Chapter 1: Overview and summary - Columbia Universitysites.apam.columbia.edu/courses/apph4990y_ITER/Nuclear Fusion 1999...Chapter 1: Overview and summary ITER Physics Basis Editors

Chapter 1: Overview and summary

ITER Physics Basis Editors∗

ITER Physics Expert Group Chairs and Co-Chairs†

ITER Joint Central Team and Physics Integration Unit‡

ITER EDA, Naka Joint Work Site, Mukouyama,Naka-machi, Naka-gun, Ibaraki-ken, Japan

Abstract. The ITER Physics Basis presents and evaluates the physics rules and methodologies for

plasma performance projections, which provide the basis for the design of a tokamak burning plasma

device whose goal is to demonstrate the scientific and technological feasibility of fusion energy for

peaceful purposes. This Chapter summarizes the physics basis for burning plasma projections, which

is developed in detail by the ITER Physics Expert Groups in subsequent chapters. To set context, the

design guidelines and requirements established in the report of ITER Special Working Group 1 are

presented, as are the specifics of the tokamak design developed in the Final Design Report of the ITER

Engineering Design Activities, which exemplifies burning tokamak plasma experiments. The behaviour

of a tokamak plasma is determined by the interaction of many diverse physics processes, all of which

bear on projections for both a burning plasma experiment and an eventual tokamak reactor. Key

processes summarized here are energy and particle confinement and the H-mode power threshold; MHD

stability, including pressure and density limits, neoclassical islands, error fields, disruptions, sawteeth,

and ELMs; power and particle exhaust, involving divertor power dispersal, helium exhaust, fuelling

and density control, H-mode edge transition region, erosion of plasma facing components, tritium

retention; energetic particle physics; auxiliary power physics; and the physics of plasma diagnostics.

Summaries of projection methodologies, together with estimates of their attendant uncertainties, are

presented in each of these areas. Since each physics element has its own scaling properties, an integrated

experimental demonstration of the balance between the combined processes which obtains in a reactor

plasma is inaccessible to contemporary experimental facilities: it requires a reactor scale device. It is

argued, moreover, that a burning plasma experiment can be sufficiently flexible to permit operation

in a steady state mode, with non-inductive plasma current drive, as well as in a pulsed mode where

current is inductively driven. Overall, the ITER Physics Basis can support a range of candidate designs

for a tokamak burning plasma facility. For each design, there will remain a significant uncertainty in

the projected performance, but the projection methodologies outlined here do suffice to specify the

major parameters of such a facility and form the basis for assuring that its phased operation will

return sufficient information to design a prototype commercial fusion power reactor, thus fulfilling the

goal of the ITER project.

∗ ITER Physics Basis Editors: F. W. Perkins (ITERJCT), D.E. Post (ITER JCT), N.A. Uckan (ORNL),

M. Azumi (JAERI), D.J. Campbell (NET), N. Ivanov (RRC-Kurchatov), N.R. Sauthoff (PPPL), M. Wakatani (KyotoUniv.). Additional contributing editors: W.M. Nevins (LLNL),M. Shimada (JAERI), J. Van Dam (Univ. Texas).

† ITER Physics Expert Group Chairs and Co-Chairs:D. Boucher (ITER JCT), G. Cordey (JET), A. Costley(ITER JCT), J. Jacquinot (JET), G. Janeschitz (ITER JCT),

S. Mirnov (Troitsk), V. Mukhovatov (ITER JCT), G. Porter(LLNL), D. Post (ITER JCT), S. Putvinski (ITER JCT),M. Shimada (JAERI), R. Stambaugh (GA), M. Wakatani(Kyoto Univ.), J. Wesley (ITER JCT), K. Young (PPPL).

‡ ITER Joint Central Team and Physics IntegrationUnit: R. Aymar, Y. Shimomura, D. Boucher, A. Costley, N.

Fujisawa, Y. Igitkhanov, G. Janeschitz, A. Kukushkin, V.Mukhovatov, F. Perkins, D. Post, S. Putvinski, M. Rosen-bluth, J. Wesley.

Contents

1. Introduction 2138

2. ITER 21392.1. ITER: background and mandate . . . 21392.2. ITER: FDR design . . . . . . . . . . . 2140

3. Tokamak physics processes andprojection principles 21423.1. General projection issues . . . . . . . . 21433.2. Core confinement and transport . . . . 2145

3.2.1. Global confinement scaling . . 21463.2.2. H-mode power threshold and

pedestal . . . . . . . . . . . . . 2146

Nuclear Fusion, Vol. 39, No. 12 c©1999, IAEA, Vienna 2137

ITER Physics Basis

3.2.3. Transport modelling and sim-ulation . . . . . . . . . . . . . . 2147

3.2.4. Confinement and magneticconfiguration . . . . . . . . . . 2148

3.3. Magnetohydrodynamic phenomena,disruptions and operational limits . . . 21493.3.1. Magnetohydrodynamic stability 21493.3.2. Magnetohydrodynamic β lim-

its and neoclassical islands . . 21503.3.3. Error field criteria . . . . . . . 21513.3.4. Disruptions . . . . . . . . . . . 21513.3.5. Sawteeth . . . . . . . . . . . . 21523.3.6. Edge localized modes . . . . . 21533.3.7. Magnetohydrodynamics of

reverse shear and steady stateconfigurations . . . . . . . . . . 2153

3.3.8. Density limit physics . . . . . . 21543.4. Particle control and power dispersal . 2155

3.4.1. Power dispersal in divertorplasmas . . . . . . . . . . . . . 2155

3.4.2. H-mode pedestal and edgeoperational space . . . . . . . . 2156

3.4.3. Erosion of plasma facing com-ponents and tritium retention . 2157

3.5. Energetic particle physics . . . . . . . 21583.6. Auxiliary power physics . . . . . . . . 21583.7. Physics of plasma diagnostics . . . . . 21593.8. Physics of plasma control and steady

state operation . . . . . . . . . . . . . 21603.9. Summary . . . . . . . . . . . . . . . . 2161

4. Reactor scale experimentalplasma physics 21614.1. Energetic particle physics . . . . . . . 21614.2. Self-heating and thermal stability . . . 21624.3. Scale dependent plasma physics . . . . 2162

5. Projecting ITER operations 21645.1. Single pulse issues . . . . . . . . . . . 21645.2. Physics performance projections . . . 21655.3. Multiple pulse and erosion issues . . . 2166

6. Concluding remarks 2166

Appendix A 2167

Appendix B. Article 1, ITER EDA Agree-ment 2170

Appendix C. ITER Special WorkingGroup 1 – Review Report 2170Preamble . . . . . . . . . . . . . . . . . . . 2170General constraints . . . . . . . . . . . . . . 2170

Performance and testing . . . . . . . . . . . 2170Design requirements . . . . . . . . . . . . . 2171Operation requirements . . . . . . . . . . . 2171Final recommendation . . . . . . . . . . . . 2172

Attachment 1 – Guideline for SWG1 . 2172

1. Introduction

Magnetic fusion energy research has reached thepoint where a tokamak burning plasma facility inwhich the thermonuclear heating balances (or is com-parable to) transport and radiation losses for peri-ods of 1000 s or longer can be seriously contem-plated as an appropriate next step. Achieving thisgoal would be a major step forward, both in sci-ence and in technology, towards the ultimate goal ofmagnetic fusion generation of electrical power withsignificant environmental advantages [1, 2]. Overall,such a facility would have a size, magnetic fieldstrength, physics phenomenology, and technologicalbasis very close to that of an eventual thermonu-clear power reactor, be it a tokamak or some othertoroidal configuration. Indeed, three aspects of theinterplay between physics and technology are com-mon to a burning plasma experiment and a reac-tor. First, the general confinement properties of atokamak device that achieves such a thermal balanceimplies a power level of ∼1 GW and a neutron wallloading of ∼1 MW·m−2: levels in the range antici-pated for commercial power production. Second, inthe proposed tokamak configuration, be it a reac-tor or burning plasma experiment, the magnitude ofthe magnetic field needed to confine stably a plasmaof sufficient pressure to generate ∼1 GW of fusionpower is comparable to the limiting magnetic fieldsthat a toroidal superconducting magnet can produce.Third, the linear size of the plasmas is sufficientlylarger than the shield thickness needed to protectsuperconducting magnets from nuclear radiation, sothat the shield occupies only a modest fraction ofthe volume available inside the confining magnetsand does not dominate the design. Appendix A addsdetails to these arguments. As a consequence, datafrom such a burning plasma facility is foreseen torequire little extrapolation to an experimental powerreactor and is essential to defining its principal oper-ational mode. For example, if a steady state opera-tional mode is to be chosen for a commercial tokamakreactor design, then this choice must rest on a robustexperimental demonstration of steady state physicsand operation in a burning plasma experiment.

2138 Nuclear Fusion, Vol. 39, No. 12 (1999)


It is, therefore, noteworthy that, in the worldwidefusion research programme, tokamak experimentshave demonstrated a common plasma physics acrossa range of device sizes, magnetic field strengths, andauxiliary heating powers. This common physics pro-vides the basis for moving ahead with a burningplasma facility by permitting development of extrap-olation principles, both theoretical and empirical,and their application to the projection of burningplasma performance. It is the role of this articleto summarize and assess the qualitative and quan-titative aspects of tokamak physics and to developrecommended extrapolation methodologies togetherwith uncertainty estimates and physics design spec-ifications for use by the designers of the burningplasma facility, which is called ITER: the Interna-tional Thermonuclear Experimental Reactor.

Assessments of projections for plasma physics per-formance carried out by the seven ITER PhysicsExpert Groups in coordination with the ITERPhysics Basis Editors and the Joint Central Team(JCT) form the core of this article: Chapters 2–6. Chapter 7 assesses plasma measurement require-ments and the extrapolation of physics principles onwhich diagnostic techniques are based. The final twochapters look forward to issues impacting the opera-tion of a burning plasma facility and its experimentalphysics programme.

The ITER Physics Basis has been compiled andwritten by a collaboration of authors that is basedupon the seven ITER Physics Expert Groups, theITER Physics Basis Editors, and physics staff fromthe ITER JCT, supplemented in the various Chap-ters by physics and technology specialists drawn fromthe plasma research programmes of the ITER Par-ties. The Expert Group Chairs and Co-chairs andthe ITER Physics Basis Editors played a key role inthe final compilation and editing. Within their ownareas of expertise, each of the Expert Groups hasbeen evaluating progress and recommending priori-ties for physics research in the Four Parties physicsresearch programmes. Consequently, their membershave acquired the physics expertise and burningplasma perspective needed to develop and assess pro-jection methodologies.

The ITER/EDA procedure has been to basedesign choices on the physics principles discussed anddocumented in this article. ITER design issues anddecisions, which are the responsibility of the JCT,are documented in the physics chapter of the ITERFinal Design Report [3,4] and in the Physics DesignDescription Documents [5].

