emission from the inner divertor. C Time evolution of β_N, defined as β_N = βaB_T/I_p[%] with β the ratio between magnetic and thermal pressure and a the plasma minor radius, H₉₈(y, 2) and $f_{Gr}={\overline{n}}_e/n_{Gr}$ the Greenwald fraction with ${\overline{n}}_e$ the average density and n_Gr the Greenwald density defined as n_Gr = I_p/πa². D Accumulated NBI and ICRF radial power deposition for the discharge #99896. E Time evolution of magnetic perturbations detected by the Mirnov coils.

Fig. 2. Comparison between the D-T discharge #99896 and the D counterpart #100871.

A Comparison between T_i and T_e for the D-T discharge #99896 and the D counterpart #100871. T_i is measured by the charge-exchange technique on impurity ions. T_e is obtained by means of LIDAR and high-resolution Thomsom scattering (HRTS). An average over 8.5 s-8.7 s is performed. ρ is defined as the square root of the normalized toroidal magnetic flux. Shaded error bars represent the standard deviation of the time-averaged signals and the systematic diagnostic uncertainties. B Comparison between electron density, n_e, for the D-T discharge #99896 and the D counterpart #100871. n_e is measured with HRTS. Shaded error bars represent standard deviation of the time-averaged signals and the systematic diagnostic uncertainties. C Comparison between χ_i and χ_e obtained by power balance analysis for the D-T discharge #99896 and the D counterpart #100871. Shaded error bars represent standard deviation.

Fig. 3. Core plasma fluctuations characteristics.

A Bicoherence analysis of the perturbations found in discharge #99896. The analysis is performed at t = 7.7 s. B Logarithmic power of the density fluctuations as obtained from reflectometry at major radius R ∼ 3.36 m, ρ ∼ 0.35, and t = 8.4 s, for the D-T discharge #99896 with only NBI heating or with full NBI and ICRF heating and comparison to the pure D discharge #100871. C Electrostatic potential, Φ, fluctuations obtained for the D-T discharge #99896 by the global code FAR3D when considering two species of energetic ions, H and D, accelerated by the ICRF power. The yellow circle represents the q = 1 surface. Inside q = 1, a n = 1 perturbation is obtained which is identified as a fishbone instability. Outside q = 1, TAEs are obtained.

Fig. 4. Alpha particle transport and losses in D-T.

A Alpha particle density profile as calculated with TRANSP assuming no alpha particle transport (n_α,eq) and comparison to the profile obtained from the FAR3D code after the full development of the fishbone instability (n_α,final). B Alpha particle loss frequency spectrum obtained by using the fast ion loss detector (FILD) with channels that are receptive to 3.5 MeV alpha particles.

Fig. 5. Zonal flow generation by energetic particle instabilities.

2D pattern of n = 0, m = 0 structures of zonal poloidal flows, V_th(0, 0), for the TAE and fishbone instabilities. V_th(0, 0) is defined as V_th(0, 0) = E_r(0, 0) × B_T with E_r(0, 0) the n, m = 0, 0 component of the perturbed radial electric field. The dependence of the zonal flow intensity on the perturbation strength is studied by scanning the energetic ion equivalent temperature (T_f) using two values, T_f = 1 MeV and T_f = 500 keV in the FAR3D code. Zonal flow generation increases with increasing perturbation intensity for both TAEs (A, B) and fishbones (D, E). The radial extension of zonal flow activity coincides with the extension of the two perturbations. The zonal flow activity in D-T is compared to the one in pure D (C and F) by artificially replacing T by D in FAR3D. The intensity of the zonal flow is lower in D than in D-T.

Fig. 6. Gyrokinetic analysis of core turbulence reduction in D-T.

A Growth rate, γ, and B frequency, ω, spectrum obtained from linear simulations with the CGYRO code. k_y is the binormal wavenumber normalized to the proton sound gyroradius ρ_s. C Energy flux obtained from gyrokinetic simulations performed with the CGYRO code for discharge #99896 at ρ = 0.31. The simulations are performed including and excluding the energetic ion component. Values of the ion thermal energy flux deduced from power balance in TRANSP (black horizontal dashed line) are only obtained when the energetic ion component is included in the simulations as a separate species. The total thermal ion energy flux obtained in D-T including energetic ions is compared to the one obtained assuming that all the thermal ions are D while keeping the rest of the parameters fixed. The energy flux in pure D is significantly higher than in D-T (a zoom of those bars is displayed in the inset on the top right). Simulations with $a/L_{T_{FI}}=0$ and T_FI/T_e = 5.25 are performed to linearly stabilize the low k_y energetic ion mode while keeping the energetic ions in the simulations. Fluxes obtained with stabilized mode and energetic ions cannot reproduce the experimental fluxes.

Fig. 7. Pedestal formation in D-T and comparison to D.

A Comparison between edge n_e for the D-T discharge #99896 and D discharge #100871. ρ_pol is defined as the normalized poloidal flux. B Comparison between edge T_e for the D-T discharge #99896 and D discharge #100871. The vertical dashed line represents the location of the plasma separatrix. The profiles are obtained from HRTS averaged in the time window 8.5 s-8.9 s for #99896 and 8.4 s-8.8 s for #100871. The evaluation of the error bars in panels A and B is done by deriving the expected signal levels at a given temperature and by calculating the standard deviation based on the photoelectron statistics, the plasma background light variation, and the detector noise.

Fig. 8. Pedestal characteristics in D-T.

Comparison between discharge #99896 and discharge #99502, with P_NBI = 12.5 MW, in H-mode with type-I ELMs, and #99776, in L-mode, with P_NBI = 5.4 MW and P_ICRF = 3.3 MW both obtained at I_p=2.5 MA, B_T=3.7 T and q₉₅=4.5. A Edge n_e. B Edge T_e. The profiles are obtained from HRTS averaged in the time window 8.5 s-8.9 s for #99896, 7.3 s-7.6 s for #99502, and 8.6 s-8.9 s for #99776. The evaluation of the error bars in panels A and B is done by deriving the expected signal levels at a given temperature and by calculating the standard deviation based on the photoelectron statistics, the plasma background light variation, and the detector noise. C Comparison of the Be_II line emission from the inner divertor for the same discharges. The vertical dashed line represents the location of the plasma separatrix.

Fig. 9. Scenario performance in D-T.

Comparison between discharge #99896, obtained at P_NBI = 3.5 MW, P_ICRF = 4.5 MW, I_p = 1.9 MA, B_T = 2.75 T, and the D-T discharge #99501, obtained at P_NBI = 9.5 MW, I_p = 2.5 MA, B_T = 3.7 T. A n_e. B T_e. C T_i. Shaded error bars represent the standard deviation of the time-averaged signals and the systematic diagnostic uncertainties in panels A–C. D χ_i/χ_GB. χ_GB is the GyroBohm diffusivity defined as ${\chi }_{GB}={T_e}^{3/2}{m_p}^{1/2}/\left(e^2{B_T}^{2}a\right)$ with m_p the proton mass, e the electron charge, and a the plasma minor radius. Shaded error bars represent standard deviation.