2. Project overview

As an open-user facility, the Space Plasma Environment Research Facility, abbreviated as SPERF, is dedicated to simulating the space plasma environment, including the near-earth and near-space plasma simulation systems (Jin et al. Reference Jin, Zhang, Ling, Chen, Wang, Hao, Dong, Lu, Li and E2023). The near-earth space plasma simulation system (abbreviated as near-earth system) is used to simulate the 3-D structure and process of the terrestrial magnetosphere. Its main scientific goals include exploring 3-D asymmetric magnetic reconnection related to the magnetopause (Mao et al. Reference Mao, Ren, Ji, E, Han, Wang, Xiao and Li2017), studying the global dynamics of the charged particles in the artificial magnetosphere (Mao et al. Reference Mao, Ma, E, Guan, Ling, Li and Li2020), investigating the evolution of the charged particles responding to the perturbation of the geomagnetic storm (Ling et al. Reference Ling, Jin, Zhu, Xu, Chen, Lu, Wu, Li and E2023b) and probing the physics of the magnetic reconnection occurring at the magnetotail (Xiao et al. Reference Xiao, Wang, Wang, Xiao, Yang and Zheng2017).

In response to the scientific objectives, the fundamental structure of the near-earth system is designed and its main parameters are determined based on the scaling laws of physical similarity (Xiao et al. Reference Xiao, Wang, Wang, Xiao, Yang and Zheng2017; E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). Figure 1 gives the diagram of the near-earth system of the SPERF, and table 1 shows its key parameters for simulating the magnetopause 3-D magnetic reconnection and radiation belt (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). A large cylindrical vacuum vessel with a length 10.5 m and radius of 2.5 m is adopted to provide an experimental region with a large spatial scale, which is crucial to simulating the magnetopause magnetic reconnection with a high Lundquist number, enlarging the effective region of the simulated radiation belt, and studying the physics in the macroscopic MHD (magnetohydrodynamic) regime. The key to replicating the physical environment of the space plasma is to simulate the plasma and magnetic field governing its dynamics (Ling et al. Reference Ling, Jin, E, Zhu, Xu, Chen, Lu, Wu and Li2022a). To make the laboratory simulation be relevant to the real space physics, it is significant to reproduce the geometrical configuration of the simulated physical process. A group of magnets of specific structure and corresponding plasma sources are designed to meet the requirement. Four flux core magnets are used to simulate the interplanetary magnetic field and induce the laboratory solar wind, and a circular magnet is adopted to mimic the earth's dipole magnetic field (Jin et al. Reference Jin, Zhang, Ling, Yang, Cheng, Feng, Wan, Wang, Lu and Li2022b). The straight interplanetary magnetic field and the dipole field combine to form a 3-D geometry analogous to that of the earth's magnetopause, making it possible to study the 3-D magnetic reconnection. Two electron cyclotron resonance (ECR) sources and a cool-cathode source are designed to create the plasma for Earth's magnetosphere simulation. The 3-D magnetic reconnection is driven by the impact of the simulated solar wind against the magnetosphere built in advance through making the currents of the flux core magnets rise rapidly. When the magnetosphere is simulated separately, the wave–particle interaction and the global dynamics of the charged particles in a dipole configuration can be explored. With an extra circular magnet concentric with the dipole magnet, the magnetic disturbance during the geomagnetic storm can be simulated for studying the response of the inner magnetosphere. Additionally, combination of the dipole field and a magnetic-mirror field will create a magnetic-reconnection configuration relevant to the earth's magnetopause. In this situation, a LaB$_{6}$ plasma source is used to simulate the magneto-sheath plasma, and the reconnection is driven by a plasma gun. Meanwhile, by using two LaB$_{6}$ plasma sources distributed symmetrically to simulate the magnetotail plasma sheets, the configuration can be easily adjusted to study the magnetotail reconnection.

Figure 1. Diagram of the near-earth system of the SPERF: (a) front view; (b) side view.

Table 1. Parameters for magnetic reconnection simulation.

Since there is an enormous difference in magnitude of the plasma characteristic parameters, such as size, magnetic field and plasma density, between the space and laboratory, it is impossible to replicate the real plasma parameters in the experimental device. However, the physics governing the space plasma can be reproduced in the laboratory with the plasma parameters scaled according to the similarity relations, which are obtained by making the dimensionless numbers in the normalized equations controlling the real and simulated plasmas constant. For magnetopause magnetic reconnection simulation, the collisionless Vlasov equations are used for the scaling and the equations of the ideal magnetohydrodynamics are applied to radiation belt simulation. The normalized Vlasov equations are expressed as follows:

(2.1)\begin{gather} \left(\frac{l_0}{v_0 t_0}\right) \frac{\partial \hat{f}_S}{\partial \hat{t}}+\hat{\boldsymbol{v}} \boldsymbol{\cdot} \frac{\partial \hat{f}_S}{\partial \hat{\boldsymbol{x}}}+\left(\frac{{e} B_0}{m_0 v_0 / l_0}\right) \frac{\widehat{q_{S}}(\hat{\boldsymbol{E}}+\hat{\boldsymbol{v}} \times \hat{\boldsymbol{B}})}{\hat{m}} \boldsymbol{\cdot} \frac{\partial \hat{f}_S}{\partial \hat{\boldsymbol{v}}}=0 , \end{gather}

(2.2)\begin{gather}\left(\frac{B_0}{l_0 n_0 {e} \mu_0 v_0}\right) \hat{\boldsymbol{\nabla}} \times \hat{\boldsymbol{B}}= \int \sum_S \widehat{q_{S}} \hat{\boldsymbol{v}}_S \hat{f}_S \, {\rm d} \hat{\boldsymbol{v}}_S , \end{gather}

(2.3)\begin{gather}\left(\frac{\varepsilon_0 E_0}{{e} n_0 l_0}\right) \hat{\boldsymbol{\nabla}} \boldsymbol{\cdot} \hat{\boldsymbol{E}}= \int \sum_S \widehat{q_{S}} \hat{f}_S \, {\rm d} \hat{\boldsymbol{v}}_S , \end{gather}

where $l_{0}$, $v_{0}$, $t_{0}$, $B_{0}$, $m_{0}$, $n_{0}$ and $E_{0}$ are the characteristic length, velocity, time, magnetic field, mass, number density and electric field; $e$, $\mu _{0}$ and $\varepsilon _{0}$ are the electron charge, vacuum permittivity and vacuum dielectric constant; $\hat {f}_S$, $\hat {t}$, $\hat {\boldsymbol {v}}$, $\hat {\boldsymbol {x}}$, $\hat {q}_S$, $\hat {\boldsymbol {E}}$, $\hat {\boldsymbol {B}}$ and $\hat {m}$ are the normalized distribution function of the charged particle of $S$ type, time, velocity, displacement, charge, electric field, magnetic field and mass.