This introductory section is written as a sum-mary of the entire article and, as such, providesan overview and integration of the separate Chap-ters. To establish context, Section 2 will describethe ITER Agreement and the current EngineeringDesign Activities (ITER/EDA) as well as the specifi-cations that the device under design must fulfil. Thedesign parameters documented in the ITER FinalDesign Report (FDR) are presented as exemplify-ing reactor scale devices. It should be stressed thatthe projection methodologies reported in this arti-cle apply to a range of parameters and form a basisfor assessing tradeoffs associated with reduced costdesigns relative to the FDR design.

Section 3 summarizes the main content of thisarticle: the identification of the various physics pro-cesses in contemporary tokamaks and their projec-tion principles. Next, Section 4 argues that the dom-inant physics in a reactor scale facility will differ inimportant ways from that in present day devices.An example is the integration of core transport andedge physics. Our discussion organizes the differencesinto three elements and outlines the scientific knowl-edge that operation of a reactor scale facility willreturn. This Introduction concludes with an assess-ment of the physics projection methodologies sup-porting design of a reactor scale experiment.

2. ITER

The importance of the step to reactor scale devicesmotivated the governments of the Four Parties –the European Union, Japan, the Russian Federationand the United States – to initiate, in 1987, theInternational Thermonuclear Experimental Reac-tor/Conceptual Design Activities (ITER/CDA). Thepromise of the Conceptual Design, which was com-pleted in 1990 [6], led, in 1992, to the present ITEREngineering Design Activities (ITER/EDA) Agree-ment [7] aimed at developing a detailed engineeringdesign for a reactor scale tokamak facility that wouldachieve controlled ignition and extended burn. Asenvisioned by the Agreement, the ITER device wouldbe the central element of an international, ‘one stepto a reactor’ strategy.

2.1. ITER: background and mandate

The overall goal of ITER/EDA, as set forth inArticle 1 of the ITER Agreement (Appendix B), isto demonstrate the scientific and technological fea-sibility of fusion energy for peaceful purposes. Spe-

Nuclear Fusion, Vol. 39, No. 12 (1999) 2139

ITER Physics Basis

cial Working Group 1 (SWG1) was chartered by theITER Council to develop detailed technical objec-tives for the ITER design to assure that the designwould fulfil this overall goal. The report of SWG1 canbe found in Appendix C. This report makes it clearthat the device that results from the EDA should notonly achieve controlled ignition and extended burn inestablished favourable confinement modes, but alsoshould be sufficiently flexible to provide access for theintroduction of advanced features and new capabili-ties and to allow for optimizing plasma performanceduring operation. Steady state experiments shouldaim at a demonstration of steady state operation inplasmas having alpha particle heating power at leastcomparable to externally applied power. The choiceof parameters should be consistent with margins thatgive confidence in achieving the required plasma per-formance.

In brief, the ITER device is to be a flexible, reac-tor scale experimental facility capable of standardand advanced operating modes. It is envisioned to bethe world’s first reactor scale magnetic fusion exper-iment, and, as such, will be the first to combine theelements discussed above: a capability for achievingsustained ignition and extended duration fusion burnin DT plasmas with reactor relevant engineeringfeatures that include superconducting magnet sys-tems, remotely maintainable in-vessel nuclear shield-ing, and plasma facing components with steady statepower and particle exhaust capabilities.

2.2. ITER: FDR design

The approximate magnitude of the parameters foran ignited, reactor scale tokamak can be derived fromsimple arguments, which are set forth in Appendix Aand based on operation in the favourable ELMy H-mode confinement regime. In this regime, plasmaturbulent heat conduction spontaneously diminishesin a thin, transport barrier layer just inside the mag-netic separatrix. This layer is commonly observedto undergo successive relaxations called edge local-ized modes (ELMs). The interest in ELMy H-modesflows from experimental observations that show thatthis mode reduces transport throughout the plasmacore. The standard working hypothesis, supportedby many observations, is that H-mode occurs whenthe power transported across the separatrix exceedsa threshold value.

Table 1 and Fig. 1 present the specifics of theITER design, which follow from the arguments ofAppendix A and supporting detailed design calcu-

lations [4]. Since the arguments are straightforward,ITER truly exemplifies a tokamak reactor facility.Quantitative calculations based on parameters closeto those of Table 1 will be representative of any reac-tor scale tokamak facility with an ignition capabil-ity. These parameters fulfil a self-consistency checkthat the power transported through the separatrixexceeds the threshold power required to maintain H-mode confinement. Table 1 takes into account thefavourable isotope effect on threshold power con-firmed in recent JET DT experiments [8]. One notesthat the optimized ignition condition depends sen-sitively on plasma size and magnetic field strength(Eq. (A4) of Appendix A), so that fusion perfor-mance degrades for device sizes less than that ofthe FDR design. Increased magnetic field strengthcan restore performance loss resulting from decreasedplasma size.

Figure 1. Poloidal plane view of the ITER FDR design.

Closed curves in the plasma region depict magnetic sur-

faces; the confining magnetic field lies on these sur-

faces. The separatrix magnetic surface (single red con-

tour) defines the boundary between magnetic surfaces

that close within the plasma region and those that inter-

sect material walls.



Table 1. ITER design features and parameters for reference ignited ELMy H-mode operation

Parameter Value

Major/minor radius 8.14 m/2.80 m

Plasma configuration Single null divertor

Plasma vertical elongation/triangularity (at 95% poloidal flux) 1.6/0.24

Plasma volume ∼2000 m3

Plasma surface area ∼1200 m2

Nominal plasma current 21 MA

Electron density 0.98 × 1020 m−3

Volume average temperature 12.9 keV

Toroidal field 5.68 T (at R = 8.14 m)

MHD safety factor (q95) ∼3.0 (at 21 MA)

Volume average β/βN 0.030/2.29

Fusion power (ignited, nominal) 1.5 GW

Plasma thermal energy content 1.07 GJ

Plasma magnetic energy content 1.1 GJ

Confinement mode ELMy H-mode

Radiation from plasma core 118 MW

Transport power loss 182 MW

Transport energy confinement time τE 5.9 s

Ptransport/PL→H 1.4

Species concentrations % He/Be/Ar 10/2/0.16

Zeff (effective ion charge) 1.9

Average neutron wall loading ∼1 MW·m−2 (at 1.5 GW)

Lifetime neutron fluence ≥1 MW· a·m−2

Burn duration (ignited, inductive current drive) ≥1000 s

Available auxiliary heating power 100–150 MW

In vessel tritium inventory safety limit 1 kg

While ITER is designed to ignite, i.e. to produceenough fusion power to overcome heat losses, auxil-iary power is required to initially raise the plasmatemperature as well as for control and current drivepurposes. Auxiliary heating power in the range 50–150 MW can take the form of negative ion based1 MeV neutral beam injection, ion cyclotron heat-ing by the fast magnetosonic Alfven wave, and elec-tron cyclotron heating. These auxiliary heating sys-tems also possess a current drive capability and elec-tron cyclotron heating is notable in that its cur-rent drive can be utilized for current profile control.Lower hybrid current drive is also under study forlater investigations of steady state operation. Neutralbeam injection is unique in its capability to introduceangular momentum.

Figure 1 presents a poloidal plane view of theITER facility and Fig. 2 gives representative den-sity and temperature profiles for an ignited ITERdischarge. Because there is essentially no ionizationoccurring inside the separatrix, the density profileis flat. Any density gradient close to the separatrixwould be sensitive to details of fuelling and not essen-tial to performance calculations, which rest on core

thermonuclear and auxiliary heating as well as coretransport.

How big a step is the ITER FDR device? Fig-ure 3 compares the fusion figure of merit M =nDT (0)Ti(0)τE for present day tokamaks with thevalues computed for ITER under minimum ignitionconditions, which require M ≈ 110. Here, τE denotesthe thermal energy confinement time in s, nDT (0)the central DT fuel density in units of 1020 m−3,and Ti(0) the central ion temperature in keV. ITERFDR parameters lie a factor of 1.5 in magnetic fieldstrength, and a factor of 2.9 in linear size, beyond thelatest JET DT ELMy H-mode discharges [9]. Theincrease in the figure of merit from ITER like dis-charges in present day devices to ITER is apprecia-ble (a factor of 40), but comparable to the range ofM spanned by ITER like ELMy H-mode dischargesin present day experiments (also a factor of 40). It islikewise in accord with the expected increase result-ing from increases in magnetic field strength and size,according to Appendix A, Eq. (A5).

Tokamaks have already entered the regime ofthermonuclear burning [8–12]. Figure 4 summarizesresults from JET and TFTR. These experiments doc-


ITER Physics Basis

6 7 8 9 10M ajor radius (m )

0

5

1 0

1 5

2 0

2 5

3 0

3 5

Figure 2. (a) Ion temperature profile (keV) correspond-

ing to the plasma of Table 1. Electron and ion tempera-

tures are close to equal.

ne

nD + nT

n He

Figure 2. (b) Electron, DT ion and He density profiles

(1019 m−3) for the plasma of Table 1.

ument that the expected heating from thermonuclearalpha particles is occurring. We note that long pulseELMy H-mode results from JET are limited by theavailable auxiliary heating power, which is insuffi-cient, at a toroidal field strength of 3.8 T, to reach

0.01

0.1

1

10

100

1000

1 10 100

JETJETTFTRTFTRDIII-DDIII-DC-ModASDEX-UJT-60UITERTEXTOR

Ti ( 0 ) - (keV)

n (0

) T

(0)

τi

iE

- (1

0 m

• k

eV•

s) -

320

Figure 3. Fusion figure of merit M = ni(0)Ti(0)τE

for selected tokamak discharges. Filled symbols repre-

sent steady discharges with Te ≈ Ti in H-mode, except

for TEXTOR, which is in a radiation enhanced mode,

and for TFTR in a pellet fuelling mode. Open symbols

represent confinement modes with Ti Te, which have

been optimized for fusion output. The ITER point rep-

resents the minimum M for steady, ignited burn and is

insensitive to Ti(0) because of β limits.

βN values characteristic of a reactor. Figure 3 alsocontains points at higher values of M based on hotion and supershot modes especially optimized forpresent day devices such as JET, JT-60U and TFTR.These high-Ti confinement modes rely on Ti Te,which, as a rule, is inaccessible to burning plasmasbecause: (1) alpha particles principally heat electronsas a result of their high energy; and (2) the electron–ion temperature equilibration time (τeq ≈ 0.5 s) isshorter than the energy confinement time (τE ≈ 6 s)in a reactor scale device. In present day experiments,the energy of injected particle beams is such thatthey principally heat ions. Moreover, since the equili-bration time is comparable to the energy confinementtime τE ≈ τeq, the resulting plasma has Ti Te.

It follows that an ignited burning plasma objective(M > 110) for the EDA design is a large step, but onethat is commensurate with the available database.

