1 Introduction

Due to the rapid climate change, major countries promote the declaration of carbon neutrality after the Paris Agreement (effective in 2016) [1] and the UN Climate Summit (September 2019) [2]. Korea announced and confirmed the “2050 carbon neutrality promotion strategy” in December 2020 [3] and will (1) reduce carbon in the economic structure, (2) create a promising low-carbon industrial ecosystem, and (3) promote a process transition to a carbon-neutral society [4]. The United States (US) returned to the Paris Agreement in January 2021 [5] and present the goal of “net zero by 2050” [6]. The European Union (EU) presented the goal of “2050 carbon neutrality” through the “Green Deal” in December 2019 [7], and the United Kingdom (UK) passed “the Carbon Neutrality Act” intending to achieve net-zero emission by 2050 in June 2019 [8]. China declared the achievement of carbon neutrality by 2060 in September 2020 [9], and Japan declared the goal of 2050 carbon neutrality in October 2020 [10].

To achieve goals of carbon neutrality and to respond to climate change, candidates of future energy sources should not emit carbon, must be clean and safe, producing sufficiently large energy to support economic growth, industrial development, and quality of life while they have minimal effects on the environments. A large-scale investment and technology development are being promoted [11]. Leading countries such as the USA, the EU, China, and Japan are strengthening investment and support for fusion energy development as part of efforts to respond to climate change, the green industrial revolution, and carbon neutrality.

Fusion energy satisfies all these requirements listed above as a future energy source, and its potential has been recognized. To secure fusion energy technology, seven countries including Korea, the USA, the EU, Russia, India, China, and Japan are participating in the construction of the International Thermonuclear Experimental Reactor (ITER) in Cadarache, France [12]. ITER is the world’s largest fusion device that will serve as a core base for fusion demonstration (DEMO) reactors around the world. More than 70% of its construction is complete, and its main device assembly began in earnest in 2020. The first plasma is scheduled for 2025.

Major countries already started to draw possible concepts and designs of DEMO reactors. Since there have been several review papers on the DEMO reactors [13,14,15,16,17], we do not intend either to make a similar technical review of the design parameters or to suggest improvements of those parameters. This paper aims to point out open questions and issues arising from points of view other than plasma physics and finally to ask an important question of how fusion energy would debut in the electricity market timely to contribute zero carbon neutrality in 2050.

This paper is organized as follows. Section 2 describes briefly the recent development strategy of the fusion technology for DEMO reactors in major countries. In Section 3, discussions on key factors of a DEMO design, such as factors related to device size, magnets (BT), plasma shape, and plasma current (Ip), will be given. Open questions and issues on the construction and operation of a DEMO reactor such as (1) identity, (2) net electricity output, (3) construction cost, and (4) public acceptance of fusion energy will be discussed in Section 4. In Section 5, a two-pathway approach towards the realization of fusion energy within the time constraint of carbon neutrality in 2050 is suggested. A conclusion is followed in Section 6.

2 Recent Research and Development Strategy on DEMO Reactors

2.1 The European Union

The EU is promoting the development road map to build an EU-DEMO reactor in the 2040s and demonstrate electricity production in the 2050s [13]. As a key element of the DEMO reactor development road map, the preconceptual design of the EU-DEMO reactor has been reviewed by experts as the first phase (2014–2020). The promotion to the next conceptual design phase (2021–2027) has been approved in November 2020 [18].

The Joint European Tokamak (JET), which is the world’s largest tokamak and core research device of the EU fusion program, completed the upgrade for the deuterium-tritium (DT) experiment scheduled for 2021 [19,20,21].

On the other hand, Wendelstein 7-X, the world’s largest stellarator, is to investigate the suitability of this type for a fusion power plant other than the tokamak concept [22]. The main goal of Wendelstein 7-X is to test an optimized magnetic field for confining the plasma without a large current flowing as in a tokamak plasma. The main assembly of Wendelstein 7-X was concluded in 2014; the first plasma was produced on 10th December 2015. Recently, Wendelstein 7-X showed that the energy losses of the plasma in the optimized magnetic field cage are reduced in the desired way [23].

For the fusion material testing and verification, the DONES-PreP project has been started as a preparatory stage for the construction of the International Fusion Materials Irradiation Facility–Demo Oriented NEutron Source (IFMIF-DONES) [24]. Sixteen EU organizations are participating in a comprehensive review of governance, safety, usability, and expected effects from a legal and economic perspective.

The UK announced the “Ten Key Plans for the Green Industrial Revolution,” which includes the plans to secure the fusion energy commercialization technology (2020). The UK jointly participated in the conceptual design of Spherical Tokamak for Energy Production (STEP) to build the world’s first fusion power plant prototype by 2040 [25]. The conceptual design is scheduled for completion in 2024. The UK Atomic Energy Authority (UKAEA) has started an open call for candidate sites for the STEP construction and plans to finalize the selection of the site in 2022 [26]. To fill the technology gap, the UK is actively seeking solutions to the challenges of developing fusion energy necessary for STEP construction. The upgrade of MAST (MAST-U) at Culham Centre for Fusion Energy (CCFE) was successfully completed in November 2020 after 7 years of effort [27]. Core infrastructures such as Remote Applications in Challenging Environments (RACE), Fusion Technology Facilities (FTF), Hydrogen-3 Advanced Technology (H3AT), and Materials Research Facility (MRF) will contribute to STEP technology development and construction [28].