To retain similarity, the dimensionless numbers in the above equations should be equivalent for the space and laboratory plasmas. Denote the subscripts of the physical quantities of the space and laboratory plasmas as 01 and 02, respectively, thus

(2.4)\begin{gather} \frac{l_{01}}{v_{01} t_{01}}=\frac{l_{02}}{v_{02} t_{02}}, \end{gather}

(2.5)\begin{gather}\frac{{e} B_{01}}{m_{01} v_{01} / l_{01}}=\frac{{e} B_{02}}{m_{02} v_{02} / l_{02}} , \end{gather}

(2.6)\begin{gather}\frac{B_{01}}{l_{01} n_{01} {e} \mu_0 v_{01}}=\frac{B_{02}}{l_{02} n_{02} {e} \mu_0 v_{02}}, \end{gather}

(2.7)\begin{gather}\frac{\varepsilon_0 E_{01}}{{e} n_{01} l_{01}}=\frac{\varepsilon_0 E_{02}}{{e} n_{02} l_{02}}. \end{gather}

Suppose the scaling factors of the length and temperatures of the charged particles satisfy

(2.8)\begin{equation} \frac{T_{01} }{T_{02}} =\left( \frac{l_{01} }{l_{02}}\right)^{\alpha } . \end{equation}

Then, since $T_{01} \propto v_{01}^{2}$, $T_{02} \propto v_{02}^{2}$, the scaling factors of the velocity fulfil

(2.9)\begin{equation} \frac{v_{01} }{v_{02}} =\left( \frac{l_{01} }{l_{02}}\right)^{\alpha /2} . \end{equation}

Meanwhile, substituting this expression into (2.4)–(2.7), we can obtain that

(2.10)\begin{gather} \frac{t_{01} }{t_{02}} =\left( \frac{l_{01} }{l_{02}}\right)^{1-\alpha /2} , \end{gather}

(2.11)\begin{gather}\frac{B_{01} }{B_{02}} =\left( \frac{l_{01} }{l_{02}}\right)^{\alpha /2-1} , \end{gather}

(2.12)\begin{gather}\frac{n_{01} }{n_{02}} =\left( \frac{l_{01} }{l_{02}}\right)^{{-}2}, \end{gather}

(2.13)\begin{gather}\frac{E_{01} }{E_{02}} =\left( \frac{l_{01} }{l_{02}}\right)^{{-}1-\alpha /2}. \end{gather}

For magnetic reconnection, there are three other parameters relevant, namely the plasma $\beta$, the inertial length $d_{S}$ of the charged particle and the Lundquist number $S$, which are expressed as follows:

(2.14)\begin{gather} \beta =\frac{n_{0}T_{0}}{B_{0}^{2}/2\mu _{0} }, \end{gather}

(2.15)\begin{gather}d_{s} ={c} \left(\frac{\epsilon_{0}m_{0} }{n_{0}q_{S}^{2} } \right)^{1/2}, \end{gather}

where $c$ is the light speed in vacuum and $q_{S}$ is the charge of the charged particle, and

(2.16)\begin{equation} S=\frac{\mu _{0}l_{0}v_{{A} }}{\eta } , \end{equation}

where $v_{{A} }$ is the Alfvén velocity,

(2.17)\begin{equation} v_{{A} }=\frac{B_{0}}{\sqrt{\mu _{0}n_{0}m_{{i} }} } , \end{equation}

where $m_{{i} }$ is the ion mass. The $\eta$ is the plasma resistivity, which is proportional to $T_{{e} }^{-3/2}$. Here, $T_{{e}}$ is the characteristic electron temperature.

With the above scaling, the $\beta$ and the ratio of $l_{0}$ to $d_{S}$ become constant, and $S$ satisfies the following scaling:

(2.18)\begin{equation} \frac{S_{1}}{S_{2}} =\left( \frac{l_{01}}{l_{02}} \right)^{2\alpha +1}. \end{equation}

The parameters for magnetic reconnection simulation can be chosen according to these scaling relations of physical similarity. Simple computation indicates that to reach global similarity, the plasma size in the laboratory should be larger than 60 m scaled from the real size of the magnetopause ($\sim 10R_{{e}}\approx 60\ 000$ km, $R_{{e}}$ is the earth radius). However, it is impossible to achieve this because of technology constraints. Thus, the SPERF focuses on local similarity. Table 1 shows the parameters of the real, scaled and designed parameters of the magnetosheath and the magnetosphere. Typical parameters of the magnetosheath and sub-solar magnetosphere are used in table 1 (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). It should be noted that for local similarity, the size of the current layer ($\sim R_{{e}}$) of the magnetopause magnetic reconnection is used in scaling, and the ion inertial length is computed for the proton. Remarkably, to make the plasma density and magnetic field lie in appropriate ranges that can be realized in the laboratory and beneficial to diagnosis of magnetic reconnection, the scaled plasma size should be set as 6 m. Due to technology and budget limitations, the SPERF is designed to have a radius of 2.5 m, and the experimental plasma size is approximately 1.5 m. Meanwhile, the needed plasma temperatures are not easily accessible, especially for magnetosphere simulation. Nonetheless, the plasma density and magnetic field can be simulated well in the laboratory, combined with the geometry similarity, which enables the SPERF to investigate the local physics of the magnetopause magnetic reconnection.

The normalized ideal magnetohydrodynamic equations for radiation belt simulation are given as follows:

(2.19)\begin{gather} \left(\frac{l_0}{t_0u_0}\right)\frac{\partial\hat{\rho}}{\partial\hat{t}}+\hat{\boldsymbol{\nabla}} \boldsymbol{\cdot} (\hat{\rho }\hat{\boldsymbol{u}})=0, \end{gather}

(2.20)\begin{gather}\hat{\rho}\frac{\partial\hat{\boldsymbol{u}}}{\partial\hat{t}}+ \left(\frac{u_0t_0}{l_0}\right)\hat{\rho}\hat{\boldsymbol{u}} \boldsymbol{\cdot}\hat{\boldsymbol{\nabla}}\hat{\boldsymbol{u}} ={-}\left(\frac{P_0t_0}{l_0\rho_0u_0}\right)\boldsymbol{\nabla}\hat{P} -\left(\frac{B_0^2t_0}{{\mu_0l}_0\rho_0u_0}\right) \hat{\boldsymbol{B}}\times\boldsymbol{\nabla}\times\hat{\boldsymbol{B}}, \end{gather}

(2.21)\begin{gather}\frac{\partial\hat{\boldsymbol{B}}}{\partial\hat{t}}= \left(\frac{u_0t_0}{l_0}\right)\hat{\boldsymbol{\nabla}}\times (\hat{\boldsymbol{u}}\times\hat{\boldsymbol{B}}), \end{gather}

(2.22)\begin{gather}\frac{\partial\hat{P}}{\partial\hat{t}}+ \left(\frac{u_0t_0}{l_0}\right)\hat{\boldsymbol{u}}\boldsymbol{\cdot} \hat{\boldsymbol{\nabla}}\hat{P}={-}\left(\frac{u_0t_0}{l_0}\right)\gamma\hat{P}\hat{\boldsymbol{\nabla}}\boldsymbol{\cdot}\hat{\boldsymbol{u}}, \end{gather}

where $u_{0}$, $P_{0}$, $\rho _{0}$ and $\gamma$ are the characteristic velocity, pressure, mass density and polytropic coefficient of the plasma fluid; $\hat {\rho }$, $\hat {\boldsymbol {u}}$ and $\hat {P}$ are the normalized mass density, velocity and pressure of the plasma fluid.

