2026 年 21 巻 論文ID: 1202033
A non-evaporable getter (NEG) pump was experimentally installed in the QUEST spherical tokamak to enhance particle exhaust capability during long-pulse, high wall temperature operation (473 K). Activation and pumping performance tests were conducted using hydrogen gas injection and quadrupole mass spectrometry. During activation at 773 K, hydrogen desorption was clearly observed, whereas impurity species (mass number 18, 28, 32, 44) desorption remained strongly suppressed. Pumping tests at 473 K demonstrated hydrogen-selective exhaust with an effective pumping speed of approximately 150 L/s. These results demonstrate that NEG pumps can provide a controllable and impurity-compatible particle exhaust method in fusion-relevant environments. In particular, the suppression of impurity release during activation suggests the possibility of in situ regeneration without significant degradation of plasma conditions due to impurity contamination, enabling continuous operation without isolation. This operational flexibility highlights the potential of NEG pumps as assist tools for particle control in steady-state plasma operation.
QUEST (Q-shu University Experiment with Steady-State Spherical Tokamak) performs long-duration plasma discharges of up to 6 hours with a high wall temperature of 473 K [1]. Under such conditions, a wall saturation phenomenon has been observed, where gas puff feedback control becomes ineffective. Enhancement of particle exhaust capability is therefore considered a key issue for stable steady-state operation. Non-evaporable getter (NEG) pumps are being considered for fusion research owing to their high pumping speed for hydrogen isotopes and compatibility with ultra-high vacuum environments [2]. In the Large Helical Device (LHD), the NEG pumps have been installed into the one toroidal section of closed helical divertor [3]. In this work, a HV400 NEG module by SAES Getters S.p.A. [4] (nominal pumping speed: 400 L/s for H2) was installed in QUEST, and its hydrogen pumping performance was evaluated.
The NEG pump was installed on an ICF152 port located in the bottom conical region behind the hot wall (BSU114 port) through a gate valve and a straight pipe of approximately 0.5 m in length as shown in Fig. 1. The main vacuum pumping system consisted of a turbo-molecular pump (TMP) and four cryo-sorption pumps. In the NEG pumping test, only TMP, in which the pumping speed for H2 gas was evaluated at approximately 650 L/s in advance, was used. Two sets of quadrupole mass spectrometers (QMS) were used to measure partial pressures.
Activation was performed at 773 K for approximately 3 hours. After activation, the NEG pump was maintained at 473 K, compatible with the QUEST wall temperature in the high temperature wall experiments. Pumping tests were carried out by isolating the vacuum vessel and introducing hydrogen gas using a mass flow controller (0.5–1.0 mL/min). The temporal decay of hydrogen partial pressure was analyzed to evaluate the effective pumping speed.
Two activation processes were conducted (Total pressure before the activation was around 3 × 10−5 Pa). During activation at 773 K, the hydrogen partial pressure increased significantly, indicating hydrogen desorption from the NEG material as shown in Fig. 2(a). In contrast, increases in impurity species such as N2, O2, and H2O were strongly suppressed. It is a common feature of a NEG pump. However, an important implication of the present result is the potential for in situ regeneration of the NEG pump during plasma operation. Since the activation process primarily releases hydrogen while suppressing impurity species, it may be possible to regenerate the NEG pump without significantly degrading plasma conditions. This opens the possibility of continuous operation without the need for isolation by gate valves, enabling direct and sustained particle exhaust. Slight increases observed in these species are attributed to outgassing from the heated ICF piping.
At 473 K operation, a clear decay of hydrogen partial pressure was observed when only the NEG pump was working, while little effect was detected for N2, O2, and H2O (less than 20% of H2) as shown in Fig. 2(b). In contrast, the TMP provided non-selective pumping for all gas species. This confirms the hydrogen-selective pumping characteristic of the NEG pump under QUEST conditions.
The pumping test was conducted just after each activation under conditions where the background partial pressure of N2 in the vacuum vessel was relatively high (Ion current ratio of N2/H2 ≈ 35, O2/H2 ≈ 4.5, H2O/H2 ≈ 2.3) as shown in Fig. 2 (b) (Total pressure after hydrogen gas introduction by the mass flow of 1.0 mL/min was around 1.5 × 10−4 Pa). The temporal decay of hydrogen partial pressure was fitted using,
| (1) |
where the decay constant
The compatibility of the NEG operating temperature (473 K) with high wall temperature operation makes it particularly suitable for long-duration steady-state experiments. In the present test, the impurity level during H2 injection was relatively high. This likely limited the achievable performance for H2, which may improve significantly under lower impurity conditions.
This work was performed with the support and under the auspices of the NIFS Collaboration Research Program (NIFS25KSPT012). The authors would like to thank SAES Getters S.p.A. for valuable technical discussions regarding the operation and characteristics of the NEG pump.