3. Tokamak physics processesand projection principles

Tokamak physics has reached a level that sup-ports the detailed design of a new, large facility. To



1.0 2.0 3.00 4.0 5.0 6.00

5

10

15

JET (1997)

TFTR (1994)

JET1991

JET (1997)Fus

ion

pow

er (

MW

)

Time (s)

Figure 4. Thermonuclear power generation in TFTR

and JET versus time (arbitrary zero). The long duration

JET power generation is for an ITER like ELMy H-mode.

a very good approximation, a tokamak is a figure ofrevolution. The resulting property of axisymmetryreduces computation of the force balance betweenplasma pressure gradients and j × B forces to thesolution of a single two dimensional partial dif-ferential equation called the Grad–Shafranov equa-tion [13]. Sophisticated computational solutions yieldaccurate and experimentally validated descriptionsof a tokamak plasma internal structure and bound-aries, and also predict precisely how they respond toexternally applied shaping fields. Stability of theseplasmas with respect to small, symmetry-breakingperturbations can also be accurately assessed byhighly developed variational techniques provided theperturbations obey the ideal magnetohydrodynamicconstraint, wherein very high plasma conductivitypermits one to neglect the component of the plasmaelectric field that lies parallel to the magnetic field.Computation of plasma heating and fuelling is alsostraightforward and sophisticated codes exist thatyield experimentally validated, first principles pro-files of energetic particle creation, plasma heating,non-inductive current drive, and particle deposition.

Experiment has shown that, once an ideally sta-ble equilibrium is assured, the plasma response toauxiliary heating and fuelling is governed by trans-port introduced by the spontaneous appearance ofboth fine scale and global symmetry-breaking fluc-tuations in which the small but finite parallel com-ponent of the plasma electric field plays an essentialrole. In the edge region, the additional complexities

of plasma atomic physics and plasma–surface inter-actions enter. Contemporary tokamak physics con-cerns itself with the consequences of these small butcrucial deviations from the fundamental axisymmet-ric equilibrium.

The physics basis for projecting reactor scaleplasma performance must begin with identifica-tion of the fundamental plasma, atomic and sur-face physics phenomena occurring in tokamak plas-mas and their supporting qualitative theoreticaldescriptions. Quantification then rests on experi-mental data from the present generation of devices,which has benefited from databases developed dur-ing the EDA. From this, one must develop theo-retical/computational (or at least well documentedempirical) methodologies for extrapolation to areactor scale device. The magnitude of the globalfusion energy research effort attests to the fact thatfusion plasmas are complex, with diverse plasmaand plasma–surface interaction phenomena occur-ring simultaneously. Each process requires an extrap-olation to ITER. In these circumstances, we canbring experimental and theoretical information tobear on identifying the qualitative features and scal-ing properties of the fundamental physics phenom-ena. Quantitative predictions flow from the normal-ization of scaling relations to data from a range oftokamaks. In systems such as tokamaks where manyindividual process are at work, a second source ofcomplexity associated with the interactions betweenfundamental processes also enters. Examples include(1) the plasma periphery where atomic radiation pro-cesses are a dominant phenomena in the plasma ther-mal balance; and (2) simulations of integrated per-formance. In such cases, modelling codes replace ana-lytic scaling relations as the preferred methodologyfor prediction of reactor plasma performance.

3.1. General projection issues

How is confidence to be established for projectionof plasma properties to ITER scale devices? Thereare two fundamental approaches. The first is theoret-ical, where the qualitative features, and sometimesquantitative aspects, of physics processes can beunderstood in terms of a theoretical model: often inthe form of a sophisticated code. An example is ener-getic particle losses caused by imperfections in theconfining magnetic fields. One can then validate themodel by comparison with data from a range of toka-maks. Conversely, the lack of a predictive theoreticalmodel, as is presently the case for the H-mode power


ITER Physics Basis

threshold, is a cause for concern in that unknownlimitations may apply. The second source of confi-dence is that projections rest on a common physicsobserved across a range of tokamak discharges span-ning a factor of 6 in linear dimension, a factor of6 in magnetic field strength, and a factor of 34 inplasma current. Uncertainties in projections can thenbe related to the degree of precision with which scal-ing formulas or predictive codes can quantitativelyrepresent the common physics. Physics that cannotbe reproduced across this spectrum of discharges isnot appropriate for use in design basis projectionsfor next step machines, except when there is a com-pelling theoretical reason to the contrary.

The starting point for our characterization ofITER physics processes is Fig. 5, which portraysa representative ‘single null’ divertor plasma equi-librium. In this figure, it is useful to identify fourregions in which different dominant physics prevails,but where there can be important interactions atthe boundaries. The four regions are: (1) the core;(2) the edge pedestal region just inside the sepa-ratrix; (3) the scrape off layer (SOL) plasma justoutside the separatrix; and (4) the divertor cham-ber plasma region, which is an extension of the SOLplasma along field lines into the divertor chamber.Within a given region, a subdivision into short scaleand global processes is also beneficial.

Further insight can be gained by recognizing thatmost of the physics processes are the result of quasi-neutral plasma physics where, to a high degree ofapproximation, ∇·j = 0 and electron and ion chargedensities can be taken as equal. Here j denotes theplasma current density. When quasi-neutrality holds,Kadomtsev [14] pointed out that general scalingscould be cast into non-dimensional forms that involveonly three dimensionless plasma quantities, in addi-tion to dimensionless geometric quantities such asthe inverse rotational transform q, elongation κ, etc.The conventional choice for these three parametershas been

ρ∗ =ion gyroradiusminor radius

=(

2Ti

Mi

)1/2Mi

eBa,

β =plasma pressure

magnetic pressure=

2µ0n(Te + Ti)B2

(1)

ν∗ =connection length

trapped particle mean free path

= νii

(Mi

Ti

)1/2(R

r

)3/2

qR (2)

I

Region I

Region II

Region III

Region IV II

Blanket andfirst wall

Core plasma

Plasma edge andH-mode

confinement barrier

Scrape-off layer

Divertor plasma

Divertorchamber

Figure 5. Poloidal plane view of ITER, illustrating four

principal regions where dominant physics differs. The

separatrix, which forms the boundary between regions II

and III, possesses a point of null poloidal field strength

where it has the ‘X’ crossing. Configurations of this type

are referred to as ‘single null’ plasmas.

νii =(

4√

π

3

)nie

4 ln Λ

M1/2i T

3/2i

. (3)

In these formulas, the temperature is expressed inenergy units (J). For global parameters, one can useT = 2W/3N , where W is the plasma energy con-tent and N its particle inventory of electrons andions.

Because of variations in magnetic field strength ina tokamak, some particles execute a bouncing typetrajectory caused by the magnetic mirroring prop-erty of particle orbits in non-uniform magnetic fields(see [13, p. 42]). Definition (2) emphasizes bounc-ing particles as the key physics that binary colli-sions alter. Other definitions of collisionality involv-ing, for example, temperature equilibration, couldbe used instead. Whenever possible, we cast ourextrapolations in dimensionless form to assure adher-ence to Kadomtsev’s principle and we refer to theseextrapolations as being ‘dimensionally correct’. Inthe plasma periphery, and especially in the diver-tor plasma, neutral atom and atomic radiation pro-



cesses become important and Kadomtsev’s principleno longer applies.

The importance of dimensionless parameters leadsto the concept of ITER Demonstration Discharges,in which region I physics is matched as closely as pos-sible to a reactor in terms of dimensionless parame-ters, including profile and magnetic geometry param-eters. It is found that discharges in present day toka-mak facilities can be formed with values of β andν∗ identical to reactor values, but with ρ∗ having avalue a factor of 5 higher. This reduces the problemof extrapolation from three parameters to a singleparameter, ρ∗, as discussed in Section 7 of Chap-ter 2. Under these constraints, the density and tem-perature scale according to

n ∝ (β2ν∗)1/3B4/3R−1/3, T ∝ (β/ν∗)1/3B2/3R1/3.

(4)

Figure 6 portrays a JET ELMy H-mode dischargewith such features and Table 2 gives relevant param-eters. Its confinement is very close to that pre-dicted by the latest ELMy H-mode scaling relationto be found in Eq. (5) and Fig. 9 of Chapter 2.The relative plasma pressure, as defined by βN =100β(aB/Ip,MA) using MKS units, has a value closeto that planned for ITER, but the non-dimensionalcollisionality ν∗ is modestly larger that the nominalITER discharge of Table 1.

The sudden relaxations of Te(0) are a generictokamak phenomenon called ‘sawteeth’, which areexplained in Section 3.3.5 of this chapter. Sawteethattest to the fact that the central q value is less thanunity. This discharge is an integrated demonstrationof the compatibility of undegraded core confinementwith core βN limits, including pressure driven modescentred on the q = 1 surface [15]. On the other hand,the ‘attached’ divertor edge physics regime for thisdischarge differs from that of a reactor, as discussedin Section 4.

Further evidence for the non-dimensional ap-proach lies in the comparison of discharges pre-pared to have identical non-dimensional parameters,but differing magnetic field, density, auxiliary power,etc. For these discharges, the Kadomtsev principlepredicts that a non-dimensional energy confinementtime defined by ΩiτE should be identical. Here Ωi

denotes the ion gyrofrequency and τE is the ther-mal energy confinement time, τE = W/P , where P

is the thermal heating power. Section 7.2 of Chap-ter 2 reports a comparison between a DIII-D dis-charge and a non-dimensionally identical JET dis-charge. The dimensionless energy confinement times

Table 2. Parameters of a JET DT ITER Demonstration

Discharge

Parameter Value

Shot number 42 756

B (T) toroidal field at axis 2.0

I (MA) plasma current 2.0

R (m) major radius 2.9

a (m) minor radius 0.93

q95 measure of magnetic twist 3.4

κ/δ elongation/triangularity 1.76/0.2–0.3∗

〈n〉/nGR (1019 m−3) 4.7/7.4

Zeff effective ion charge 1.9

P (MW) auxiliary heating power 17.3

Pfusion (MW) fusion power 2.1

Wth (MJ) plasma energy content 4.5

τth (s) heat confinement time 0.26

Bτth 0.52

HH 1.04

ν∗/ν∗ITER 2.1

βN,th normalized pressure 2.25

n(0) (1020 m−3) plasma density 0.59

n(0)ITER† 1.6

Ti(0), Te(0) (keV) 5.5, 5.0

Ti,ITER, Te,ITER (keV) 16.0, 14.5

Divertor status Attached

∗Range of values during pulse.†Scaled at constant β and ν∗.

are identical within 5%, which establishes the valid-ity of the Kadomtsev scaling principle over the sizerange between DIII-D and JET.

Next we turn to the central purpose of this Sec-tion: identification of the various plasma phenomenaoccurring in tokamaks and of the projection princi-ples that apply to them.

3.2. Core confinement and transport

Chapter 2 addresses the anomalous core thermaltransport arising in region I from fine scale plasmaturbulence, whose characteristic scale size is smallcompared to the device size. The working hypothesisis that core transport is governed by core dimension-less physics variables through the core density, tem-perature, and magnetic field values. It is recognizedthat the properties of the region II edge plasma couldalso affect core energy content, particularly if thecore logarithmic temperature gradient is constrainedto lie near marginal stability values.


ITER Physics Basis

Figure 6. JET DT ELMy H-mode ITER Demonstration

Discharge. Normalized β, line average electron density

(1019 m−3), central electron temperature Te(0) in keV,

Dα and total power (MW) versus time for JET pulse

42 756. See Table 2 for parameters. Divertor regime was

attached.