Being analogous to the magnetic reconnection situation, to retain similarity, the dimensionless numbers must be identical, which means that

(2.23)\begin{gather} \frac{l_{01}}{t_{01}u_{01}}=\frac{l_{02}}{t_{02}u_{02}}, \end{gather}

(2.24)\begin{gather}\frac{P_{01}t_{01}}{l_{01}\rho_{01}u_{01}}= \frac{P_{02}t_{02}}{l_{02}\rho_{02}u_{02}}, \end{gather}

(2.25)\begin{gather}\frac{B_{01}^2t_{01}}{l_{01}\rho_{01}u_{01}}= \frac{B_{02}^2t_{02}}{l_{02}\rho_{02}u_{02}}. \end{gather}

Suppose the characteristic mass density, length and pressure satisfy the following scaling relation:

(2.26a–c)\begin{equation} \frac{l_{01}}{l_{02}}=a,\quad \frac{\rho_{01}}{\rho_{02}}=b,\quad \frac{P_{01}}{P_{02}}=c . \end{equation}

Then, to keep the similarity, the characteristic time, fluid velocity and magnetic field must comply with

(2.27a–c)\begin{equation} \frac{t_{01}}{t_{02}}=a\sqrt{\frac{b}{c}},\quad \frac{u_{01}}{u_{02}}=\sqrt{\frac{c}{b}},\quad \frac{B_{01}}{B_{02}}=\sqrt c. \end{equation}

With the scaling, the plasma $\beta$ is kept constant and the Alfvén velocity transfers according to

(2.28)\begin{equation} \frac{V_{{A1}}}{V_{{A2}}}=\sqrt{\frac{c}{a}} . \end{equation}

Based on these scaling relations, the experimental parameters are determined and listed in table 2 (Xiao et al. Reference Xiao, Wang, Wang, Xiao, Yang and Zheng2017; E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). It is noted that the designed experimental parameters totally cover the scaled parameters. Thus, based on the geometry and parameter similarities, the radiation belt can be simulated well by the SPERF facility.

Table 2. Main parameters for radiation belt simulation.

For any experimental plasma device, it is significant to perform parameter diagnosis to capture the plasma state and its evolution for physical research. Hence, multiple plasma diagnostics tools, such as magnetic probe arrays, soft X-ray detector arrays and a polarization interferometer, are deployed in the facility to diagnose the main plasma parameters, including plasma density, electron temperature, magnetic field as well as their fluctuations. The resolutions of the diagnostic tools are needed to adapt to magnetic reconnection research. The SPERF $(\log (l_{0}/d_{{i}})=0.82\unicode{x2013}1.3$, $\log (S)=2.8\unicode{x2013}4.1$, computed with the experimental magnetosheath parameters) lies in the single X-line collisionless regime of the magnetic reconnection phase diagram (Ji & Daughton Reference Ji and Daughton2011). To study physics in the electron diffusion region, the spatial resolution should be of the order of the electron skin depth of the device ($d_{{e}}=1.7 \unicode{x2013}5.4$ mm). Meanwhile, the temporal resolution is required to be higher than the Alfvén transit time, which in our facility is of the order of 1 $\mathrm {\mu }$s.

3. Key components of the near-earth system of the SPERF

The near-earth system of the SPERF is mainly composed of the vacuum system, pulsed magnet system, pulsed power supply system, plasma source system and plasma diagnostic system. In the construction of these systems, several critical technologies have been broken through. In this section, the design and construction of these systems will be presented with the primary focus on the design goal, core parameters, crucial technical issues and solution scheme.

3.1. Vacuum system

The vacuum system aims to provide a pure working vacuum for the facility (Jin et al. Reference Jin, Zhang, Ling, Chen, Wang, Hao, Dong, Lu, Li and E2023). To achieve the goal, an ultimate background vacuum is built first and then working gases are injected into the vacuum chamber to obtain the experimental pressure. The main challenges posed on the vacuum system lie in the fact that the vacuum vessel is large in volume, has a lot of outgassing objects in it (Jin et al. Reference Jin, Zhang, Ling, Chen, Wang, Hao, Dong, Lu, Li and E2023), and possesses hundreds of interfaces on it for supporting the magnets, feeding the electrical currents into the magnets, installing the plasma sources, diagnosing the plasma, etc. As shown in figure 2, a multi-stage pumping unit is designed to achieve the ultimate background vacuum, and a pulsed gas-injection unit controlled by piezoelectric ceramic valves is adopted to establish the experimental vacuum. Meanwhile, the leakage rates of the main components in the vacuum vessel are reduced and considered in the design. The operation test shows that an ultimate vacuum of pressure at $10^{-4}$ Pa can be attained within 24 h based on the design, and the total leakage is smaller than $7.1\times 10^{-7}\ {\rm Pa} \cdot {\rm L}\ {\rm s}^{-1}$ (Jin et al. Reference Jin, Zhang, Ling, Chen, Wang, Hao, Dong, Lu, Li and E2023). The reader is referred to the previous papers (Jin et al. Reference Jin, Zhang, Ling, Liu, E, Chen, Li, Peng, Lu and Li2022a,Reference Jin, Zhang, Ling, Yang, Cheng, Feng, Wan, Wang, Lu and Lib, Reference Jin, Zhang, Ling, Chen, Wang, Hao, Dong, Lu, Li and E2023) to grasp the technology details of the vacuum system.

Figure 2. Schematic view of the vacuum system (Jin et al. Reference Jin, Zhang, Ling, Chen, Wang, Hao, Dong, Lu, Li and E2023).