2.2 Japan

Japan is promoting the DEMO development road map to build a DEMO reactor in the 2040s and demonstrate electricity production in the 2050s. The DEMO development road maps of Japan and the EU have very similar features because of the Japan-EU Broader Approach (BA) activity [29]. BA activity is one of the main streams to lead the direction of the Japanese fusion program. BA activity started in 2007 and completed the first phase (June 2007–March 2020) in April 2020. On 2 March 2020, the European Atomic Energy Community (Euratom) and the Japanese government signed a joint declaration for the extension and entered the second stage [30]. As a part of BA activity, the assembly of JT-60SA has been completed in March 2020 [31, 32] and is undergoing integrated test operation. The preconcept design of J-DEMO is in progress, and a preliminary conceptual design and comprehensive R&D review are planned in 2021 before the transition to the concept design phase [33]. For fusion material research, the IFMIF prototype linear accelerator, LIPAc, is being built and verified through IFMIF/EVEDA [34, 35]. Japan is also in the process of designing for the construction of the A-FNS (neutron irradiation facility) based on the technology gained from the R&D of IFMIF/EVEDA [36].

2.3 The USA

The Fusion Energy Sciences Advisory Committee (FESAC) presented a long-range plan to deliver fusion energy and to advance plasma science emphasizing that the Fusion Science and Technology area should focus on establishing the scientific and technical basis for a fusion pilot plant by the 2040s. Overall, special emphasis has been made to advance solutions to remaining technical and critical gaps. The approach of the USA revealed in the report of the FESAC pointed out the importance of public-private partnerships on the development of Fusion Science and Technology in 2020 [37].

In addition to the FESAC recommendation, the National Academies of Sciences, Engineering, and Medicine organized a committee to develop a strategic plan for US fusion research and published the Final Report of the Committee on a Strategic Plan for U.S. Burning Plasma Research [38]. In this report, two main recommendations were drawn: (1) the USA should remain an ITER partner as the most cost-effective way to gain experience with a burning plasma at the scale of a power plant. (2) The USA should start a national program of accompanying research and technology leading to the construction of a compact pilot plant that produces electricity from fusion at the lowest possible capital cost [38].

Advanced Research Projects Agency-Energy (ARPA-E) and Office of Science–Fusion Energy Sciences (SC-FES) launched a joint research program, so-called Galvanizing Advances in Market-aligned fusion for an Overabundance of Watts (GAMOW), which will prioritize R&D in (1) technologies and subsystems between the fusion plasma and balance of plant, (2) cost-effective, high-efficiency, high-duty-cycle driver technologies, and (3) cross-cutting areas such as novel fusion materials and advanced and additive manufacturing for fusion-relevant materials and components. Fourteen projects are supported through this program in total [39].

To support research in artificial intelligence (AI) and machine learning (ML) for fusion energy, the US DOE announced a plan to provide up to $21 million. Planned funding for projects will be invested over the next 3 years [40].

The development project of MPEX (Material Plasma Exposure eXperiment), a next-generation linear plasma device that will support the study of long-term plasma-material interaction in future fusion reactors, in particular, at the divertor, which is the power and particle exhaust system, has also been started [41].

2.4 China

China is promoting the development road map with the goal of the construction of the China Fusion Engineering Test Reactor (CFETR) in the 2020s and operating the first phase of the CFETR in the 2030s [42]. It is known that the engineering design of the CFETR has been completed in 2020 and is currently pursuing safety regulation and tritium regulation research and technology development.

Comprehensive Research Facilities for Fusion Technology (CRAFT), a comprehensive research facility for CFETR construction support and fusion technology development, is under construction with the goal of completion in 2025 [43]. At CRAFT, superconducting magnet research facilities (material, conductor, magnets), cryogenic testing facility, power supply testing facility, large linear plasma testing facility, CFETR development facilities, EAST divertor upgrade facility, heating and current drive research facilities (negative NBI, ECRH, LHCD, ICRF), and remote handling testing facility will be built. About 6 billion CNY is expected to be invested in the construction of the CRAFT on a site of about 0.4 km2.

EAST, operated by ASIPP (Plasma Physics Research Institute under the Chinese Academy of Sciences), is currently in the process of upgrading its divertor to develop a longtime high-performance scenario for CFETR and support CFETR [44].

HL-2M of SWIP (Southwest Physics Research Institute under China Atomic Energy Corporation) completed the upgrade from HL-2A to HL-2M in December 2020 [45]. With a major radius of 1.78 m and a magnetic field of up to 2.2 T, the machine can reach a plasma current of up to 2.5 MA with a plasma ion temperature Ti up to 13 keV.

2.5 Korea

Korea announced a National Fusion Roadmap released in 2005, and Fusion Energy Development Promotion Law (FEDPL), the first legal act in the world for fusion energy development, was enacted in 2007 to promote long-term cooperative fusion research and development among participating industries, universities, and research institutes as shown in Fig. 1 [46]. To meet this target schedule, a conceptual study of K-DEMO has been started in 2012 to explore the parameters and operational capabilities of Korean fusion demonstration reactor (K-DEMO) and prepare the basis for supporting the activity identifying key research and R&D directions and provide the vision for the Korea fusion power plant program. The first comprehensive summary of the conceptual study of the K-DEMO reactor was published in a scientific journal [47]. In December 2019, three volumes of intermediate conceptual study reports were published (written in Korean) [48]. The main purpose of the publication of the conceptual reports was to prepare for the regulation and licensing of fusion energy. Therefore, the study for the intermediate reports was largely extended to include safety, regulation, and environment. The first volume is “physics and operation,” which deals with an overview of DEMO reactor research, plasma physics for K-DEMO design, heating and current drive, and operation. The second volume is “K-DEMO tokamak and plant system,” which deals with the main components of the K-DEMO tokamak (superconducting magnet, diverter, breeding blanket, fusion reactor material, vacuum vessel and cryostat, remote maintenance) and plant system. The third volume is “safety, regulation, and environment” which deals with safety regulation, licensing, and environmental impact of K-DEMO. Note that the specifications of the K-DEMO reactor in the reports and plant systems are not clearly defined as they are in the design concept research stage. Therefore, many specifications, assumptions, and calculations were borrowed and/or extrapolated from the results of the medium-sized tokamaks and ITER. Final conceptual study reports were published in October 2021 (also written in Korean) [49].