3.2.1. Global confinement scaling

Confinement properties of tokamak plasmas havelong been characterized by their global confinementtime: the ratio of thermal energy content to heat-ing power (in steady conditions). Two approachesare used to project values measured on present daytokamaks to reactor scale devices: regression anal-ysis of a confinement time database; and the ρ∗

scaling of ITER Demonstration Discharges. Globalregression projections for ITER rest on a databaseof ELMy H-mode discharges, which has been con-siderably expanded and improved during the EDA.Log-linear (power law) regression analyses appliedto this database generate confinement time scal-ing relations. Kadomtsev’s non-dimensional con-siderations impose a constraint equation on thepower law exponents. A free fit power law scal-ing relation satisfies this constraint to within sta-tistical uncertainties. Consequently, this constraintis applied to the recommended power law scalingrelations to assure that they are dimensionally cor-rect. As shown in Section 6 of Chapter 2, the scal-ing relation IPB98(y,1) based on the most com-plete set of ELMy H-mode data from 11 differenttokamaks, including all heating methods, takes the

form

τELMyE = 0.0503HHI0.91B0.15P−0.65n0.44

×M0.13R2.05ε0.57κ0.72 (5)

where the units are s, MA, T, MW, 1019 m−3,amu and m, respectively. HH denotes a constantnormally taken to be unity, and the elongationκ is defined as κ = S0/(πa2) with S0 beingthe plasma poloidal cross-section area. Variationsin HH about unity are used in modelling stud-ies [3] to ascertain the sensitivity of fusion perfor-mance to changes in confinement. During the EDA,much attention has been focused on the uncertaintyintervals associated with the recommended scalingrelation. Section 6.4 of Chapter 2 addresses theseissues.

The ELMy H-mode regression analyses are sup-plemented by ITER Demonstration Discharges pre-pared to have core non-dimensional parameters asclose to a reactor as possible. These discharges haveβ and ν∗ values similar to a reactor but differ in ρ∗.H-mode scaling experiments that vary ρ∗ at fixed β

and ν∗ find that confinement invariably lies close tothe regression analysis prediction (5) whose ρ∗ scal-ing is almost that of the ‘natural’ gyroBohm scalingtheoretically predicted by simple dimensional anal-ysis of equations for microinstability transport. Thedimensional analysis argument rests on the fact thatthe scale of the turbulent fluctuations will exhibita separation of spatial scale from the overall deviceand vary according to the ion gyroradius. Almostall first principles microinstability simulations havegyroBohm scaling.

In addition to heat transport, transport of heliumash and angular momentum are important for reac-tor operations. The source of angular momentum istangential neutral beam injection, which, togetherwith the momentum diffusivity, determines the dif-ferential toroidal rotation rate. Section 10 of Chap-ter 2 reports that observations suggest that momen-tum diffusivity is close to heat diffusivity. It is asubject of current research whether differential rota-tion can then close a loop and influence the diffusiv-ity via differential rotation effects on microinstabilitygrowth rates and turbulence levels.

3.2.2. H-mode power threshold and pedestal

ITER confinement projections assume operationin the ELMy H-mode. Transport power losses fromregion I must exceed the H-mode threshold power to



assure that an edge transport barrier and pressurepedestal occur in region II. Such a power thresh-old requirement constitutes an important constrainton the operational space available to a fusion reac-tor [3,4]. Work during the EDA has created an exten-sive database of H-mode power thresholds [16–18].Section 4 of Chapter 2 presents a resulting fam-ily of empirical scalings, Eq. (10) of Chapter 2, forthe L→H power threshold, which are dimensionallycorrect, but which contain appreciable uncertaintycaused by the necessity to determine a functionalform for extrapolation from experimental data.

A considerable reduction in the power thresholduncertainty would result if a viable theoretical mech-anism and theory based scaling relation were avail-able for extrapolation. Although half of the H-modepuzzle has been solved (we know that the trans-port reduction in the barrier arises from electric fieldshear [19]), these considerations have yet to pro-duce a predictive theory for the scaling of the powerthreshold that triggers the evolution from L-mode toH-mode. Recent theoretical and computational sim-ulation work [20–22] has introduced finite-β physicsinto threshold physics and the transport physics ofregions II and III in general. Dimensional analysisarguments indicate that this step is essential for the-ory to recover the empirical threshold scalings. Thesimulation models do exhibit qualitative features ofthe region II/III plasma, but do not yet have a fullseparatrix magnetic geometry needed for an accurateprediction of the power threshold.

Pedestal values of density and temperature justinside the H-mode transport barrier of region IIserve as boundary conditions for the region I anoma-lous transport process. Transport and turbulencewithin region II, which determine the pedestal den-sity and temperature values, are regarded as partof edge physics and treated in Section 3.7 of Chap-ter 4, mostly from the perspective of a databaseto determine pedestal values. Pedestal temperaturescan be very important if region I temperature gra-dients are constrained to lie near a marginally sta-ble logarithmic temperature gradient. First princi-ples simulations of fine scale turbulence are currentlyinvestigating whether region I temperatures will beso constrained or will be relatively independent ofpedestal boundary conditions. Even if region I tem-peratures are reasonably independent of pedestaltemperatures, the region II pedestal energy content isgenerally not negligible compared to region I energycontent and can have a scaling that differs from thecore scaling. In particular, pedestal energy content

may be the source of the ‘isotope’ effect common inconfinement scalings such as (5) [23]. Indeed, workduring the EDA has led to a greater general appre-ciation of the limitations that different scalings ofdifferent physics in different regions inevitably placeon direct experimental investigation of the compati-bility of the desired core and edge physics processesin reactor scale plasmas.

3.2.3. Transport modelling and simulation

Figure 1 depicts representative magnetic surfacesin a tokamak plasma. The very rapid transport ofheat and particles along a magnetic surface relativeto the slow transport across surfaces has lead to amodel of plasma transport wherein magnetic surfacesare regarded as iso-density, iso-temperature surfaces,so transport only need be computed across magneticsurfaces. For cross surface transport, the full shapeof the magnetic surface is used in defining the vol-ume element. Codes constructed in this approxima-tion are called 1.5 dimensional transport models. Sec-tion 8 of Chapter 2 describes two different ways topredict the local energy transport coefficients (ther-mal diffusivity) for ITER from within a 1.5 dimen-sional transport modelling code. The first way, usedin the PRETOR [24] code for FDR projections [4,5],consists of adjusting the thermal diffusivity in sucha manner that the global energy confinement timecomputed by the code is constrained to be equal tothat given by a global scaling relation. The spatialprofile of diffusivity is chosen so that temperatureprofiles are close to those observed in ITER Demon-stration Discharges. This combined use of local trans-port coefficients adjusted to global scaling relationsand of a 1.5 dimensional predictive transport codethat can compute sources, sinks and boundary con-ditions – including some aspects of divertor physics– self-consistently with the predicted profiles is themost direct and reliable way to extrapolate the per-formance of a reactor scale tokamak from present dayexperiments.

A second and more fundamental choice for localheat transport coefficients consists in using a modelfor the diffusivity that does not depend on a globalscaling relation, but instead uses expressions for dif-fusivity and other transport coefficients that aredrawn from theory based considerations such asquasi-linear theory, numerical plasma turbulencesimulations, or simply dimensionally correct formu-las motivated by observations. These models, onceimplemented in transport codes, can be used to pre-


ITER Physics Basis

dict temperature profiles, which are then comparedto experimental measurements available in the ITERProfile Database created during the EDA [25]. Pre-dictions for some models are quite sensitive to thepedestal boundary temperature, because they are‘stiff’, meaning that the heat flux increases rapidlyonce the logarithmic temperature gradient exceeds acritical value. Currently, many models are either stillevolving, with new terms being added, or present toolarge a dispersion when compared to experimentalresults to allow reliable projection of ITER perfor-mances. On the other hand, a theory based expres-sion such as the multi-mode model [26] is shown toachieve reasonable success when compared to exper-imental data and may, therefore, present a crediblealternative to global scaling for ITER predictions.Presently, predictions for ITER using this modelcome close to that using a diffusivity normalizedto the global scaling relation, thereby providing anadditional level of confidence to the overall perfor-mance projections.

Still more fundamental is computational simula-tion of turbulent transport coefficients or, more gen-erally, the non-linear heat flux temperature gradientrelation. Two approaches are used: a straightforwardgyrokinetic particle simulation method [27]; and thegyrofluid approach [28]. The gyrokinetic approach ismore fundamental and uses particle simulation com-putational techniques in the five dimensional phaseof microinstability turbulence: three spatial dimen-sions and two velocity space dimensions, energy andmagnetic moment. Gyrofluid computations rest onvelocity space closure schemes that mimic kineticeffects and reduce the dimensionality of the compu-tational space to three spatial dimensions. Differentcomputational domains are used as well [29]. At thiswriting, the various approaches differ by up to a fac-tor of eight in the heat flux for a given gradient. Res-olution of these differences is in the research stage.

A fundamental understanding is needed to assessthe prospects of the radiation improved (RI) confine-ment mode studied on TEXTOR [30]. These obser-vations find that confinement equals or exceeds thatpredicted by scaling relation (5) without the require-ment to establish an H-mode edge barrier. Two cen-tral physics questions are as follows.

(1) What critical concentration of impurities isneeded to alter microinstability turbulence toproduce the characteristic RI-mode peakeddensity profiles and lower overall thermal lossesthan ordinary L-mode turbulent transport,

even while permitting densities in excess of theGreenwald value (discussed in Section 3.3.8 ofthis Chapter)?

(2) Are the high fractional radiated powers asso-ciated with RI-mode impurity concentrationsessential to altering the turbulent transport?

If a specific impurity concentration is required, then,in a reactor scale device, radiation from the outerportion of region I, called the mantle, may exceedthe available power because of the lower heating perunit volume associated with a fusion energy sourcecompared with auxiliary power deposition levels inpresent day experiments. Of course, a reactor scalefacility will provide a test bed for experimental inves-tigation of possible confinement improvements frominjection of high-Z material, but a common physicsover a variety of tokamaks remains to be establishedbefore the RI-mode can be used as a design basis fora reactor scale facility.

3.2.4. Confinement and magnetic configuration

Many recent experiments in tokamaks indicatethat transport arising from fine scale turbulence isstrongly influenced by the global magnetic configu-ration. Reverse shear configurations are an evidentexample [31, 32]. Even for the ELMy H-mode, theempirical scaling relations indicate a high sensitivityto elongation. Yet more dramatic are the numerousobservations of internal transport barriers (ITBs) –documented in Section 3.4 of Chapter 2 and Sec-tion 2.7 of Chapter 3 – whose duration appears tobe limited by resistive evolution of the q profile. Fig-ure 7 portrays a representative example from JT-60Uin an almost steady state, reverse shear operation.

r/a

ITB layer

0

2

4

6

8

10

12

0

1

2

3

4

5

6

0 0.2 0.4 0.6 0.8 1

Te

, Ti (

keV

)

ne

(101

9 m

-3)

neTMS

TiCXR

TeTMS

Figure 7. An ITB in JT-60U. For details, see Figure 8

of Ref. [31].