3.2. Pulsed magnet system

Figure 3 shows the diagram of the pulsed magnet system. The structure and distribution of the magnets are designed to replicate the field geometry of the near-earth space plasma, and their parameters are dictated by the requirement of the magnetic field to keep the physical similarity. The magnet system is composed of five kinds of magnets (E et al. Reference E, Ling, Mao, Xiao, Guan and Zhang2020), i.e. the flux core magnets, a dipole magnet, magnetopause shape-control magnets (also named CK magnets), magnetic perturbation magnets (also named CRD magnets) and magnetic mirror magnets (also named CJC magnets). Four coaxial flux core magnets, each composed of four poloidal coils (abbreviated by PF) in parallel and four toroidal coils (abbreviated by TF) in parallel, are used to simulate the solar wind. The dipole magnet creates an approximated dipole magnetic field analogous to that of the Earth. Two groups of CK magnets are employed to adjust the field topology formed by the flux core magnets and the dipole magnet to increase experimental flexibility (Ling et al. Reference Ling, Jin, Yin, Guan, Zhu, Xu, Chen, Lu, Wu and Li2023a). Two CRD magnets are designed for modelling the magnetic disturbance during the geomagnetic storm. Two CJC magnets can produce a field to mimic the interplanetary magnetic field of the terrestrial magnetopause or magnetotail.

Figure 3. Schematic view of the pulsed magnet system (Ling et al. Reference Ling, Jin, Mao, E, Wu, Zhu, Chen, Lu and Li2022b).

The structural and electromagnetic parameters of the magnets are designed based on the constraints in physics and technology (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d), such as the field geometry, magnetic field magnitude, pulsed waveform of the field, temperature rise and conductor strain (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). Notably, the design of the pulsed magnet is coupled with that of the pulsed power supply, which is presented later. Table 3 shows the critical parameters of the pulsed magnet system. Several challenges are presented to the construction of the magnets, ranging from high-voltage insulation, vacuum sealing, support and motion in vacuum, resistance to the magnetic-force impact (E et al. Reference E, Ling, Mao, Xiao, Guan and Zhang2020), and cooling to conducting strong current through the interfaces. Many specific measures are proposed to rise to these challenges. Their details are presented in the papers continuously published in the Vacuum journal (E et al. Reference E, Ling, Mao, Jin, Zhu, Tan, Chen and Lu2021e; Ling et al. Reference Ling, Jin, E, Zhu, Xu, Chen, Lu, Wu and Li2022a,Reference Ling, Jin, Mao, E, Wu, Zhu, Chen, Lu and Lib, Reference Ling, Jin, Yin, Guan, Zhu, Xu, Chen, Lu, Wu and Li2023a,Reference Ling, Jin, Zhu, Xu, Chen, Lu, Wu, Li and Eb). Here, we only sketch the main points of the key technologies adopted in the construction of the magnets.

Table 3. Critical parameters of the pulsed magnet system.

The magnets can be grouped into two categories. One is the short-pulse magnet whose magnetic field reaches the target value at a typical time of the order of 0.1 ms, and the other is the long-pulse magnet that generates a magnetic field sustaining at the target value no smaller than 5 ms or 10 ms. The flux core magnets and CK magnets belong to the short-pulse magnets, and the dipole magnet and CJC magnets pertain to the long-pulse magnets. It is noted that the CRD I magnet has a short pulse, and the CRD II magnet is between the short-pulse and long-pulse magnets since its rise time can vary from 0.65 ms to 1.4 ms discretely. The short-pulse magnets except for the CRD I are all made from custom-made soft cables composed of enamel-insulated filaments for reducing the skin effect, and the long-pulse magnets and two CRD magnets are wound by the solid copper wire with a rectangular cross-section. The construction processes of the magnets are different due to the differences in their structure.

For the magnets made from the soft cables, the cables are inserted into a frame and guided through the stainless-steel tubes to the outside of the vacuum chamber. Epoxy resin is poured into the tubes for vacuum sealing, fixing the cables and improving the electrical insulation. The frame of the flux core magnets is made from G10 material. After being installed in the frame and wrapped with a layer of fibreglass tape soaked by epoxy resin (E et al. Reference E, Ling, Mao, Jin, Zhu, Tan, Chen and Lu2021e), the flux core magnets are enclosed by a thin aluminium shell for smoothing the toroidal magnetic field generated by their TF coils. Then, after being wrapped with a layer of fibreglass tape soaked by epoxy resin again (E et al. Reference E, Ling, Mao, Jin, Zhu, Tan, Chen and Lu2021e) and coated with the sol-gel coat for reducing the leakage rate, they are covered with a layer of Inconel to avoid the tip discharge that may pollute the plasma. For the CK magnets, since the frames containing their cables are stainless-steel tubes with extremely low leakage rates, there is no need for wet wrapping of fibreglass tapes and Inconel spraying.

For the magnets made from the solid conductor wire, the wire is first wound together to form a conductive body. Then, it is covered with an insulation layer relative to the ground. Notably, before this, the bodies of the dipole and CRD I magnet are embedded in a G10 frame for support. Afterwards, the magnet is addressed by the epoxy-resin impregnation method, which is important to decrease leakage, improve electrical insulation and increase mechanical strength (Ling et al. Reference Ling, Jin, E, Zhu, Xu, Chen, Lu, Wu and Li2022a). After that, the magnet is wrapped with the sol-gel coating and followed by creating a sand layer on the outer surface of the magnet (Ling et al. Reference Ling, Jin, E, Zhu, Xu, Chen, Lu, Wu and Li2022a). Finally, an Inconel layer is put over the surface of the magnet.

The support and motion mechanisms are specifically designed for all the magnets to bear intensive magnetic-force impact. The dipole magnet and CRD magnets are directly supported on a base that is connected to a pedestal outside the vacuum chamber through the penetration stanchions. The magnetic force exerted on the dipole magnet can be transferred to the pedestal through the stanchions without deforming the vacuum vessel. Multi-freedom motion is realized by a mechanism combining a spherical joint hinge with a guided screw, and an elastic shaft sleeve is adopted to buffer the strong magnetic-force impact (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d; Ling et al. Reference Ling, Jin, Mao, E, Wu, Zhu, Chen, Lu and Li2022b). The other magnets including the flux core magnets, CK magnets and CJC magnets are all supported and moved via the supporting tubes, which go through the vacuum vessel and are fixed to and driven by a mechanism installed on the vacuum vessel. Meanwhile, buffer plates in the transverse and vertical directions are added to mitigate the strong magnetic-force impact.

For the flux core magnets and CK magnets, the cables are directly led through the vacuum vessel with an interface sealed by the solidified epoxy resin. However, for other magnets, the solid copper wires are first converted to coaxial cables, which are then changed to the solid wires penetrating the vacuum vessel. The solid wires traversing the interface are further converted to the coaxial cables to connect with the pulsed power supplies. A dedicated structure is designed to achieve transition between the soft coaxial cables and the solid copper wires or the ordinary cables. The water through the cooling channel in the wires can be separated from the electrical connection by the transition structure. During the construction of the pulsed magnet system, a series of tests in terms of high-voltage insulation, magnetic field, inductance, resistance, leakage rate and hydraulic characteristics are done to validate the design. Now, as shown in figure 4, all the magnets have been constructed and installed in the facility. First tests with operation voltage smaller than 10 kV have been done for them, and results show that the magnet system is robust to fulfil its objective.