Fig. 1
figure 1

Basic promotion plan for fusion energy development in Korea [36]

Full size image

Recently, the Korean government has set out to establish a long-term R&D roadmap to demonstrate fusion power production in 2050. The Ministry of Science and ICT announced the finalization of the 4th Basic Plan of nuclear fusion energy development (2022–2026) [50]. To systematically promote long-term and large-scale fusion R&D, a solid concept of the K-DEMO reactor will be established in 2022, and then, a long-term R&D road map including essential core routes will be derived by 2023.

In parallel to the K-DEMO design activity, Korea Superconducting Tokamak Advanced Research (KSTAR) had focused on the development of DEMO relevant long-pulse steady-state operation scenarios. KSTAR achieved the world’s first 100 million degree (ion temperature) ultra-high-temperature plasma operation for 30 s in November 2021 by improving the performance of the internal transport barrier (ITB) [51]. Furthermore, the operation of the new supercomputer Kairos (1.56PF) has begun. Kairos is the largest supercomputer in Korea dedicated to the fusion simulation to develop simulation codes for the interpretation and prediction of experiments related to ITER and K-DEMO [52].

2.6 Private sectors

The MIT Plasma Science & Fusion Center in collaboration with private fusion startup Commonwealth Fusion Systems (CFS) has been developing a compact but with high magnetic field tokamak by using high-temperature superconducting magnet technology [53]. The SPARC could produce 50–100 MW of fusion power, achieving a scientific Q fusion gain greater than 10. A conceptual design of the SPARC has been made with a 1.65 m major radius and 0.5 m minor radius operating at a toroidal field of 12 T and a plasma current of 7.5 MA. The SPARC plans to demonstrate net energy for the first time by about 2025.

A new-generation high-field spherical tokamak ST40 at Tokamak Energy Ltd. (TE) in Oxford, UK, is newly upgraded in 2021 and on track to make a hot plasma of a temperature of 100 million degrees [54]. The ST40 aims to demonstrate burning plasma condition parameters (nTτE) and may also be suitable for D-T operations in the future. ST40 has main machine design parameters such as a major radius of 0.4–0.6 m, operating at a toroidal field of 3 T (LN2 cooled copper), a plasma current of 2 MA, and κ = 2.5. TE is also developing high-temperature superconducting magnet technology for the future version of compact spherical tokamak to achieve first electricity by 2025.

General fusion was established in 2002 based in Vancouver, Canada, with locations in London, UK, and Oak Ridge, Tennessee, USA. General fusion utilized forceful and precisely shaped pistons around a vacuum vessel to compress liquid lithium symmetrically within tens of milliseconds. Fusion occurs inside the compressed liquid lithium vortex. The company recently announced they have achieved successful performance commission of plasma compression prototype in early January 2021. General fusion emphasized that “achieving this milestone with the prototype significantly reduces engineering and technical risks for General Fusion’s fusion demonstration plant” [55]. Recently, the UK Atomic Energy Authority (UKAEA) and general fusion announced an agreement to build and operate Fusion Demonstration Plant (FDP) in Culham, UK. The construction will begin in 2022, and the operation of the FDP is planned to be in 2025 [56].

TAE Technologies was founded in 1998. The company aims to develop commercial fusion power with the cleanest environmental profile and represents the fastest, most practical, and economically competitive solution to bring abundant energy to the grid. TAE is based in California and operates international offices in the UK and Switzerland. TAE technology is based on the advanced accelerator beam-driven field-reversed configuration (FRC) for compact and efficient operations. TAE showed in experiments that the machine has already reached more than 50 million °C. A new prototype will operate at more than 100 million °C. TAE Technologies works closely together with Google to solve issues arising from the experiments. By using Google’s vast computing power and machine learning, TAE Technologies has cut down tasks that once took 2 months to just a few hours [57]. TAE Technologies predicts the commercialization of its reactors within 10 years [58].

3 Discussion on key factors of a DEMO reactor design

Typically, a demo reactor design starts with 0-D/1-D tokamak system design code by solving a set of simplified physics equation and finding solutions to meet the requirements of a DEMO reactor as an energy source. There are several system codes to design a DEMO reactor such as SUPERCODE [59], GASC [60], TESC [61], ARIES system code [62], HELIOS [63], TPC [64], PROCESS [65], and SYCOMORE [66]. By scanning key parameters such as R0 and a (major/minor radius), BT (toroidal magnetic field), IP (plasma current), q95 (safety factor), κ (elongation), δ (triangularity), n/nGr, (ratio to Greenwald limit), βN (beta normalized), Q (= Pfus/Pinput), n(0)/< n >, T(0)/< T >, τp*/τE (ratio of confinement times), ηCD (current drive efficiency), and fimp (impurity fraction), one could obtain net electricity, thermal energy, energy/S, neutron wall load, etc., maximum and minimum design values as output parameters. Note that the design parameters obtained by those system codes are sets of possible parameters from the plasma physics point of view, and several other factors must be carefully considered because the construction and operation of a DEMO reactor will be critically dependent on the engineering, safety limits, and regulations. Table 1 shows key design parameters of major fusion machines and DEMO reactors proposed by the fusion community in the world [47, 49, 67,68,69,70,71,72,73].