Experiments must now focus on establishing acommon ITB physics, consistent with reactor con-straints (e.g. Te ≈ Ti), to provide a basis for confi-dence that such advanced performance modes can berealized on a reactor scale device. Issues concern therole of plasma shaping by elongation and triangular-ity, velocity shear stabilization of microinstabilities,and the importance of deep interior, neutral beaminjection (NBI) fuelling. The physics of Chapters 2and 3 both bear on the prospects for operation ofITER in a transient transport barrier confinementmode.

Similar remarks apply to high bootstrap frac-tion, reverse shear, steady state modes, often called‘advanced tokamak’ operation. Ideal magnetohydro-dynamic (MHD) stability calculations find theseplasmas to be unstable to global, n = 1, externalkink modes for useful values of the plasma pressure(β > 0.03), unless a perfectly conducting shell closelysurrounds the plasma. In principle, a finitely con-ducting shell will suffice, provided the plasma rotatessufficiently fast with respect to the shell so that theskin depth is small compared to the shell thickness.Thus, data regarding plasma rotation and the associ-ated wall stabilization of global MHD kink and resis-tive wall modes [33] appears to be essential to demon-strating useful plasma pressures in steady state dis-charges. As explained in Section 3.3.7 below, thealternative is stabilization by active n = 1 coils [34].Section 3.3.7 of this Chapter and Section 2.4 of Chap-ter 3 report the status of this physics; a firm rotationrequirement has yet to emerge.

3.3. Magnetohydrodynamic phenomena,disruptions and operational limits

In magnetically confined plasmas, MHD phenom-ena that have a global character play a defining rolein determining the accessible parameter space, and,thereby, setting the limits of fusion performance.Global physics processes in region I govern opera-tional limits for the core of a tokamak discharge.Chapter 3 summarizes our current knowledge of suchprocesses, which encompasses ideal MHD stability,determination of the plasma pressure limit via slowgeneration of magnetic island structures driven bybootstrap current, as well as potential methods fortheir control, sawtooth relaxations of the inner core,tolerable error field limits, positional and shape con-trol, and disruption phenomenology, including verti-cal displacement events (VDEs) and runaway elec-tron generation. The steep gradients in region II,

which are characteristic of H-mode operation, causea sequence of relaxation phenomena called ELMs,which are global on the scale of region II. Section 2.6of Chapter 3 addresses their interpretation as MHDphenomena, while Section 3.8 of Chapter 4 evaluatesthe role of ELMs in power and particle control.

3.3.1. Magnetohydrodynamic stability

The principal global stability limits relate tothe maximum plasma current, plasma density andplasma pressure, or β, that can be achieved. At-tempting to exceed these limits often gives rise tomajor disruptions, which leads to a loss of the plasmathermal energy and a dissipation of magnetic energyon rapid time-scales, typically 100 µs and 10 ms,respectively, in present day experiments. In addi-tion, local stability limits give rise to MHD insta-bilities, such as sawteeth in the plasma centre andELMs at the plasma edge, which can have a lesssevere, but nevertheless important, impact on fusionperformance. In Chapter 3, it is shown that whileideal MHD theory, in which parallel electric fields areneglected, is very highly developed and sets the ulti-mate limits on current and β (Section 3.2.1), resis-tive effects must generally be invoked to describethe global instabilities most commonly observed intokamak experiments. Moreover, additional destabi-lizing or stabilizing effects arising from the presenceof a bootstrap current, interactions with energeticparticle populations, and the existence of low levelnon-axisymmetric error fields, can have a significantinfluence on MHD activity in present day tokamaksand are expected to be important in reactor scaleplasmas.

Although it can be shown that fundamental con-siderations deriving from ideal MHD theory deter-mine the limiting parameters for the magnetic equi-librium, principally the plasma current and verticalelongation, the choice of equilibrium parameters forthe ITER reference scenario, a plasma current, Ip,of 21 MA, and elongation, κ, of 1.6, are based onexperimental evidence and practical considerationsthat should apply to reactor scale plasmas in general.It is known that a hard disruptive limit exists whenthe edge safety factor q95 ≈ 2. However, extensiveoperational experience has shown that operation atq95 = 3, as is foreseen in ITER, is a good compromisebetween the desire to maximize energy confinementby operating at high current and the increasing sus-ceptibility to instability as q = 2 is approached. Thischoice, furthermore, allows some margin for increas-


ITER Physics Basis

ing current, if necessary, to offset degraded confine-ment close to operating limits. Operation of elon-gated plasmas, desirable to increase confinement andideal β limits, requires continuous feedback controlof otherwise vertically unstable plasmas. Althoughoperation at higher elongation – κ ≥ 2 – is wellestablished in present day experiments, considera-tions relating to power requirements for feedback sta-bilization of vertical displacements, constraints aris-ing from a reactor relevant poloidal field coil config-uration, and the limitation of forces in VDEs con-strains the choice of κ.

3.3.2. Magnetohydrodynamic β limitsand neoclassical islands

Because fusion power production scales approxi-mately as β2B4, there is a substantial incentive tooperate at the highest attainable β, a point empha-sized by the requirements of steady state opera-tion and attractive economics in a reactor. In Sec-tion 3.2.1 it is shown that the β limit arisingfrom ideal MHD stability, corresponding to βN =β/(Ip/aB) ∼ 3.5 for simple, monotonic q profilescharacteristic of inductive operation, is well validatedby existing experiments and allows ITER a consid-erable margin for operation at its design operatingpoint of βN = 2.2. However, as discussed in detail inSection 2.3 of Chapter 3, the observation in numer-ous experiments in recent years of neoclassical tear-ing modes at βN values well below the ideal limitposes a more significant constraint for ITER opera-tion [35].

Since the configuration is no longer rigorouslyaxisymmetric, these modes can change the topologyof the magnetic field in the vicinity of low order ratio-nal magnetic surfaces to have an island structure bestrepresented in helical flux [35]. Figure 8 shows theisland topology in helical flux while Fig. 9 portrays arepresentative waveform of neoclassical island devel-opment from a sawtooth trigger to a saturated island.

The growth of neoclassical islands arises from aninstability caused by a deficit of bootstrap currentinside a magnetic island due to the flattening ofthe pressure profile across the island, and is gen-erally initiated by so-called ‘seed’ island topolog-ical changes produced by other MHD instabilitiessuch as sawteeth or ELMs. Figure 9 portrays a saw-tooth crash triggering growth of an (m, n) = (3, 2)island. In general, the seed island needed to initiateisland growth is substantially smaller than the sat-urated island size, which is governed by βN and ν∗

0.7 0.8 0.9 1 1.1 1.2 1.3

0.5

0.4

0.3

0.2

0.1

0

0.1

0.2

0.3

0.4

0.5

Figure 8. Neoclassical island topology. An (m, n) =

(2, 1) mode is shown in helical flux [35].

through their effect on the bootstrap current density.Recently, experiments have supported a polarizationdrift theory regarding the ‘onset β’ value at whichneoclassical islands could grow once the relative seedisland exceeds a value of order ρ∗ [38]. The onset β

scales as ρ∗ and depends on collisionality through thecombination ν∗/ρ∗.

Thus, neoclassical islands occur with a rangeof saturated island sizes corresponding to a rangeof βN values, but their effect on plasma perfor-mance depends on βN . In low-βN discharges withsmall saturated islands characteristic of onset β val-ues, neoclassical islands do not affect global con-finement, while, in the most severe cases, generallythose involving m = 2, n = 1 saturated islands atβN ≥ 2.5, they satisfy an island overlap criterion,which can lead to major disruptions. Although theycan be observed at βN values in the vicinity of 2.2in existing experiments, the present understandingof confinement degradation versus saturated island



2.SXR CEN CHOR

W (CM) 3/2

EFI

(W) MOD

SXR EDG CHOR

8617

1.

1.

0.

6.

5.

4.

3.

2.

1.

0.1.25

1.00

0.75

0.50

0.25

0.002600 2800 3000 3200

TIME (ms)3400 3600 3800

0.

Figure 9. Time waveform for neoclassical islands from

DIII-D [36]. At the time of the sawtooth drop in the

central soft X-ray chord (top panel), an (m, n) = (3, 2)

magnetic island growth is triggered (middle panel) and

attains a saturated width, causing a decrease in plasma

energy content (bottom panel) inferred from βp as deter-

mined by EFIT [37] and modelled by an annular region

of high thermal conductivity.

size is not adequate to allow an accurate predictionof the limiting βN in ITER. Successful operation oflong pulse discharges in JET at the required valuesof βN = 2.2 and ITER like ν∗ values (cf. Fig. 6)supports the ITER reference scenario. Higher val-ues of βN (≈3.0) are experimentally accessible in theabsence of sawteeth [39]. Moreover, the slow growthtime of such modes in ITER, ∼50 s, in principlepermits stabilization via electron cyclotron currentdrive [40, 41]. Planned experiments will investigatethis scheme in the near future. Initial results areencouraging [42]. Eliminating neoclassical island lim-itations on β could permit fusion power levels of upto 3000 MW in the ITER FDR facility for periodsof 50–100 s, and establish βN ≈ 3.0 as the nominallimit for inductive tokamaks.

3.3.3. Error field criteria

An operational limit not strictly related to theplasma state arises from the existence of small ampli-tude non-axisymmetric error fields, produced byresidual asymmetries in the toroidal and poloidalmagnet sets, which can cause a growth of resistiveislands and lead to major disruptions. This phe-nomenon is described in Section 2.5 of Chapter 3.The low-m (m = 1, 2, 3), n = 1 components ofthese error fields can instigate the growth of magneticislands in existing experiments when the amplitudeof the total error field, Br, is ∼10−4 of the toroidalfield, B. The mechanism for the initial developmentof the islands is understood in terms of magneticbraking of the plasma mode rotation by interactionof the error field components with the relevant res-onant q surfaces. However, at error field levels typ-ical of moderate to large experiments, only small,essentially harmless, islands should be induced and,to date, there is no satisfactory theoretical explana-tion of why such modes grow to amplitudes capa-ble of causing disruption. An empirical scaling of thethreshold field, required to initiate island growth inohmic discharges, has been assembled that indicatesthat error fields having Br/B ∼ 5 × 10−6–5 × 10−5

may be critical in ITER. Current experiments alsoindicate that the plasma may be most susceptibleto this effect during the initial low density ohmicphase and at the highest β values. Successful cor-rection of such field errors by additional coil setson several tokamaks underpins the proposed installa-tion of a correction coil set capable of cancelling them = 1, 2, 3, n = 1 error field components on ITER.This, together with the observation that externalmomentum injection by neutral beams improves theresilience of plasmas to such modes [43] gives confi-dence that error field induced modes will not limitplasma operation in ITER.