Figure 4. The magnet system of the SPERF.

3.3. Pulsed power supply system

The objective of the pulsed power supply (PPS) system is to provide excitation currents to drive the pulsed magnet system. The core challenge of the system is how to achieve two kinds of strong pulsed currents reliably, one with rise time and peak of the order of 0.1 ms and $100 \unicode{x2013} 600\ {\rm kA}$, and the other with flattop time and peak of the order of 10 ms and $10 \unicode{x2013} 20\ {\rm kA}$, for different inductive loads of the magnet system. The PPS generating the fast current is labelled with fast PPS and that providing the current with a flat top is marked with flat PPS. To accomplish the pulsed currents in these parameter regimes, the scheme of discharging a capacitor into an inductor with a crowbar to shape the pulse in the descending phase is employed due to the advantages in energy density, budget, reliability, feasibility, maintenance, etc. Meanwhile, a fundamental framework based on modularization is adopted for the PPS system to improve the reliability, operation and maintenance efficiency, expandability, and flexibility of pulse forming (Mao et al. Reference Mao, Ma, E, Guan, Ling, Li and Li2020). The module number and parameters of the capacitor and crowbar resistor are determined based on the principle of the LCR discharging circuit (Mao et al. Reference Mao, Ma, E, Guan, Ling, Li and Li2020), and two types of modules composed of the charging unit and discharging unit are designed (E et al. Reference E, Guan, Ling, Ma, Mao, Deng, Ding, Li, Kang and Li2021b).

Many critical technologies have been attacked to construct the pulsed power supply, ranging from switch assembly, crowbar branch, electromagnetic compatibility and connection of strong current. A series of papers have been published by our team to discuss these key points (Mao et al. Reference Mao, Ma, E, Guan, Ling, Li and Li2020; E et al. Reference E, Guan, Jin, Mao, Ma, Deng, Ding, Kang, Li and Xiao2021a,Reference E, Guan, Ling, Ma, Mao, Deng, Ding, Li, Kang and Lib,Reference E, Guan, Ma, Mao, Deng, Ding, Kang, Li, Xiao and Lic; Guan et al. Reference Guan, E, Ma, Mao, Deng, Ding, Kang, Li, Xiao and Tong2021a,Reference Guan, E, Mao, Ma, Jin, Deng, Ding, Li, Kang and Xiaob, Reference Guan, Ling, Ma, Li, Tong and E2022), and here we only summarize them simply. The thyristors are adopted as the discharging switch, which is a significant element of the pulsed power supply, due to high reliability, long lifetime and easy controllability. To conduct the strong pulsed current in the condition of high voltage, a switch assembly with five thyristors connected in series is designed along with the trigger circuit, protection circuit and monitoring unit. The thyristors with similar switch performance are used in one assembly to make them turn on synchronously, and tests have been done to validate this before integrating the assembly into the module. Meanwhile, a protective inductor between the capacitor and discharging switch has been optimized for the flat PPSs to reduce the magnitude and changing rate of the current through the thyristors to prevent them from damage by the short-circuit fault (E et al. Reference E, Guan, Ling, Ma, Mao, Deng, Ding, Li, Kang and Li2021b).

The crowbar branch is composed of a diode assembly and a crowbar resistor, which is used to adjust the descending time of the current and alleviate the thermal burden of the magnet. The diode assembly is constructed by five power diodes connected in series to support the high voltage stress, and they have been chosen to have similar off-state impedance for preferable voltage-sharing performance. The crowbar resistor is optimized to be not melt by the temperature rise and have low inductance, and the cooling through forced convection of the air is adopted to raise the repetitive frequency of the PPS. Electromagnetic compatibility must be considered to reduce the influences of the electromagnetic interference (EMI) from the strong current and voltage. A distributed control system is built for the PPS with the EMI immunity done. A floating ground is used for the controller, and the coupling through distributed capacitance with the discharging circuit is mainly restrained by the isolation transformers used in the circuit powering the controller and data-acquisition elements and the optical fibre communication. Meanwhile, the original acquired voltage signal connecting to the high-voltage circuit is sent to the controller by transforming it first to the frequency signal, then to an optical signal and finally to the voltage signal. In addition, the strong pulsed current is transmitted by the coaxial cables with a shielding layer to inhibit electromagnetic interference.

The connection structure of the strong pulsed current is designed to collect the coaxial cables of the discharging modules and convert the coaxial cables to the wires of the magnets. The polarity reversal can be achieved by it, and the connection modes of the coils of the lux core magnets and CK magnets can also be changed by it. In addition, a synchronous control system with a resolution of 5 ns and a safety interlock system are designed to achieve the integrated control and operation of all the power supply modules. Based on these crucial technologies, a pulsed power supply with a total energy of 18.3 MJ has been built and passed tests on the dummy loads. Meanwhile, initial tests on the magnet system with a voltage below 10 kV show good results. Parameters of the PPSs for the pulsed magnets are shown in table 4. Figure 5 shows the physical picture of the PPS system of the SPERF.

Table 4. Parameters of the pulsed power supplies (Mao et al. Reference Mao, Ma, E, Guan, Ling, Li and Li2020)

$T_{\rm tp}$ and $I_{\rm tp}$ mean the typical time and typical current of the fast PPSs. $T_{\rm flat}$ and $I_{\rm peak}$ mean the flattop time and peak current of the flat PPSs. $T_{\rm D}$ and $W$ stand for the descending time of the current and the PPS's stored energy. It should be noted that the CRD II PPS can provide five kinds of current with different parameters, which are not given here, and the reader can find them in the paper by Guan et al. (Reference Guan, E, Mao, Ma, Jin, Deng, Ding, Li, Kang and Xiao2021b).

Figure 5. The PPS system of the SPERF.

3.4. Plasma source system

Several plasma sources are adopted on the SPERF near-earth system to create the plasma, including the TF coils of the flux core magnets to simulate the solar wind, two ECR plasma sources and one cool-cathode source to simulate the magnetosphere's plasma, a thermal-cathode plasma source simulating the solar wind or the plasma of the terrestrial magnetotail, and a plasma gun to drive magnetic reconnection. The parameters of these plasma sources are given in table 5. It is of extreme significance to achieve the plasmas reliably by these sources for further physical research. A lot of critical technological issues must be tackled to reach the goal, and it is necessary to optimize their performance and obtain the parameter regime before formal experiments.