Table 1 Key design factors of major fusion machines and DEMO reactors [47, 49, 67,68,69,70,71,72,73]
Full size table

In this section, we would like to discuss the key factors listed in Table 1 from the engineering and safety point of view.

3.1 Factors related to device size — major (R0) and minor (a) radius

If the main purpose of a tokamak in a research facility is to study the physics of fusion plasma and to gain corresponding engineering technology, there will be no specific size-limiting factor other than the construction and operation costs in principle. Plasma confinement time increases as the size of the fusion device increases, which makes it easier to achieve high-performance plasma conditions. The increase in the size of tokamak results in the increase of the overall size of engineering components such as vacuum vessel, superconducting magnets, heating, and current drive devices. As a trade-off for the size increase, the construction and operation costs increase. Furthermore, vacuum vessel and superconducting magnets larger than ITER will be very challenging to fabricate, transport, and assemble.

Different from current fusion machines, ITER and DEMO reactors have a necessary condition in the confinement time to achieve the critical value because those facilities must produce a certain level of fusion power as output. The in-vessel components — first wall and blankets — will be heavily bombarded by 14 MeV high-energy neutrons resulting in the activation of the materials. During the lifetime of a DEMO reactor, in-vessel components have to be replaced every 2 to 5 years depending on the neutron damage level of the materials [74]. Consequently, the generation of radioactive wastes from the replacement of in-vessel components will make another issue [75]. The larger the machine is, the more the radioactive waste will be. Therefore, even though they will be classified as low-level radioactive wastes, much larger hot cell and disposal sites will be inevitable.

3.2 Factors related to toroidal magnetic field (BT)

The strength of the magnetic field has a direct relationship with the size of the fusion tokamak. To increase plasma volume and confinement time, a larger size tokamak can be designed and built as described in the previous section. On the other hand, the confinement of ions in the plasma can be improved by increasing the strength of the toroidal magnetic field; hence, fusion power increases [76]. Nonetheless, two factors limit the use of high magnetic fields:

  1. 1)

    High magnetic field can be achieved either by applying a much higher current or much more windings of magnetic coils. Nevertheless, all conductors have their maximum current limit, and if excessive currents would be applied to the superconducting coils, the thermal quench will occur which transforms the superconducting state into a normal conductor state at a small local place inducing serious damage to the coil [77]. Furthermore, since the spaces for the superconducting magnets in a DEMO reactor are finite and due to the remote maintenance, poloidal coils can be positioned at unfavorable locations which need a much higher current to match the required magnetic field strength to control plasma.

  2. 2)

    Depending on the winding methods and configuration of cable in conduit conductors (CICC), magnetic ripples may affect the plasma condition, and the mechanical stress on the supporting structures may exceed the limit of the allowable stress of the structure materials. For instance, the static allowable values for SS316LN at 4K are 547 MPa for membrane stress alone and 820 MPa for membrane and bending stress [78].

3.3 Factors related to plasma shape — κ, δ, and double/single null

In the case of tokamaks for research purposes, the selection of the design parameters was made to obtain favorable conditions for high-performance plasmas with high flexibility of research under a limited budget. For instance, KSTAR aimed to have a high elongation (κ) and a triangularity (δ) to obtain a high plasma beta value through improved safety for the ideal MHD ballooning mode [79]. On the other hand, a high triangularity value can also contribute to the stabilization of the edge localized mode.

Unlike other DEMO reactor designs, the double-null plasma shape was adopted as the baseline mode of the K-DEMO reactor. With the double-null divertor configuration, the heat load applied to the diverter targets can be reduced by approximately half through a top-down symmetric structure compared to a single-null diverter configuration. However, the most important constraint that a DEMO reactor must secure is the self-sufficiency of tritium which requires a global tritium breeding ratio (TBR) greater than 1.05 [80]. TBR will be proportional to the volume occupied by the blankets inside the vacuum vessel. In the case of the double-null divertor type, the volume occupied by the upper and lower diverter structures will be twice larger than the single-null divertor resulting in reducing the global TBR [81, 82]. In addition, the remote maintenance strategy of single-null and double-null divertor will be different [83]. Therefore, the plasma shape in a DEMO reactor has to be determined in a way, considering not only the plasma performance but also the strategies on how to increase the global TBR, how to remove heat load from the divertor, and how to perform the remote maintenance.

In addition to the global TBR, the divertor design concept will be affected by the power conversion strategy, whether the heat exchange in the diverter is to be used for power generation [49]. If RAFM steel is to be used as cooling pipe material, the allowable temperature is less than 550 °C, and the divertor cooling water temperature is within the pressurized water reactor (PWR) condition. However, the cooling tube wall has to be very thin, and it will cause a problem in terms of mechanical strength. If copper alloy (CuCrZr) is to be used as in the case of ITER, the temperature of the divertor cooling water must be kept below ~150 °C due to the operating temperature limit (< 350 °C). In that case, the cooling water temperature does not satisfy the PWR condition, meaning that the diverter cooling water cannot be used as the input water of the steam generator. Consequently, the overall thermal efficiency of the plant will be low. Furthermore, the use of dual cooling systems — one for blankets and the other for divertor — will make the engineering of the cooling system more complicated.

3.4 Factors related to plasma current

As the magnitude of the plasma current increases, it contributes to the improvement of plasma performance such as confinement time [84]. To increase the plasma current, the size of the CS coil must be enlarged, which increases the overall size of the tokamak. Furthermore, the capacity of the heating and current driving device must be increased to achieve and maintain the noninductive current drive at high plasma current. With a higher magnetic field, the heating and current drive based on electromagnetic waves may have difficulties penetrating deeply into the plasma and to couple the power effectively. Therefore, the overall cost must be increased significantly.