3.3.4. Disruptions

All tokamaks suffer from abrupt, uncontrolledevents, involving rapid cooling and loss of plasmacurrent, which have come to be known as disrup-tions. Accommodating the consequences of disrup-tions imposes significant design constraints on reac-tor scale tokamaks. Section 4 of Chapter 3 detailsthe extensive progress that has been made in under-standing and quantifying the impact of disruptionson a reactor scale device based on data from presentday tokamak facilities. However, modelling studiesindicate that new and significant aspects of disrup-


ITER Physics Basis

tion physics will arise in reactor scale devices.Major disruptions most often occur as a terminat-

ing event when the growth of resistive MHD modes orthermal instabilities has evolved plasma parametersclose to an ideal MHD limit or stochasticity thresh-old. On occasion, they can occur explosively – with-out a resistive precursor mode – when the axisym-metric equilibrium lies in close proximity to an idealβ limit. VDEs resulting from loss of vertical positioncontrol of the plasma constitute both a cause and aconsequence of major disruptions.

It is generally accepted that disruptions occur intwo stages: the thermal quench stage followed by acurrent quench stage [44, 45]. The thermal quenchstage is initiated by the growth of large ampli-tude magnetic islands (often (2, 1) modes) withinthe plasma, which overlap to produce large scaleergodization of the magnetic structure, leading toa catastrophic loss of confinement. This, togetherwith a massive influx of impurities [46], producesa rapid loss of thermal energy, expected to occuron the 1 ms time-scale in ITER, which causes theplasma temperature to fall to as low as 3 eV. A majorfraction of the plasma thermal energy content isdeposited onto the divertor chamber, causing meltingand vaporization of plasma facing materials, whichserves as a source of impurities for the subsequentcurrent quench stage. In the current quench stage,the plasma current decays, on a predicted time-scaleof∼50 ms in ITER, consistent with the very substan-tial increase in plasma resistance. Detailed physicsinvestigations and a large scaling database of ther-mal and current quench time-scales, assembled dur-ing the EDA, give confidence that these projectionsare well founded.

A new phenomenon expected to occur in reac-tor scale devices is the ablation of significant mate-rial from the divertor surface in the thermal quenchstage due to the large thermal energy of the plasma,∼1 GJ [47, 48]. Calculations show that a vapourshield should form in front of the divertor targets,dispersing the majority of the incident energy fluxto the divertor chamber walls via radiation, which,in turn, causes thin melt and vaporization layers toform over the entire chamber. Divertor wall vaporiza-tion serves as an impurity source for the subsequentcurrent quench stage.

Reactor scale plasmas also differ from contempo-rary devices in the evolution of runaway electronsduring the current quench phase. Analysis of ener-getic electron behaviour in the cold, highly impureplasmas produced by the disruption predict that a

substantial runaway electron current, possibly reach-ing 16 MA, can be generated by an avalanche pro-cess involving Coulomb scattering of thermal elec-trons [49]. Large skin currents might also form dur-ing this phase at the boundary between field linesin the plasma and those that contact material sur-faces, giving rise to a potential further source of heli-cal instabilities and magnetic fluctuations that couldserve to inhibit runaway electron generation.

During the current quench phase, control of theplasma vertical position is generally lost and the ver-tical drift of the plasma induces both eddy currentsin the vessel structures and so-called ‘halo’ currents,which flow partly in a ‘halo’ surrounding the plasmaand partly in those elements of the vessel structurein contact with the halo plasma. The resultant elec-tromagnetic forces on the mechanical structure canbe very large, ∼15 000 t in the most severe cases,and, in addition to a predominantly vertical forcecomponent, radial forces and toroidally asymmetricforces can occur. An extensive database of halo cur-rent observations from existing experiments has beenassembled during the EDA and has guided the speci-fication of a mechanical design capable of withstand-ing such substantial forces.

The trigger event for effectively all disruptions canbe identified [45, 50] and disruptions are, therefore,potentially avoidable. Various disruption mitigationschemes have been studied with the aim of dissipat-ing the thermal and poloidal energies of the plasmain a way that avoids the most serious consequences ofthe disruption. Section 4.6 of Chapter 3 summarizesthese schemes. Several of these have been appliedin present day experiments with some success. Forexample, the recognition that the plasma verticalstability can be maintained following a disruptionif the plasma is located at the ‘neutral point’ of thesurrounding conducting structure, has been success-fully exploited in JT-60U to avoid post-disruptiveVDEs [51]. Nevertheless, further R&D is required todevelop mitigation techniques capable of satisfyingthe very demanding requirements imposed by thedisruption consequences on the reactor scale.

3.3.5. Sawteeth

More localized MHD instabilities will also occurin reactor scale devices that, on the basis of expe-rience in present day experiments, are expected tohave a largely benign influence. The sawtooth, whichis characterized by periodic relaxations of the cen-tral electron temperature as portrayed in Fig. 6,



falls into this class. Sawtooth activity is associatedwith an instability that occurs when the central q

value falls below unity and that involves a flatten-ing of the central plasma profiles, but no global lossof plasma energy or particle content. Little impacton global plasma performance is foreseen (Chap-ter 2, Section 5.1), but sawteeth may be the dom-inant mechanism producing seed islands needed totrigger neoclassical tearing modes (as illustrated inFig. 9). A detailed theoretical model of the underly-ing m = n = 1 MHD instability, incorporating non-ideal effects such as resistivity and finite ion Larmorradius and including the stabilizing role of fast par-ticles and thermal trapped ions, has been developedduring the EDA [15] and is discussed in Section 2.2of Chapter 3. For a reactor scale plasma, this theorypredicts that the relaxation events will occur witha repetition time of 50–100 s. Although such longperiod sawteeth might provide the seed island forneoclassical modes, there exist various approachesto modifying sawtooth behaviour, by exploiting theheating and current drive methods available, whichprovide confidence that direct sawteeth effects willnot limit plasma performance in ITER. The dis-charge of Fig. 6 presents an example of an ITERDemonstration Discharge whose global propertiesremain unaffected by sawteeth.

3.3.6. Edge localized modes

ELMs are instabilities of the plasma edge (i.e.region II) associated with H-mode confinement,which result in regular relaxations of the edge tem-perature and density profiles. ELMs limit the maxi-mum pressure gradient that can be reached in thenarrow edge pedestal region that is a distinctivefeature of the H-mode. Section 2.6 of Chapter 3details the various forms this instability can take– called type I, type II, type III, ICRF and grassy– and outlines the various theories that have beenadvanced to explain them [52]. Although a theo-retical description of ELM behaviour is just begin-ning [53, 54] and is focusing on moderate-n ‘peelingmodes’, ELMs have been well characterized empir-ically. The most common type of operation is withtype I ELMs and the majority of the confinementdatabase comes from this operating mode. It has thebeneficial effects of regulating impurity content anddensity rise in a manner that has essentially allowedsteady state operation. The most significant concernis that the energy pulse produced by type I ELMscan enhance erosion of the divertor targets to the

point where component lifetime becomes unaccept-ably short. The present database on type I ELMamplitudes, when projected to ITER, spans a rangefrom unacceptable to acceptable in this regard. Sincethe ITER reference operating point has transportlosses close to the H-mode power threshold, the morebenign type III activity, acceptable from an erosionpoint of view, will in all likelihood prevail. Recently,it has been found that the type III regime dividesinto two branches:

(i) the high density branch of primary interest toITER (DIII-D, ASDEX-Upgrade) has confine-ment quality that is possibly comparable to thetype I regime [55]; and

(ii) the low density branch (DIII-D, JET) can haveconfinement quality strongly diminished fromtype I levels.

It is also observed that ELMs in ICRF heated plas-mas [9] are less severe than with NBI heating andacceptable for ITER operations. Alcator C-Mod,another ICRF heated device, does not observe ELMsat all [56], but rather a region of enhanced Dα

associated with the edge of H-mode plasmas. Sec-tion 3.8 of Chapter 4 evaluates the implications ofELM behaviour for power and particle control.

3.3.7. Magnetohydrodynamics of reverse shearand steady state configurations

During the course of the EDA there have beenrapid advances in confinement regimes that exploitmodifications of the current density profile – such asreversal of the central shear – to optimize plasmaperformance [57–59]. Two general classes of resultshave been obtained. At moderate q values, negativecentral shear operation has stabilized both microin-stabilities and neoclassical tearing modes. The result-ing ITBs (as portrayed in Fig. 7) are quite strikingand could form the basis for transient ignition in areactor scale device. Indeed, recent experiments havereported ITBs enduring many energy confinementtimes [60].

Second, the leading scenarios for high-Q, steadystate operation of a tokamak reactor are low current,high bootstrap fraction, reverse shear discharges. Insuch discharges, the bootstrap current density profilecan be well matched to the desired plasma currentdensity profile. Discharges of this type have been suc-cessfully maintained at low toroidal β [60, 61]. Thelow current of these discharges acts to increase the


ITER Physics Basis

bootstrap fraction but increases their susceptibilityto global instabilities, because their internal induc-tance and normalized current (Ip/aB ≈ 0.8) are low.Many of the limitations encountered in plasma per-formance in such regimes are associated with globalMHD modes, both ideal and resistive. The principalMHD modes observed in such regimes are discussedin Section 2.7 of Chapter 3. Infernal modes (corelocalized ideal kink modes), double tearing modes,and global ideal kink modes have all been identi-fied as performance limiting instabilities in presentday experiments. Attainment of steady state oper-ation in these low normalized current dischargesat toroidal β values comparable to the ELMy H-mode requires high βN values, and low-n externalkink modes must be stabilized by nearby conductingshells. This raises the issues of finite shell resistiv-ity and ‘resistive wall modes’. In principle, the low-n kinks can be suppressed by means of sufficientlyfast rotation of the plasma. However, the analysisis complicated by the existence of a slowly rotatingmode whose growth depends on the proximity of aresistive shell, hence the nomenclature resistive wallmode (RWM). Recent experiments [62] have foundthat plasmas in this wall stabilized regime sponta-neously lose their rotation, in spite of continuingangular momentum input by NBI, and ultimatelysuffer an external kink mode that grows on a wallpenetration time-scale. Active n = 1 feedback coilsto control the spin-down and/or kink mode appearnecessary [34]. The physics of this phenomenon andits implications for steady state operation are dis-cussed in Section 2.4 of Chapter 3.

Further experimental investigation is required todemonstrate the necessary feedback control of theplasma current and pressure profiles necessary toavoid such instabilities and to sustain the high per-formance in steady state required to meet ITER’sultimate goal of non-inductive steady state opera-tion.

3.3.8. Density limit physics

As discussed in Section 3 of Chapter 3, the ulti-mate limit for the plasma density is set by the growthof resistive instabilities leading to a major disruption.Non-MHD effects, principally the growth of edgeand divertor radiation, which can produce radiativeinstabilities known as MARFEs, cause cooling of theplasma edge, contraction of the plasma current pro-file, and destabilization of low-m, n MHD activity,which causes a major disruption in a well established

sequence of events. The role of radiation imbalance inthe limiting process indicates that this limit dependson the plasma input power (including fusion power),a result demonstrated in several tokamaks. A reac-tor scale plasma will differ from present day exper-iments by highly baffled divertor configurations andlarge size, which serve to appreciably reduce neutralparticle densities inside the separatrix and raise theseparatrix electron temperature, thereby potentiallyeliminating MARFE formation, and its effect on thecurrent density profile.