Table 5. Parameters of the plasma sources on the near-earth system of the SPERF.

3.4.1. TF coils of the flux core magnets for solar-wind plasma simulation

The simulated solar wind is formed by the TF coils of the flux core magnets through inductive coupling (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d,Reference E, Ling, Mao, Jin, Zhu, Tan, Chen and Lue). An azimuthal electric field induced by the ramping toroidal magnetic field driven by a pulsed current through the TF coils will ionize the gas around the small circle of the flux core magnets and produce the plasma. The first plasma and its motion pushed by the poloidal field of the PF coils of the flux core magnets have been achieved (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). The plasma density was measured by a probe collecting the ion saturation current in an initial test. The pressure is set as 0.5 Pa and the discharging gas is argon. Both the discharging voltages of the TF and PF coils are 3 kV, and the PF and TF coils are triggered at 1 ms and 1.05 ms, respectively. The probe is located at the axis of the SPERF, and the distance from the axis of the flux core magnets is 1.5 m. The probe's effective collection area is equal to $4.8\times 10^{-6}\ {\rm m}^{2}$. The measured plasma density and the currents of the PF and TF magnetic coils are shown in figure 6. We can see that evolution of the plasma density is complicated. The first peak of the density indicates plasma motion driven by the magnetic pressure of the PF coils and the thermal pressure of the plasma itself. Meanwhile, the second peak is believed to be caused by the outflow of the magnetic reconnection occurring between the middle two flux core magnets. The peak plasma density approaches $2\times 10^{18}\ {\rm m}^{-3}$, which is sufficient for magnetic reconnection research. The plasma density can be increased by improving the discharging voltages of the magnetic coils or increasing the gas pressure. Therefore, the solar-wind plasma can be modelled by the flux core magnets and it is possible to drive magnetic reconnection by pressing the plasma outward using the poloidal field of the magnets.

Figure 6. Plasma density of the TF coils.

3.4.2. Cool-cathode plasma source

The simulated magnetosphere plasma is created by a cold-cathode plasma source with the ability to form a high-density (${>}1\times 10^{13}\ {\rm cm}^{-3}$) plasma. It consists of a round anode and a meshed cathode, which is made from molybdenum owing to its advantages of high emission efficiency of electrons, high strength and fine temperature resistance. A high-voltage pulse is exerted on the cathode relative to the anode connecting with the ground to excite glow discharge based on Paschen's law. The created plasma has a low electron temperature of 1 eV to 5 eV. The main challenge on the cool-cathode source is to trigger the plasma, and improve its stability and azimuthal uniformity (Xiao et al. Reference Xiao, Wang, Wang, Xiao, Yang and Zheng2017). The source will work in the prescribed dipole field of the modelled magnetosphere. Sizes of the electrodes of the cool-cathode plasma source are defined based on the magnetic field geometry to generate large-volume plasma simulating the radiation belt of the earth. The radius of the cathode and anode is designed as 1.75 m and 1.2 m, respectively, and the distance between them can be adjusted to control the region of the plasma. A capacitor-based pulsed high-voltage (up to 20 kV) supply is designed to drive the source to achieve the large-volume plasma sustaining no less than 1 ms. The existence of the magnetic field may substantially increase the breakdown threshold, which can be diminished by the pre-ionization provided by an ECR plasma source which is also deployed at the SPERF. The uniformity of the plasma can be reached by making the electric field uniform through reducing the installation error and improving the smoothness of the electrodes. Now, as shown in figure 7(a), the cool-cathode plasma source has been constructed and integrated into the vacuum vessel, and it is being under test.

Figure 7. Cathode plasma sources: (a) cool-cathode plasma source; (b) thermal-cathode plasma source.

3.4.3. $LaB_{6}$ thermal-cathode plasma source

A ${\rm LaB}_{6}$ thermal-cathode plasma source is used to generate a large-surface plasma having a density of up to $1\times 10^{13}\ {\rm cm}^{-3}$, which enters the experimental region by diffusion along the magnetic field. The continuous operation time of the source is required to be larger than 1000 s. The main challenge on the plasma source is to make it operate stably for a long time without fragmenting the ${\rm LaB}_{6}$ cathode. The core task is to achieve enough uniform heating of the ${\rm LaB}_{6}$ cathode surface. A ${\rm LaB}_{6}$ cathode with a large area of 150 ${\rm cm}^{2}$ ($30\ {\rm cm} \times 5\ {\rm cm}$) is manufactured by joining many ${\rm LaB}_{6}$ segments to avoid cracks due to thermal stress arising from the temperature change. To match with the design, the cathode is indirectly heated by the S-shaped tungsten wire for uniform heating. Meanwhile, a heat-shielding room is specifically designed to include the cathode to prevent heat loss. The temperature should be in the range of $1000\unicode{x2013} 1500\,^\circ {\rm C}$ to obtain a high electron-emission efficiency of the thermal cathode and impede tungsten evaporation. The plasma density is decided by the working temperature of the thermal cathode, gas condition and biased voltage of the anode. The anode is made from oxygen-free copper with high electrical and thermal conductivity, and cooling has been designed for it. In addition, the ${\rm LaB}_{6}$ thermal-cathode plasma source can move in vertical and horizontal directions to change the plasma region. At present, as shown in figure 7(b), construction of the ${\rm LaB}_{6}$ thermal-cathode plasma source has been completed, and it is now also under installation and test.

