Development of the HCN heterodyne collective scattering system on J-TEXT

1.   Introduction

Plasma transport across the magnetic field is one of the principle elements of magnetic confinement fusion. Anomalous transport driven by turbulence in magnetically confined plasma typically far exceeds the level of neoclassical transport and brings a great challenge to plasma confinement [1, 2].

Collective Thomson scattering is a kind of inelastic scattering and is characterized by scattering wave frequency that is different from the incident wave, and the difference frequency ω and difference wave vector k between the scattering wave and the incident wave are consistent with the frequency and wave vector of fluctuation wave in plasma. The relationship between scattering signal and fluctuation wave can be simply described as [3, 4]

where P_s is scattering wave power, P_i is incident wave power and δn_e is electron density fluctuation.

As an effective and non-perturbing method to determine the spatial and temporal distribution of electron density fluctuation in plasma, the collective Thomson scattering diagnostic has been widely applied on various magnetic confinement fusion devices, such as EAST [5], NSTX [6], DIII-D [7], KSTAR [8], etc.

In J-TEXT tokamak, turbulence transport has been studied by diagnostic systems such as the edge Langmuir probe (LP) system [9], Doppler reflectometer [10] and polarimeter-interferometer system [11]. Recently, a dual-HCN collective scattering system has been rebuilt from the original HCN interferometer on J-TEXT [12] to measure density fluctuations of plasma. With the collective scattering system, we can obtain the spectra and spatial distribution of density fluctuation and observe physics processes such as turbulence transition during high-density discharge on J-TEXT.

Typical J-TEXT plasmas are characterized by plasma densities n_e of (1–7)×10¹⁹ m^-3, toroidal field B_t of ~2.0 T, and electron temperature T_e of ~1 keV [13]. Ion temperature gradient (ITG) and trapped electron mode (TEM) turbulence are characterized by k_⊥ρ_i≤1 and the typical ion gyroradius on J-TEXT are 0.1‒0.2 cm. Therefore, the collective scattering system on J-TEXT design aims to measure fluctuations at a wavenumber of 2‒12 cm^-1 with three channels and mainly investigate ion-scale turbulence. A dual-HCN laser has been developed as the probe source of the scattering system. The laser source provides ~50 mW of power at 893 GHz, more details will be introduced in section 3.

2.   Design of the scattering system

A heterodyne collective scattering system has been designed on J-TEXT. The layout of the scattering system is shown in figure 1 and a schematic of the optics is described in figure 2.

Figure  1.  Layout of the J-TEXT scattering system.

DownLoad: Full-Size Img PowerPoint

Figure  2.  Schematic of the J-TEXT scattering system.

DownLoad: Full-Size Img PowerPoint

The dual-HCN laser which is employed as the laser source of the scattering system is installed on a vibration isolation platform. Small parts of the two laser beams are split ahead of the output window of lasers and then combined on the platform as the reference signal. The probe beam is brought to the top of tokamak by an aluminized mirror mounted on the optical breadboard and focused onto the centre of plasma via a methylpentene (TPX) lens. An elongated port of 76 mm wide and 760 mm long is available for the scattering system.

The configuration of the scattering cross-section is shown in figure 3. The launch mirror is mounted on a sliding table and the interaction volume can be located at r=0~10 cm. As shown in figure 2, the scattered beams and local oscillator (LO) beams are combined by beamsplitters and focused into mixers by an aluminized concave mirror. The collection mirrors are positioned to resolve the scattered angles θ_s≈0.6°, 2.1°, 3.7° which are corresponding to wavenumbers of 2 cm^-1, 7 cm^-1, 12 cm^-1 for a source at 337 μm. The measured wavenumber consists of poloidal and radial components, with scales related to the scattering position.

Figure  3.  The configuration of scattering cross-section.

DownLoad: Full-Size Img PowerPoint

The width of the probe beam and scattering beam is ~1 cm and located at the core of scattering volume so the wavenumber resolution of the scattering system is ~2 cm^-1. The spatial resolution perpendicular to the incident wave equals the width ~1 cm and the length of the scattering volume along the incident wave can be estimated by

The length is closed to the chord length for the 2 cm^-1 and 7 cm^-1 channels and ~30 cm for 12 cm^-1 channel. Although poor spatial resolution prevents us from directly resolving turbulence from the interior and edge of the plasma, we can resolve the interior component and edge component in the scattering signal by applying resonant magnetic perturbations (RMP) to change the plasma potential and direction of edge radial electric field. Then we can distinguish density fluctuation from the plasma interior and edge.

The calibration of the scattering system will be carried out later. An acoustic cell will be utilized to scatter the incident wave at an angle determined by the acoustic wavelength and both scattering length and the wavenumber resolution can be calibrated by this method [14]. Amplifier gain and sensitivity of mixers will be calibrated before each round of experiments, then the absolute magnitude of density fluctuations can be determined.

The detectors of the scattering system are Schottky barrier diode mixers (VDI WM-250(~893 GHz)) with an IF amplifier gain of 220@50 Ω. The responsivity of the detector is more than 650 V W^-1 and the IF bandwidth is 20 kHz~10 MHz. With P_i =20 mW, P_LO = 10 mW, scattering length ~30 cm and noise power ~10^-6 W, the minimum detectable density fluctuation resolution is in the order of 10¹⁶ m^-3.