However, of more relevance to the determina-tion of the operating space for ignited operation inITER is the common observation, across many toka-maks, that it is difficult to maintain H-mode confine-ment while increasing the density above the so-calledGreenwald value nG(1020 m−3) = IMA/πa2 with gaspuff fuelling [63]. Indeed, the density limit is charac-terized by the lack of response of plasma density toappreciably increased gas puff fuelling rates. Eventhough this limit is empirical, it is a surprisinglyrobust characterization of the experimental operat-ing space. Essentially by coincidence, a reactor scaletokamak, operating at an optimum average temper-ature of close to 10 keV and at its βN limit, has adensity almost equal to the Greenwald value.

The physics mechanism of the density limit ispresently not known. Pellet injection, especially highfield side launch, can lead to region I densities 20–50% above the Greenwald value [64] indicating thatthe Greenwald value is not a fundamental limit forthe core. Indeed, on DIII-D, outside pellet launchcoupled with divertor pumping leads to plasmas witha density 50% above the Greenwald value and con-finement somewhat better than the ITER93H scal-ing [65]. More recent work points to the importanceof pumping in the divertor private flux region. There-fore, an understanding is sought in terms of region IIand region III global physics, including atomic radia-tion and ionization processes. Section 3 of Chapter 3reviews the data and conceptual models influencingdensity limit physics, while Section 3.7 of Chapter 4describes the density limit in terms of an edge opera-tional space. A limit based on reversion from H-modeto L-mode is proposed, but, in light of the lack of pre-dictive theoretical models for the L→H transition, aquantitative scaling is not available.

Scaling arguments indicate that investigation ofdensity limits in present day experiments resultsin core plasmas with collisionalities ν∗ apprecia-bly above ITER values and potentially differentconfinement physics. This serves as an example of



the difficulties of carrying out integrated core–edgeexperiments in present day machines. The parame-ter spaces of interest to ITER for core and diver-tor physics are disjoint in density in present daytokamaks. These two lines evidently do meet inITER; there, when the density reaches the Green-wald value, divertor modelling calculations show thatthe divertor detaches, while the core plasma colli-sionality remains in the low regime characteristicof ITER and the database presently being used toproject confinement. Establishment of reliable opera-tions at trans-Greenwald densities could appreciablyraise ITER’s fusion power output and neutron wallflux, increasing the attractiveness of the inductivelydriven tokamaks as a reactor. Experiments aimed atincreasing density without confinement degradationusing inside pellet launch and highly baffled diver-tor configurations (the ITER design approach) areneeded to establish a common physics of high densitytokamak operations. The essential difference foreseenbetween a reactor and present day experiments suchas ASDEX-Upgrade and JET is a region of finitedensity gradient between the H-mode pedestal andthe plasma core, supported by an inside pellet launchplasma source. The ability of inside pellet launch andshaping of the plasma periphery to support such agradient is an area of crucial research. Reactor scaledensity limits (if any) will be a key output of burningplasma physics studies. This is a rapidly developingsubject.

3.4. Particle control and power dispersal

Power and particle control are central to the suc-cessful operation of a magnetic fusion reactor and arethe subject of Chapter 4. Control of region I densityis the key method for regulating the fusion poweroutput under ignited conditions. Pellet injection inturn, although it penetrates only into the peripheryof region I, is the principal approach to core den-sity control. High field side pellet launch [64] appearspreferable, but is only beginning to be exploited on aregular basis. Power dispersal is required as it is wellknown from the ITER/CDA studies [6] that unatten-uated power outflow from a fusion reactor will leadto heat fluxes in narrow ribbons around the divertorstrike points that, as a rule, cannot be accommo-dated by material surfaces. The ITER divertor con-figuration portrayed below [66] addresses both powerdispersal and particle control issues.

With respect to particle control, the configura-tion features a high degree of baffling with the intent

Figure 10. The ITER final design report divertor con-

figuration. The divertor dome is in the private flux region

where magnetic field lines remain in the divertor cham-

ber. The dome acts to prevent neutral particles from

entering the x point region from the private flux region,

thus reducing the prospects of x point MARFE forma-

tion.

of minimizing main chamber neutral gas pressures(thereby avoiding a potential source of confinementdegradation) while permitting high neutral pressuresin the private flux region (region IV), which serve toincrease pumping flows and exhaust helium, removemomentum from the plasma component, providecontrol over radiating impurity species concentra-tions, and effect detachment of the divertor plasma,wherein heat and particle fluxes to the divertor tar-get plates are greatly reduced in the neighbourhoodof the separatrix strike points. Contemporary exper-iments are moving towards baffled divertor configu-rations, although they still lack the long leg lengthexposed to private flux neutral pressure that charac-terizes the ITER divertor, even in relative terms.

3.4.1. Power dispersal in divertor plasmas

Dispersal of the power outflow from a tokamakreactor in the region III and IV plasma is requiredto reduce heat flux to a level that material sur-faces designed with adequate erosion lifetime canaccommodate: about 5–10 MW·m−2. Although highheat flux components can be designed to withstandup to 20 MW·m−2 by making the material betweenthe plasmas and the coolant thinner, their lifetimeagainst erosion then becomes unsatisfactory. Thestrategy for power dispersal is to introduce impu-rity noble gas ions – such as neon, argon or kryp-ton (via controlled feedback loops for either pellet


ITER Physics Basis

injection into region I or gas puff into region III) –with the objective of attenuating the heat flux byradiation yet still providing sufficient power acrossthe separatrix to remain in H-mode confinement.Divertor codes find that carbon impurities, sputteredphysically and chemically from the graphite divertorstrike plates, also radiate as part of an uncontrolledbut self-regulating loop. Power dispersal by impu-rity radiation has been successfully implemented incontemporary tokamak experiments in which con-trolled impurity radiation in the mantle (the outerperiphery of region I) and in regions II–IV is suffi-cient to partly detach the divertor plasma from thestrike plates, greatly attenuating the heat flow to thestrike plates [67,68]. Instead, the heat flux takes theform of VUV radiation, which impinges more or lessuniformly on the entire divertor chamber wall andthe baffle region of the first wall near the divertorchamber. The acceptable level of mantle radiation isconstrained by a requirement to maintain powerflowthrough the separatrix above the H-mode threshold.

Two dimensional modelling codes applied toregions III and IV have shown, for a reasonable rangeof input parameters, excellent agreement with diver-tor region detachment observations and a varietyof other sophisticated divertor diagnostics [69, 70].The codes have even predicted recombination ratesand electron temperature values in advance of theirobservation experimentally. Thus, use of two dimen-sional codes constitutes the key method for project-ing the performance of reactor scale divertors interms of ‘free’ input parameters, such as the crossfield diffusivity and the ‘upstream’ region III den-sity. These same divertor codes predict that ITERwill attain the desired partly detached divertor plas-mas, again for a reasonable range of input parame-ters [71, 72]. A theoretical scaling for the regions IIIand IV cross field diffusivity would serve to reducethe number of free parameters. A quantitative con-nection with region II calls for a model of densitydrop across the H-mode transport barrier, which isnot yet available.

Because of different scaling for core and edgephysics, the core plasmas associated with detacheddivertor operations in present day devices are usu-ally significantly colder and more collisional thanthe corresponding reactor cores. In other words,the database of plasmas that exhibit detachment islargely disjoint from the database of plasmas havingcore non-dimensional parameters β and ν∗ similar tothose of ITER. As Chapter 9 argues, the issue of inte-gration of edge and core plasma physics can only be

addressed experimentally in a reactor scale device.Since direct experimental investigation of integratedcore–edge physics is not presently possible, it is a spe-cial challenge to contemporary experiments and the-ory to provide a sufficiently detailed physics model sothat the effect of divertor detachment and core den-sity on core confinement can be reliably projected forpresent day and reactor scale devices.

3.4.2. H-mode pedestal and edge operational space

The physics of region II links the SOL plasma tothe main core plasma. Under nominal H-mode oper-ations, this region contains high density and temper-ature gradients, which are a manifestation of the H-mode transport barrier. Observations show that thetotal pressure gradient is limited by a criterion toremain approximately stable against ideal MHD bal-looning modes. The width of the high gradient regiondetermines the ‘pedestal’ pressure just inside thetransport barrier [73]. This pressure, in turn, deter-mines the energy content of the region II plasma,which can contribute noticeably to the total plasmaenergy by dint of the large volume associated withflux surfaces close to the separatrix. And, as the pre-vious Section indicates, density flow through the highgradient H-mode transport barrier region determinesthe relation between (pellet) fuelling, pedestal den-sity, and the region III ‘upstream’ SOL density.

A key physics issue is: what mechanism sets thewidth of the steep gradient region and therefore thepedestal pressure? Recent data from JET [74] indi-cate that this depends on the hydrogen isotope, asdoes the type 1 ELM frequency. Because theoreti-cal models are just beginning to be developed [75],this issue is being addressed by a database approachdescribed in Sections 3.5–3.7 of Chapter 4. Moredata is required before definitive extrapolations canbe made, and, therefore, the boundary conditionsfor region I transport calculations remain uncertain.Consequently, results of core transport simulationsare usually presented with edge temperature as aparameter.

The ASDEX-Upgrade team [76] has recentlyintroduced an ‘edge operational space diagram’ (por-trayed in Fig. 11) that depicts the physics phe-nomenology of the pedestal part of region II.

Regions are identified for various ELM types, H-mode transitions, and radiation instabilities such asMARFEs. Evidence for a constant maximum pres-sure pedestal for edge plasmas with type I ELMsand a fixed magnetic configuration is quite strong.



0

0.1

0.2

0.3

0.4

0.5

0.6

0 2 4 6 8 10n [1019 m-3]

T [k

eV]

ASDEX Upgrade

Ideal MHDunstable

L - H

ELM-freeH-mode

type IIIELMs

Radiation unstable zone

H-modetype IELMs

B=2.5 T I=1 MA δ=0.1 ∆exp

=2.2 cm

Figure 11. The edge operational space of ASDEX-

Upgrade with a triangularity (sep.) < 0.1. Squares are

just after the H-mode transition, triangles are type III

ELMs, circles are type I ELMs, crosses are radiation

unstable.

A conceptual picture for an H-mode density limitis based on the confluence of type I and type IIIELM regions as density increases. However, magneticfield and size scalings for these regions have not yetbeen established. The next step is to construct a non-dimensional edge operational space diagram that willcorrelate several tokamaks, with the aim of establish-ing the desired common physics.