3.4.4. ECR plasma source

On the near-earth system of the SPERF, two ECR plasma sources are employed to generate the plasma with high electron energy, shape the plasma distribution and provide the seed electrons for the cool-cathode plasma source. Their microwave frequencies are 2.45 GHz and 6.4 GHz (Xiao et al. Reference Xiao, Wang, Wang, Xiao, Yang and Zheng2017), and the maximum power of both sources is equal to 10 kW. The density threshold is $7.5\times 10^{10}\ {\rm cm}^{-3}$ for the 2.45-GHz plasma source and it is $5.1\times 10^{11}\ {\rm cm}^{-3}$ for the 6.4 GHz plasma source (Xiao et al. Reference Xiao, Wang, Wang, Xiao, Yang and Zheng2017). The continuous operation time of the plasma sources is needed to surpass 10 ms (E et al. Reference E, Ling, Mao, Jin, Xiao, Xu, Chen, Bai, Hao and Lu2021d). The primary challenge posed on the ECR plasma sources is how to couple the microwave power into the plasma confined in the simulated dipole field of the magnetosphere efficiently. Microwave power is mainly absorbed at the resonant surface, in which the gyrating electron resonates with the wave. For the near-earth system of the SPERF, the microwave can be absorbed by the plasma through multiple reflections on the vacuum vessel, which serves as an ideal conductor. This absorption principle is used in many similar facilities including LDX (levitated dipole experiment) (Garnier et al. Reference Garnier, Hansen, Kesner, Mauel, Michael, Minervini, Radovinsky, Zhukovsky, Boxer and Ellsworth2006), CTX (collisionless terrella experiment) (Maslovsky, Levitt & Mauel Reference Maslovsky, Levitt and Mauel2003) and RT-1 (ring trap 1) (Nishiura et al. Reference Nishiura, Yoshida, Saitoh, Yano, Kawazura, Nogami, Yamasaki, Mushiake and Kashyap2015; Ling et al. Reference Ling, Jin, Zhu, Xu, Chen, Lu, Wu, Li and E2023b). Despite this, since the size of the vacuum vessel is large and there are many mechanical components in it, the effectiveness of this absorption mechanism may descend. Thus, to increase the microwave coupling, we designed two curved reflectometers, located above and below the dipole magnet, respectively (Ling et al. Reference Ling, Jin, Zhu, Xu, Chen, Lu, Wu, Li and E2023b). In addition, the angle of the injecting antenna can be adjusted in the range of $-5^{\circ }\unicode{x2013}5^{\circ }$ to regulate the injection direction of the microwave. Meanwhile, the plasma distribution can be shaped through adjusting the powers of two ECR sources and changing their antenna's angles. Presently, as shown in figure 8, the 2.45-GHz ECR source has been installed in the facility and is being tested. The first plasma without the arc reflectometers has been created by the 2.45-GHz plasma source successfully, and the 6.4-GHz source is under construction.

Figure 8. Microwave source and transfer system of the 2.45-GHz ECR plasma source.

3.4.5. Plasma gun of capacitive discharge

A plasma gun installed on the right end of the vacuum vessel is adopted to generate a hypersonic plasma jet perpendicular to the field of the magnetic-mirror magnets to drive 3-D magnetic reconnection, whose configuration is created by the dipole magnet and the magnetic mirror magnets as well as the corresponding plasmas confined by them. The plasma is accelerated out from the cavity between the coaxial electrodes of the gun under the self-consistent magnetic force. The gun has been integrated into the SPERF and tested. Test results show that the plasma is generated by the gun successfully, with a maximum plasma density larger than $1\times 10^{12}\ {\rm cm}^{-3}$, velocity of the plasma jet surpassing 80 km s$^{-1}$, diameter of the plasma with a density no smaller than $1\times 10^{11}\ {\rm cm}^{-3}$ reaching 20 cm and plasma duration longer than 200 $\mathrm {\mu }$s.

3.5. Plasma diagnostic system

Several plasma diagnostic tools are deployed on the near-earth system of the SPERF to diagnose the plasma number density, electron temperature, magnetic field and their fluctuations. The main tools of the diagnostic system adopted on the near-earth system of the SPERF, and their time and spatial resolutions are shown in table 6. The basic set-up and parameters of the system will be introduced in this section.

Table 6. Main parameters of the diagnostic tools on the near-earth system of the SPERF.

$n_{{e}}, T_{{e}}, V_{{p}}$, $B$ are the plasma concentration, electron temperature, plasma potential and magnetic field; the quantities with $\sim$ stand for the fluctuation of the corresponding equilibrium quantities; $\int {n_{{e} } }\,\mathrm {d}s$ and $\int {n_{{e} }B_{\parallel } }\,\mathrm {d}s$ are the chord-integrals of the $n_{{e}}$ and $n_{{e}}B_{\parallel }$, where $B_{\parallel }$ is the magnetic field in the direction of the light chord.

3.5.1. Electrostatic probes and magnetic probe arrays

A single probe and a triple probe are used to obtain the plasma concentration, electron temperature (Ling et al. Reference Ling, Jin, Zhu, Xu, Chen, Lu, Wu, Li and E2023b), space potential and their fluctuations at a point with spatial resolution smaller than 1 cm. The time resolution is 5 ${\mathrm {\mu }}$s and 0.5 ${\mathrm {\mu }}$s, respectively. The purpose of diagnosis with the electrostatic probes is to assist in deducing the profile of the plasma distribution by analysing the chord-averaged data from the optical diagnostic instruments.

As a significant diagnostic tool for studying magnetized plasma behaviours, a group of magnetic-probe arrays is deployed on the SPERF to diagnose the magnetic field as well as its fluctuation. These arrays can be categorized into two types, namely the high-resolution array and standard-resolution array. The diagram of the two kinds of arrays is shown in figure 9(a). For the high-resolution array, the probes are arranged as follows. Four one-dimensional (1-D) probes are placed in the centre of the array with an interval of 2 mm, and around them, ten 3-D probes are distributed symmetrically on two sides of the array, with the first three probes 10 mm apart and the last two probes 30 mm apart on one side. As to the standard-resolution array, there are ten 3-D probes in total distributed symmetrically on two sides of the array about its centre. On one side of the array, the distance of the first three probes is 10 mm and it is 30 mm for the last two probes. For both arrays, the time resolutions of measuring the magnetic field and its fluctuations are 1 ${\mathrm {\mu }}$s and 20 ns, respectively, and the respective spatial resolutions for the high-resolution and standard-resolution arrays can attain 2 and 5 mm. A mechanism is designed to support five arrays 60 mm apart and change their position to achieve measurement in a large region. Figure 9(b) shows that the magnetic-probe arrays have been completed, and they are now under calibration and test.

Figure 9. (a) Diagram of the magnetic-probe arrays and (b) their practical pictures.

3.5.2. X-ray detectors

The distribution of the electron temperature on the circular cross-section of the near-earth system of the SPERF is reconstructed from a group of soft X-ray detectors arranged in 12 fans, which are approximately equally spaced in angle on the circle 2.5 m in radius. Each fan includes 10 chords, and the emission along each chord is measured by two detectors blocked by respective slits, both of which has a beryllium film of different thickness for filtering. The average spatial resolution of the adjacent chords in each fan is approximately 13 cm on the central plane of the device when the width of the slits is 0.5 mm. The locations of the X-ray detectors and the chords of one fan are shown in figure 1. There are 120 channels of X-ray signals for both cases with different beryllium-film filtering, which is sufficient for the reconstruction of the 2-D emission profile. The electron temperature can be obtained by the ratio of the emissions reconstructed based on the different beryllium-film filtering. The spatial resolution of the induced electron temperature can attain a high value of approximately 1 cm. The measurable range of the electron temperature is $20\unicode{x2013} 100\ {\rm eV}$ with the plasma concentration larger than $1\times 10^{12}\ {\rm cm}^{-3}$, and the time resolution of the diagnosis can reach 10 ${\mathrm {\mu }}$s. Additionally, a FAST SDD detector is employed to probe the X-ray spectrum in the range of 500 eV–10 keV for analysing the emission characteristic of the plasma. The X-ray detectors are very useful in revealing the physics of magnetic reconnection and studying the transport of the energetic particles of the modelled magnetosphere.