3.   Development of the dual-HCN laser

In consideration of the principle of collective scattering, a laser source with high output power and stable frequency is essential for the scattering system. Hence our first task is to develop a dual HCN laser meeting the requirements of the scattering system.

Figure 4 shows the schematic drawing of the dual-HCN laser. The resonator of the laser consists of a metal grid and a gold-plated mirror. The excitation source of the laser is a high voltage constant current source with 6 kV voltage and 0.5 A current output. The temperature stability of the laser is controlled by oil bath circulator system which consists of an oil jacket and thermostat. The laser system consists of two separately pumped HCN gas lasers which work at a wavelength of 337 μm, and generates an IF of about 1 MHz when there is a ~4 μm cavity length difference. More details of HCN laser on J-TEXT had been well documented in previous article [12] and will not be described here.

Figure  4.  The schematic drawing of dual-HCN laser.

DownLoad: Full-Size Img PowerPoint

Figure 5 shows the output power of dual-HCN laser with different IF. The power and frequency of laser are detected by differential detectors (XZDET-S) made by SINANO, Chinese

Figure  5.  The output power of dual-HCN laser versus intermediate frequency.

DownLoad: Full-Size Img PowerPoint

Academy of Sciences. The amplitude of the IF signal decreases from 1.5 V to 1 V with the beat frequency increasing from 600 kHz to 2 MHz. Since the normal frequency of turbulence on J-TEXT is 20 kHz–1 MHz, IF ~1 MHz is an eligible choice.

The stability of intermediate frequency is essential for the analysis of the scattering spectrum. As shown in figure 6(a), the IF of lasers without quartz in front of the metal mesh varies with an amplitude of ~300 kHz and frequency of ~1 kHz. The 1 kHz variation is caused by the vibration of the metal mesh. To optimize the stability of intermediate frequency, one more quartz plate, with a thickness of ~1 mm, is inserted in front of the metal mesh in the cavity head of the HCN laser to suppress the vibration of metal mesh. With the quartz plate added, the IF stability has been significantly improved as shown in figure 6(b). The 1 kHz vibration has been eliminated and the bandwidth of IF during 0.9 s narrows from 800 kHz to 200 kHz.

Figure  6.  (a) Time-frequency spectrum without quartz plate. (b) Time-frequency spectrum with quartz plate.

DownLoad: Full-Size Img PowerPoint

Figure 7 shows time traces of the output power and IF over 300 min. After ~90 min of warm-up, the output power and frequency of the laser become stable after 2 h. During whole 5 h, the dual-HCN laser can maintain stability without adjustment by controlling the temperature with an oil bath circulator system. The attenuation rate of output power is less than 10% and the drift of IF is within 200 kHz.

Figure  7.  Time traces of the output power and IF over 5 h.

DownLoad: Full-Size Img PowerPoint

4.   Initial experimental result

Scattering signals of three channels have been obtained after developing dual-HCN laser and installing a light path. Considering the small scattering angle ~0.6° and wavenumber resolution ~2 cm^-1, we used the k=2 cm^-1 channel for interferometric measurements before the calibration of the system and have obtained the far-forward collective scattering (FFCS) signal.

The scattering signal obtained from FFCS is presented in figure 8(c). The discharge is performed in Ohmic heating with toroidal field B_t=1.84 T. IF-sized frequency shifts are applied to each window of the spectrogram to eliminate the influence of vibration of IF as shown in figure 8(c). The dash curve shown in figure 8(d) indicates the reference signal on the platform and the solid curve indicates the scattering signal. There are distinct broadband fluctuation signals on both sides of the IF peak. The asymmetry of the spectrum may arise from different levels of density fluctuations above and below the midplane as observed on TEXT tokamak [15].

Figure  8.  (a) The total plasma current, (b) line-average density at r=0 cm, (c) spectrogram of the FFCS signal, (d) spectra of the FFCS signal. Negative (positive) frequency corresponds to the electron (ion) diamagnetic drift direction in the laboratory frame of reference.

DownLoad: Full-Size Img PowerPoint

The scattering signal of k=12 cm^-1 is shown in figure 9. The discharges are performed in Ohmic heating and parameters of two shots are the same with plasma current I_p = 140 kA, toroidal field B_t=2.0 T and line-average density ${\overline{n}}_{\mathrm{e}}$ at r=0 cm ~2.4×10¹⁹ m^-3 in the plateau. The frequency of the scattered signal undergoes significant periodic changes in opposite directions as positive or negative bias is applied. The phenomenon proves that the scattering signals are modulated by the plasma rotation.

Figure  9.  Scattering signal of wavenumber k = 12 cm^-1 changes with positive (negative) bias. (a), (c) Spectrogram of scattering signal. (b), (d) The voltage of bias electrode.

DownLoad: Full-Size Img PowerPoint

Unfortunately, we found that due to the diffraction effects of the optics, the probe light cross-tracks into the optical path of the local oscillator beam, so there is an interference component in the signal. We plan to change the material and size of optics and to optimize the light path to eliminate interferometric components in the signal before the next campaign.

5.   Summary

A collective scattering system has been designed and initially installed on J-TEXT to investigate turbulence transport. A dual-HCN laser that can maintain stable ~5 h has been developed as the source of the scattering system. The bandwidth of IF is obviously narrower after inserting a quartz plate to suppress the vibration of metal mesh. Scattering signals of three channels have been obtained. More results can be expected in the near future after optimizing the system.