Moreover, the safety factor q = RaBT/R0BP α Ra2BT/R0IP decreases as the plasma current increases resulting in MHD instability leading to plasma collapse. Such disruptions with high plasma current must be avoided in a DEMO reactor to protect the machine including in-vessel components; thus, a robust disruption mitigation strategy must be set, and a disruption mitigation system has to be equipped within the plasma control system.

4 Open questions and issues on the construction and operation of a DEMO reactor

In the previous section, we have briefly discussed a few key factors of a DEMO design in terms of engineering and safety issues. As mentioned above, most DEMO reactor designs were focused on the physics side only, but not seriously considered the restrictions in engineering, safety, and regulation. In this section, we will discuss the questions and issues on the construction and operation of a DEMO reactor such as (1) identity, (2) net electricity output, (3) construction cost, and (4) public acceptance of fusion energy. Those questions are not simple to answer, but they must be answered as soon as possible along with the DEMO design activity to contribute to the 2050 carbon neutrality in time.

4.1 Identity of a DEMO reactor

Most DEMO reactors designed in the past shown in Table 1 will generate net electricity outputs of 0.5–0.7 GW (2–3 GW thermal power) to fulfill the condition of competitive cost of electricity (COE) as they are aiming for commercial services. A competitive COE under commercial services assumes that fusion energy can contribute to the market as baseload electricity very soon. If a DEMO reactor will be a facility providing commercial services, the following issues must be clarified:

  1. 1)

    Is a fusion DEMO reactor defined and registered in the nuclear regulation or will be defined separately in a “fusion regulation”?

There is no definition of a “fusion reactor” in the nuclear regulation of any country, yet. Mostly, an experimental fusion device has a nuclear license as a radiation generator [49], and ITER has been licensed to be built as a nuclear facility in France [85]. A fusion DEMO reactor must be registered within the nuclear regulation, but current nuclear regulations are based on the construction, operation, and maintenance of fission reactors, which are not relevant to fusion DEMO reactors in many ways, e.g., pressurized vessel vs vacuum vessel, fission fuel (uranium) vs fusion fuel (tritium), etc. If a fusion regulation would be set very soon, will it be completely different from the fission nuclear regulation? Recently, the UK government proposed the regulation of fusion energy in the UK to lead the development of international fusion standards and regulations [86]. To start the commercial service utilizing fusion energy, the fusion regulation must be ready.

  1. 2)

    Do we have all technologies ready to build a DEMO reactor and to make it commercial?

Commercial service means “everything is ready to sell.” We all know that commercial service will be immediately fail if it is not ready to provide the proper services expected from customers. The term “all technologies” in the question mean not only technology related to constructing and operating a tokamak but also the power conversion system in the balance of plant (BoP), the remote handling technology, the fuel cycle technology including tritium extraction, the technology connecting fusion grids to existing power grids, the marketing, the disposal of radioactive wastes from first walls, etc.

Furthermore, is the code and standard for the construction of a DEMO reactor ready? A code and standard is an industry standard to be kept for commercial uses of stuff for safety and quality such as ASME (American Society of Mechanical Engineers) [87] and RCC-MRx [88]. Material properties, methods of fabrication and preparation, treatment, safety regulations, etc. are present in the code and standard. For a commercial fusion reactor, all material properties including 14 MeV irradiation data must be presented in the material database of ASME and/or RCC-MRx code to obtain a license for the construction.

After obtaining the construction license, a license for the D-T operation (the use of tritium) has to be approved [89, 90]. Similar to the case of ITER, the total amount of tritium stored in the fuel cycle and inside the vacuum vessel, and on-site inventory, has to be determined and regulated by a nuclear (fusion) authority [91, 92]. Please note that a DEMO reactor producing 1 GW net electricity will consume 12 kg of tritium per month [93, 94]. Therefore, tritium must be produced from the beginning of the DEMO reactor operation, as early as possible with sufficiently high TBR for the commercial services. As indicated above, the tritium self-sufficiency is one of the fundamental requirements of DEMO reactor; thus, the tritium breeder technologies including full fuel cycle, e.g., manufacturing of breeder materials, tritium extracting, cooling of blankets along with BoP functions to produce electricity, the remote maintenance, and the hot-cell operation strategy, must be ready for the full commercial operation.

  1. 3)

    Finally, the COE of fusion energy has to be clearly shown as an attractive commercial service.

If a DEMO reactor cannot provide commercial services, i.e., above questions (1), (2), and (3) cannot be fulfilled yet, a DEMO reactor has to be declared as a research facility. Then, we need to answer following the questions:

  1. 1)

    What research topics remain before the commercialization?

We already know that there are lists of technological gaps between current machines including ITER and a DEMO reactor such as burning plasma control, particle and heat flux handling at divertor, tritium fuel handling including tritium breeding, extraction, and power conversion in blankets, remote maintenance. And this question is directly connected to the second question.

  1. 2)

    Why do we need another research machine other than ITER?

Those two questions are not simple to answer. ITER is an integrated experimental facility aiming to test the engineering feasibility of fusion energy, identifying and filling the technological gaps needed for DEMO reactors. Nonetheless, it is not clear at this moment whether the ITER could fill entire scientific and technological gaps towards the DEMO reactor timely before the zero-carbon neutrality in 2050. For instance, since ITER test blankets will not be usable in any DEMO reactor, the blanket technology for DEMO reactors has to be developed separately. Furthermore, there is no volumetric 14 MeV neutron source in the world which emits a neutron flux comparable to a DEMO reactor for any test of breeding blanket functions. Although IFMIF plans to investigate tritium breeding in low fluence module region [95], the neutron irradiation zone is only 20 cm width and 5 cm height, which is even smaller than one breeder unit of a size of 20 cm width and 20 cm height designed for EU HCPB test blanket module system in ITER [96].