3.5.3. Polarization interferometer

A far-infrared three-wave polarization interferometer is deployed for obtaining the chord-average plasma density and Faraday angle (Yang et al. Reference Yang, Gao, Shi, Zhou, Wang and Zhou2021). The chord of the interferometer is in the radial direction of the near-earth system of the SPERF, and the spatial resolution of the chord is smaller than 4 cm. Now, though only one chord is designed, the system can be extended to include many chords easily in the future. The interferometer can operate in a continuous mode with a time resolution smaller than 1 ${\mathrm {\mu }}$s. Three waves are generated by three solid microwave sources with frequency differences of the order of 1 MHz to approximately 320 GHz. The intermediate-frequency signal from a mixture of the left-hand and right-hand polarized probe waves contains information on the Faraday angle (Hare et al. Reference Hare, Lebedev, Suttle, Loureiro, Ciardi, Burdiak, Chittenden, Clayson, Eardley and Garcia2017b), while the chord-average density is deduced from the intermediate-frequency signals generated by mixing the reference wave with the left-hand and right-hand circularly polarized probe waves. The Faraday angle is used to give the chord integral of the plasma density times the magnetic field in the direction of the chord. Additionally, by adjusting the location of the detectors, the signal of super-thermal collective scattering can be obtained to diagnose the plasma density fluctuation. The polarimeter-interferometer system has been constructed and now is under test and calibration.

4. Possible physical experiments

Several physical experiments are accessible in the SPERF facility. As follows, we only sketch some typical ones. The physics of 3-D asymmetric magnetic reconnection relevant to the Earth's magnetopause will be investigated at the SPERF for the first time in the world. The most important feature of the SPERF is that it can achieve the magnetic reconnection with 3-D magnetic field geometry similar to that of the terrestrial magnetopause. This enables research on physics of 3-D magnetic reconnection and expands our abilities to understand the magnetopause magnetic reconnection using controlled experiments. According to the parameters shown in table 1, the reconnection available in the facility $((\log (L/d_{{i}})=0.82 \unicode{x2013}1.3, \log (S)=2.8\unicode{x2013}4.1)$ belongs to the single X-line collisionless regime of the reconnection phase diagram. Meanwhile, the ratio of the magnetic field of the simulated magnetosheath to that of the simulated sub-solar magnetosphere is equal to $0.25 \unicode{x2013} 1$, and that of the plasma density lies in the range of $1 \unicode{x2013} 100$. Thus, the asymmetry degree of the magnetic reconnection at the earth's magnetopause, for which the ratios of the magnetic field and plasma density between both sides of the reconnection region are approximately 0.3 and 10, respectively, can be reproduced well in the SPERF.

Several crucial physical issues in the 3-D collisionless regime of magnetic reconnection can be attacked based on the unique geometry and parameter space of the facility. The structures of the diffusion regions and the dominant physical mechanisms accounting for them for 3-D magnetic reconnection will be explored to test the validity of the theory of the multiple-layer structure that has been established well for the 2-D reconnection. Especially, it is possible to use the facility to verify the existence of the 3-D magnetic-null points, which have been observed for the magnetopause magnetic reconnection (Dunlop et al. Reference Dunlop, Zhang, Xiao, He, Pu, Fear, Shen and Escoubet2009; Fu et al. Reference Fu, Cao, Cao, Wang, Vaivads, Khotyaintsev, Burch and Huang2019; Guo et al. Reference Guo, Pu, Wang, Xiao and He2022) but not validated in experimental devices. If the 3-D magnetic-null points can be found in the facility, related issues such as structure of the null points, conditions of their existence and physical processes governing them will be able to be studied. Meanwhile, even though a lot of research focuses on the dynamics of the electron diffusion region, it is still not enough to totally uncover its mystery, particularly from the perspective of the laboratory experiments. The SPERF is appropriate to study the physics of the electron diffusion region. For example, the electron pressure anisotropy is accessible to the SPERF as is done by the TREX device (Forest et al. Reference Forest, Flanagan, Brookhart, Clark, Cooper, Désangles, Egedal, Endrizzi, Khalzov and Li2015) so that its role on the reconnection dynamics can be investigated in the 3-D case. Furthermore, the energy conversion and partitioning process in the 3-D reconnection can be studied, which is valuable for evaluating the efficiency of energy coupling into the magnetosphere caused by the magnetic reconnection at the Earth's magnetopause. In addition, the SPERF may create turbulent reconnection due to its large experimental region, which enables research on the interaction between the wave-induce turbulence and the reconnection. In the future, the multiscale coupling issue may also be explored in the SPERF as the regime is expanded with the development of the plasma-source and diagnostic technologies.

Another crucial ability of the SPERF is to achieve the simulation of the Earth's radiation belt, which creates an environment for studying the physics relevant to the magnetosphere dynamics. The issue of the wave–particle interaction is under the central concern of scientists to explain the generation, transport and loss of the energetic particles in the magnetosphere. However, some critical points relevant to the issue wait for elaborate experimental research, some of which can be explored in the SPERF. A significant one is to study the nonlinear wave–particle interaction and its implication on the global dynamics of the radiation belt. At present, it is recognized that the whistler wave plays an essential role in the wave–particle interaction in the radiation belt, while the key physics of the wave is still elusive (Frantsuzov et al. Reference Frantsuzov, Artemyev, Zhang, Allanson, Shustov and Petrukovich2023). The SPERF facility can be used to perform some research to bridge the gap. For example, it is possible to clarify the influence of the non-uniformity of the background magnetic field on the generation and propagation of the whistler wave. Meanwhile, the mechanism of exciting the chorus wave in the dipole configuration can also be studied in the facility. Additionally, the parameters of the facility are chosen such that the physics on the MHD scale is also accessible. Thus, the Alfvén wave can be driven in the SPERF for exploring its role in influencing the energy and momentum transport of the magnetosphere.

In addition, the facility has a unique feature to study the response of the inner magnetosphere to the magnetic disturbance during a geomagnetic storm, which has not been reported elsewhere. This function makes it possible to reveal the evolution dynamics of the energetic particles during a geomagnetic storm in the laboratory. In addition to the magnetic reconnection driven by the flux core magnets, the magnetic reconnection triggered by a plasma gun can also be achieved in the facility. The intention of this reconnection scheme is to model the magnetic reconnection at the Earth's magnetotail. The physical issues related to the magnetotail reconnection can thus be probed in the facility, one of which is the relationship between the dipolarization front and the earthward plasma jet. In summary, the SPERF will serve as an exceptional facility for research on space plasma physics due to its expanded parameter regimes and diverse functions.