Those are the reasons why the DEMO reactor design activities assume that there might be a period for further research and development at the beginning of a DEMO reactor operation. For instance, the natural path assumed in Korea was to follow the technological development of KSTAR, ITER, and then finally a DEMO reactor. Therefore, K-DEMO suggested two-phase approaches (see Fig. 2) to mitigate risks in the course of K-DEMO technology development [49]. In its first stage, K-DEMO will be a technology test facility for the preparation of a commercial reactor. Operation stage I assumes a fusion power PFusion = 2200 MW but is not considered as a final DEMO, yet. Several ports will be designated for diagnostics relevant to burning plasma operations, and at least one port will be designated for the Component Testing Facility (CTF) including blanket and material tests. The main goals of this phase are to demonstrate the net electricity generation (Qeng > 1) and the self-sufficient tritium cycle (TBR > 1.05). After about 10 years’ operation of the first-stage K-DEMO, a major upgrade of the machine, which involves a replacement of in-vessel components including blankets, divertor system, and affected interfacing systems and services, will be followed, then, the second-stage K-DEMO operation will be started. In operation stage II, the main goals are to demonstrate the net electricity generation > 400 MWe and to demonstrate the competitiveness in COE with fusion power of PFusion = ,000 MW. This brings the third question.

  1. 3)

    Could a DEMO reactor licensed as a research machine be “upgradable” or “modifiable” as a commercial reactor?

Fig. 2
figure 2

Two-phased approach in K-DEMO development in Korea [37, 39]

Full size image

As described in the K-DEMO case, is there any possibility that we design a DEMO reactor for research purposes and then upgrade and/or modify it as a commercial facility later? The licensing process of research and commercial facility might be very different country by country. But in general, approval of an upgrade and/or modification of a nuclear reactor would be very difficult: For instance, licensing process of a small fission reactor for research purposes is the same as that for a commercial reactor in Korea [97]. It means, even though the reactor will be used for research, all nuclear safety and regulations must be fulfilled, and the reactor control system, as well as safety features, must be equipped as in the case of a commercial reactor. Furthermore, an upgrade or modification of a licensed reactor is not allowed in general, because a small change of the neutron irradiation condition will have a large impact on many aspects of safety, operation of the machine itself, and surrounding buildings and environments. Therefore, any modification, even though it would be very minor, is strictly controlled by nuclear authority [97].

  1. 4)

    Do we have enough time for the development AND can contribute to 2050 carbon neutrality?

Construction of a medium-size fusion machine usually takes about 5 years, if a construction site is already assigned. It takes much longer in the case of larger machines, about 10 to 15 years. For instance, JT-60SA began to manufacture the twisted superconducting wires for poloidal coils started in October 2008 as the first equipment, and the assembly of the cryostat base was started in January 2013. The assembly of the JT-60SA tokamak main body was completed in March 2020. In the case of ITER, land clearing and leveling were started in 2007, and construction of the ITER plant and auxiliary buildings for the first plasma is scheduled from 2010 to 2021, and the first plasma will be in Dec 2025. The main mission of both machines is to deliver scientific and engineering inputs needed for a DEMO reactor, not producing net electricity. To contribute to the 2050 carbon neutrality in time, plans for the design, construction, and operation of a fusion power plant must be established as soon as possible.

4.2 Net electricity output from a DEMO reactor

It was never discussed seriously before, but how much net electricity output would be expected from a DEMO reactor? Most of the DEMO reactors designed so far have assumed an electricity output of 1 GW as a default value [47, 67,68,69,70,71,72,73, 76]. Theoretically, the net electricity to the grid increases as the fusion power increases and as the recirculating power decreases. A DEMO reactor could produce net electricity up to 700 MW out of 3 GW fusion power including about 300 MW of recirculating power (assuming Rankine cycle and 30% conversion efficiency). At this point, we need to identify what the output power from a fusion DEMO reactor means, not just as a design value but also as a contribution to the market.

A 1 GW stable electricity output from any power source means that it can be one of the baseload electricity providers. The baseload is practically the minimum level of unvarying power over a period providing electricity with high availability and reliability at low cost [98]. Therefore, it has to be carefully examined, whether a DEMO reactor could provide such a commercial service from the beginning of its operation, although it will be the first kind. Furthermore, as indicated above, the tritium fuel cycle including TBR must be confirmed as the primary requirement.

If a DEMO reactor would produce net electricity less than 1 GW, how much output power would be acceptable? The USA recently proposed a small reactor based on high-temperature superconducting magnets that could produce 50 MW of output power [99]. The case of the first commercial fission power plant — the Calder Hall nuclear power plant — had four Magnox reactors, and each reactor produced 50 MW making a total of 200 MW of output power. As in the case of a fission power plant, a fusion power plant consisting of several small fusion reactors producing a net output power of 1 GW might be possible. Of course, the construction cost must be considered seriously.

4.3 Construction cost of a DEMO reactor

Although ITER does not produce any net output power, the construction cost of ITER is known as already US $20 billion [100]. How much should be the construction cost for a DEMO reactor? Since the construction cost of a commercial power plant is directly connected to the COE, the cost must be minimized. If the construction cost of a DEMO reactor would be similar to that of ITER and produce 1 GW of output power (100–200 MW net electricity), would it be acceptable?

Recently, a study on the construction cost estimation of a DEMO reactor has been performed [101]. Authors have carried out a broad range of parametric studies of tokamak fusion reactors utilizing a coupled system analysis that included a simple cost model. After the analyses of a total of 2,700,000 cases, authors claim that a COE between 109 and 140 mills/kWh with a direct capital cost between US $5 and 6 billion, and net electricity between 1000 and 1600 MW was possible (with the central solenoid design), which is comparable with that of average electricity cost in the USA (125 mills/kWh).

As a comparison, the construction costs of fission reactors in various countries are as follows: 1.6 GW reactor costs 9 billion USD in France [102], 1.1 GW reactor costs US $4.5 billion in China [103], and 1.2 GW reactor costs US $7 billion in Russia [104].

4.4 Public acceptance of fusion energy

The public is very sensitive to the word “nuclear” and “radioactivity” and afraid of the radiation from activated materials and fuels. Especially, many people have a concern about tritiated water. A new facility treating radioactive fuels, thus producing radioactive wastes will be a very sensitive issue to the public if it would be constructed and operated nearby.

Fusion energy is known to the public as clean and limitless energy, producing no long-lived radioactive wastes. Nevertheless, ITER is already declared as a nuclear facility in France, although it has very limited use of tritium and neutron yields. As a nuclear facility, a DEMO reactor will utilize more than a hundred kg of tritium per year with an order of magnitude higher neutron flux than that of ITER and, thus, will produce a large number of low-level radioactive wastes from the replacement of blankets and divertors. Those wastes must be stored in hot cells over 100 years before they are to be treated and stored in an outside disposal site [105]. Those factors are not commonly known to the public.

Gaining public acceptance for fusion power plants would be similar to that of fission reactors: The most important concern will be radioactive wastes and nuclear safety [106]. Public acceptance varies from place to place, from time to time, and from country to country. For instance, radioactivity is one of the major social issues in Japan. Therefore, the Japanese stellarator in Toki, Japan, LHD (large helical device) has started its D-D operation about 20 years after the first plasma was conducted in March 1998 due to poor public acceptance [107, 108]. On the other hand, KSTAR in Daejeon, Korea, took only 2 years to perform the deuterium campaign in 2010 after the first plasma in July 2008 with collaborative public acceptance [109].

To get a clear public acceptance for the construction and the operation of DEMO reactors, long discussions and public hearings will be needed.

5 Two-pathway approach towards the realization of fusion energy

ITER is following the traditional pathway for the development of fusion energy aiming for 500 MW fusion power. The design of ITER is based on physics and experimental evidence obtained from several decades of research at a large number of devices in the world. Therefore, ITER will be the final test field of realization of fusion energy. This pathway must be completed as it was planned. We have no doubt that ITER will be successfully operated and demonstrated its goals including the burning plasmas, as we have witnessed the successful operation of 59 MJ D-T plasma at JET [110].

On the other hand, as we have discussed in Section 2, the USA and UK move forward to build small reactors using high-temperature superconducting magnets. Recently, MIT announced that they have successfully developed, for the first time, a 20 T large high-temperature superconducting magnet [111]. This pathway would be an alternative way to bring fusion energy much closer and faster to our everyday life.

5.1 Small-size DEMO reactor

As we have indicated at the beginning of the paper, we are entering now into the zero-carbon emission era, meaning that fusion energy has to contribute to the electricity market as soon as possible. We have discussed that DEMO reactors larger than ITER size would produce 1 GW of electricity. From the engineering point of view, DEMO reactors do not need to be the baseload electricity providers immediately after their birth. As in the case of the first commercial fission reactor, we could start the commercialization of fusion energy with 50–100 MW small DEMO reactors. Small reactors will produce around 50–200 MW of electricity, which could serve as a power source covering the intermediate peaking load. The operation of a DEMO reactor for the intermediate peaking load allows the operation scheme in both steady and pulsed operation discussed as the baseline operation scenarios. Note that the intermediate peaking load in a day would be from a couple of hours up to about 6 h normally, which could exceed 12 h in summer and winter. In other words, DEMO reactors need to be operated for 2 to 6 h normally and should operate more than 12 h when needed. The small DEMO reactor concept is also well fitted to large countries like the USA and Russia, where many small cities are located far from power stations. A cluster of 4 or 6 small DEMO reactors can provide electricity up to 1 GW to support large cities. In this way, 75% of the availability of power plants can be easily fulfilled. Furthermore, mass production of components will be needed to maintain clusters of small DEMO reactors which will lower the construction costs and will allow the change of in-vessel components frequently to control activation level before they are too much activated. This will shorten the hot-cell storage time and recycling period. Finally, it will open a new market and populate related industrial companies as a new blue ocean.

5.2 Plasma operation and disruption avoidance strategy

Plasma operation in the DEMO reactor will not be flexible as it must yield a certain number of neutrons during plasma operation with high Q to produce the desired level of electricity. This means that the plasma operation in a DEMO reactor must be optimized in a way to maximize neutron production, i.e., maximize the tritium breeding as well as the power conversion. Nevertheless, there is an engineering limit of neutron flux (thus nuclear heating) towards structures, especially vacuum vessel and superconducting magnets. Those structures are not meant to be maintained or replaced during the whole lifetime of the DEMO reactor once they are placed. Neutron irradiation on the superconducting magnets causes not only the activation of the materials but also degradation of the magnetic field strength (decrease of critical temperature Tc, increase of critical current Jc) [112]. Furthermore, nuclear heating from fast neutrons (En > 1MeV) and decay heat from activated materials inside the superconducting magnets could cause an increase of local temperature resulting in the thermal quench. The limits of the fast neutron flux and nuclear heating on a toroidal superconducting magnet are 109 n/cm2s and 5 × 10-5 W/cm3 [113]. Superconducting magnets of small DEMO reactors are much closer to burning plasmas resulting in much higher neutron flux. Therefore, the engineering limit sets the upper limit of plasma operation accordingly, even though higher performance plasma could be achieved. Depending on the situation of the electric grid of the day, several different plasma operation scenarios aiming to produce different levels of neutron yields will be employed.

While maintaining high-performance burning plasmas, plasma disruption at high plasma current will be an unavoidable threat. Furthermore, the disruption in small reactors is more dangerous because the energy density is much higher, and they have smaller areas that could mitigate the plasma stored energy. Furthermore, in order to maximize the volume occupied by blankets, only restricted space will be available for real-time diagnostics. Once a DEMO reactor is to be built, it will have two different operation phases: (1) the commissioning phase in which all systems will be checked and (2) D-T operation to produce electricity. A “full set” of diagnostics installed on a dedicated port will be utilized for the initial commissioning and for the characterization of the plasma behavior in preparation for D-T operation. In this phase, AI learns the behavior of plasma by using the full set. Once commissioning is finished, the full set will be replaced with a “small essential set” of diagnostics. In the D-T phase, the PCS will rely on the small essential set, and AI will provide “missing” information learned from the previous phase to the PCS. Nevertheless, there is no absolute guarantee that PCS works perfectly without any fault, and thus, there will be no perfect way to avoid disruptions. Instead, we must focus on the safe landing of plasmas whenever an off-normal event triggers because machine/human safety is the most important factor in a commercial facility. Once PCS coupled with diagnostics and AI notices any sign of an off-normal event that is potentially hazardous for the machine’s safety, the plasma must be terminated immediately by the PCS interlock. Upon an inspection of the situation and machine status, it will start a new plasma. This disruption avoidance strategy may be acceptable for the intermediate peaking load, especially with a cluster of small DEMO reactors.

5.3 Tritium production and supplement strategy

The amount of tritium in the work would be around 24 Kg, and it will increase up to about 27 Kg and then will decrease rapidly as a function of time. Traditionally, tritium is produced by Candu reactors, and the number of Candu reactors in operation also decreases in time. The prediction shows that there will be around 5 Kg left in 2040 as ITER consumes it [114].

To operate a small DEMO reactor along with the operation of ITER, we need to set up a tritium production and supplement strategy at least until the DEMO reactors produce enough tritium for fuel self-sufficiency. As we already know, 6Li reacts with thermal neutrons due to its higher cross-section (En < 1eV). We may start to produce tritium by utilizing thermal neutrons from fission reactors by using breeding materials developed for blankets: For instance, capsule-type storage containing 6Li breeder pebbles could be placed in a dedicated high neutron flux area in a fission reactor. Furthermore, the USA and Russia have thousands of hydrogen bombs containing a small amount of tritium. In order to maintain those bombs, there are government weapons production plants that produce and store tritium for military use [115]. If we can access and use them for peaceful purposes to achieve carbon neutrality and to overcome the global warming which threatens the existence of humanity on Earth, we might have enough tritium for the startup of the clean and safe fusion energy era.

5.4 Remote maintenance and radioactive waste management

Remote maintenance of a DEMO reactor is a critical issue for commercial service. In-vessel components such as blanket modules and divertor have to be replaced periodically to maintain high plasma performance. DEMO reactors designed for 1 GW output in Table 1 have a maintenance period of 2.5 years for divertor and 5 years for blankets depending on the displacement per atom. As already known, the replacement of those in-vessel components in a tokamak device equipped with toroidal and poloidal superconducting magnets is a very challenging task due to the accessibility. There are several remote maintenance concepts developed for DEMO reactors, and it can be summarized as follows [116, 117]: (1) the sector transport using all horizontal maintenance ports, (2) the sector transport using the limited number of vertical maintenance ports, and (3) the sector transport using the limited number of horizontal maintenance ports. Those remote maintenance schemes need large horizontal and/or vertical ports and need to rotate toroidally one or two sectors inside the vacuum vessel, which is extremely challenging because each sector has a weight of more than 100 tons.

One of the advantages of high-temperature superconducting magnets which will be introduced to ARC is that the magnets could be demountable [118]. Through this technology, the main body of ARC can be divided into two sections, and the vacuum vessel and blanket structures can be replaced [118]. Note that the demountable and replaceable vacuum vessel concept makes the remote maintenance revolutionary easier than the conventional concept listed above. Furthermore, as we have mentioned above, the activation level of the structure materials can be actively controlled before they are too much activated. Replaced vacuum vessel and blanket structure will be transported and stored in hot cell. Finally, they are recycled as soon as they are classified as nonradioactive materials. In this way, the total amount of radioactive wastes that remain in the hot cell and disposal site can be reduced.

6 Conclusion

There were many DEMO reactor concepts suggested by many authors for long. However, most of the design concepts deduced by 0D/1D system codes assumed the best plasma performance and effective power conversion rate producing 1 GW of electricity. The most important goal to be achieved is we would like to solve climate change and to achieve “2050 carbon neutrality” by the contribution of fusion energy to the power grid system.

Following physics developed so far, a larger fusion reactor produces more fusion power. Therefore, the size of designed DEMO reactors is similar or larger than ITER as seen above. Nevertheless, due to the technological breakthrough in the field of high-temperature superconducting tokamaks, there might be a chance to reduce the size of a DEMO reactor significantly. Although such a DEMO reactor might produce smaller power than a large size reactor, clustering of 4–6 small reactors may produce 500–1000 MW net electricity. Clustering of DEMO reactors means that mass production of components would be possible, which lowers the construction cost of DEMO reactors and thus COE. Furthermore, the components can be much more frequently and cheaply replaced resulting in the reduction of the radioactivity of in-vessel components. This leads to a much shorter hot-cell storage time and gaining much broader public acceptance. Meanwhile, technological gaps to larger machines could be filled [118,119,120].