3.4.3. Erosion of plasma facing componentsand tritium retention

In the present generation of tokamaks, erosion ofthe divertor strike plates serves as a source for impu-rities in the discharge, but has minimal impact on thelifetime of plasma facing components (PFCs). How-ever, in reactor devices, with their much longer expo-sure time, erosion and redeposition will combine tolimit PFC lifetimes [77]. Erosion/redeposition esti-mates are a major factor in the choice of PFC mate-rials for ITER. Such estimates involve several classesof physics, including steady, partly detached divertoroperation, energy pulses arising from ELMs, vapor-ization and melting caused by the disruption ther-mal quench, and slow thermal cycles where divertordetachment is lost for ∼10 s. Figure 12 illustrates thechoice of plasma facing materials in the ITER FDRdivertor design. Tungsten is the preferred plasma fac-ing material except near the divertor strike plates

because of its low sputtering rate, while CFC (car-bon fibre composite) is chosen for the strike pointregion because it sublimes, rather than melts, dur-ing disruption thermal quenches, thereby avoidingsurface irregularities that might later form hot-spotsin normal steady heat flux operation. Section 5 ofChapter 4 summarizes our knowledge of this area,which has special impact on design choices.

Figure 12. Plasma facing materials of the FDR diver-

tor.

The ultimate fate of tritium fuelling is of keyimportance to the issues of maintaining the in-vesseltritium inventory below the ≈1 kg safety limit andthe sustainment of tritium self-sufficiency for a reac-tor. Reactor blanket neutronics designs yield a tri-tium breeding capability that assures tritium self-sufficiency provided 90% of the tritons injected intothe core are burned, even though they may have togo through a pellet–plasma–neutral-gas-pump–pelletcycle several times. Thus, if there is any retention oftritons in the wall or codeposited layers during thesecycles, such as occurred in TFTR [78] and JET [11],tritium recovery and self-sufficiency become issues.Section 6.6 of Chapter 4 identifies codeposition witheroded carbon from the divertor strike points asthe most likely mechanism for tritium retention andoutlines potential tritium recovery techniques. Sec-tion 1.3 of Chapter 8 outlines the operational aspectsof tritium retention.

At present, our predictive capability regardingerosion and tritium retention leaves substantialuncertainties in our estimates of the erosion rates ofplasma facing surfaces and of the level of codepositedtritium in reactor class experiments [79]. We there-fore foresee a significant experimental programmeduring the EDA transition phase on the present gen-eration of tokamaks and during the initial protonplasmas phase of ITER operations, when access tothe machine is unhampered by activation, to bettercharacterize and understand erosion and hydrogenretention, to develop and test techniques to mini-mize hydrogen retention, and to efficiently recoverhydrogen retained in the plasma facing components.


ITER Physics Basis

3.5. Energetic particle physics

In ITER, it is expected that the dominant heat-ing will be via thermonuclear generation of 3.5 MeValpha particles whose energy is well above that asso-ciated with the characteristic Alfven speed EAlf =MαB2/2µ0nMDT ≈ 1.3 MeV. A self-heating reac-tor must confine these particles, and Chapter 5 dis-cusses and evaluates potential loss paths, as well asthe effect of a dilute energetic particle species indestabilizing or stabilizing collective modes of theplasma. Direct particle loss by toroidal field rippleappears controllable by design. As noted for otherissues in this Chapter, present day experiments arenot a direct replica of ITER with respect to α par-ticle effects on collective modes, which range fromsawteeth to Alfven eigenmodes. (The latter are dis-crete modes of the MHD spectrum that arise becauseof the periodic nature of poloidal variations of mag-netic field strength, such as toroidal Alfven eigen-modes (TAEs)). The general properties are thatunstable Alfven eigenmodes in ITER will have sig-nificantly higher toroidal mode numbers and a rela-tively weaker drive than in current experiments. Dur-ing the EDA, linear stability theory for Alfven eigen-modes has developed to the point where remarkablyaccurate predictions for low toroidal mode number n

excitations are possible [80,81]. Extensions to highern find that, in ITER class devices, Alfven eigenmodesare just unstable and become quite stable with minorspreading (involving no losses) of the α particle pres-sure profile [82]. Non-linear computational programsto fully assess Alfven eigenmode physics on a reactorscale tokamak are under development. ITER classexperiments have the correct parameters to returnimportant new physics needed by reactor designers.Experimental evidence also indicates that a relatedm = n = 1 instability, ‘fishbone’ activity, which ischaracterized by regular coherent bursts of internalMHD modes and which is destabilized by fast parti-cles, should not prove significant in ITER [83].

3.6. Auxiliary power physics

Reliable and effective sources of auxiliary powerare an element in almost all the advances in tokamakplasma physics. Not only is the evident heating func-tion important, but auxiliary power can also drivenon-inductive current, which has maintained smalltokamak discharges for more than two hours [84]and is crucial to the realization of steady state toka-mak operation. Differential toroidal rotation can bedriven by neutral beam angular momentum injec-

tion and arguments are advanced that the resultingflow shear could play a significant role in determin-ing microinstability turbulence levels and the result-ing transport [19, 85]. In current experiments, neu-tral injection can also provide significant fuelling ofhot plasma cores and resultant high density gradi-ents within region I. For ITER, appreciable NBIfuelling at 1 MeV per nucleus is prohibitive froman energy budget point of view. Finally, selectivetransfer of auxiliary power to ions has led, duringthe EDA, to plasmas with Ti Te, which havegreatly improved ion thermal confinement and highthermonuclear fusion rates [86]. Such regimes willbe available to ITER only at low densities wherethe energy confinement time is comparable to theelectron–ion temperature equilibration time.

Chapter 6 describes four approaches to auxiliarypower in ITER as follows.

(1) The advantages of fast wave auxiliary powersystems [87] rest on a well developed commercialpower generation technology in the 40–100 MHz fre-quency range, and a validated ability to calculatewave propagation and absorption by straightforwardlinear and quasi-linear methods, respectively. ThePION code [88] summarizes this ability and canbe used as a module in transport codes like PRE-TOR, which simulate the entire tokamak plasma.Fast wave current drive capabilities are restrictedto central current drive where attractive efficiencieshave been observed. When used in connection withion cyclotron absorption, fast wave methods also pro-duce a good degree of control over heating profiles.The principal limitation of the ITER fast wave sys-tem is related to the requirement of operating witha large gap (∼15 cm) between antenna and plasmaseparatrix, which prevents a close antenna/plasmacoupling, limits the power/area that can be trans-mitted through a port, and enhances the sensitivityof antenna radiation resistance to variation in theregion III density caused by ELMs.

(2) NBI [89] benefits from the spatial separa-tion of plasma physics and beam technology. Anattractive feature of NBI is that the beam veloc-ity distribution is known and its power level canbe accurately measured. In reactor scale tokamaksthat require ∼1 MeV beams to assure penetration,the beam must be created by negative ion beamtechnology. Ionization of beam particles in a toka-mak plasma then leads to sources of density, energyand angular momentum. The cross-section for ioniza-tion processes are known and involve multi-step pro-



cesses. Once created in the plasma, ions from neutralbeams slow down and scatter in angle according toclassical binary collision physics. Confined ions exe-cute conventional banana orbits and toroidal preces-sional drifts. As an energetic particle species, beamions can undergo or excite many of the collective pro-cess identified in Chapter 5. Current drive theory iswell developed. Indeed, the high level of detail withwhich codes can calculate beam interaction with theplasma has made NBI the power source of choice inmany plasma physics experiments. By displacementof the target plasma, one can realize a good measureof control over the spatial deposition profiles. Froma plasma physics viewpoint, neutral beam power isunique in two ways: as a source of angular momen-tum; and as a source for charge exchange processesin region I from which line radiation can be exploitedby diagnostics.

(3) Electron cyclotron auxiliary power systems[90] in the 130–200 GHz frequency range can providereliable and predictive localized heating and currentdrive. Ray tracing wave propagation codes and rel-ativistic quasi-linear wave absorption codes are welldeveloped and validated against experimental data.ECH methods are insensitive to region III densityand the proximity of the plasma to the antennastructure. The resonance nature of the interaction,coupled with the narrow beamwidths of ECH anten-nas leads to very localized wave–plasma interactionregions. This provides the basis for a number ofimportant plasma control functions such as drivingoff-axis current density to extend the duration ofreverse shear magnetic configurations and generatinglocalized, modulated current density to suppress neo-classical tearing modes. At present, ECH is under-utilized in contemporary tokamaks because reliablesources are only just now becoming available. Giventhe slow pace of tokamak upgrades, it will take 5–10years of experimental research before the capabilitiesof ECH on tokamaks are demonstrated at the powerlevels now enjoyed by neutral beam systems.

(4) In ITER like plasmas, the principal role forlower hybrid auxiliary power systems near a fre-quency of 5 GHz is off-axis current drive. A wellknown accessibility criterion prevents wave propa-gation into the central regions, basically eliminatinga heating role for lower hybrid auxiliary power. Thephysics governing lower hybrid/plasma interactionson an ITER scale device is, like ECH, straightforwardray tracing wave propagation and quasi-linear wave–plasma interaction physics. In contrast with present

day experiments, single pass absorption is the normfor lower hybrid in reactor scale plasmas. The morequalitative stochastic ray formalism is needed onlyfor smaller plasmas of current devices, where mul-tipass absorption prevails. A corresponding increasein confidence of ray tracing results follows. Becauseboth the frequency and parallel wavenumber arefixed in the proposed lower hybrid launching sys-tem [91], and because flat region I density profilesare expected, the temperature profile will determinethe off-axis current drive profile. Driven off-axis cur-rent is an essential element in maintaining steadystate magnetic reverse shear profiles.

Coupling of lower hybrid power from the launcherinto propagating plasma waves depends on theplasma density profile in the immediate vicinity ofthe launcher. This will be known only after ITERoperates. Alternative measures to create plasma closeto the antenna, such as local gas puffing, must beexamined for compatibility with the general strat-egy of maintaining low neutral densities in the mainchamber. Again, these experiments must be per-formed on ITER itself.

3.7. Physics of plasma diagnostics

A different set of physics processes, plasma mea-surement physics in contrast to plasma performancephysics, governs our ability to measure plasmaparameters in reactor scale tokamak discharges. TheITER diagnostics programme has established, viaextensive consultation during the EDA, detailedrequirements for plasma measurements, which aresummarized in Chapter 7 [92]. These requirementsreflect the higher levels of feedback control loops andinterlocks anticipated for a reactor scale facility. Byand large, the physics processes utilized for diagnos-tics on the current generation of tokamaks will trans-late to a reactor scale device and fulfil the require-ments. Indeed, most of the diagnostic techniques inroutine use today will be available to ITER. Thereare, of course, substantial technical differences, forexample radiation hardening of components, protec-tion of mirror systems from plasma erosion and dis-ruption debris, and the need for all in-vessel compo-nents to be maintainable with remote handling tools.Nevertheless, there remain some highly desirable, butnot crucial, measurements, which need new physicsmeasurement principles to fulfil the requirements.The density of light ions, including the thermalizedhelium ash, is an example where, in a reactor scaleplasma, the neutral beams at a velocity appropri-


ITER Physics Basis

ate to charge exchange recombination spectroscopydo not penetrate to the plasma core. Other examplesinclude the energy and density of confined and escap-ing alpha particles. Possible new diagnostic physicsconcepts to meet some of these demands are in theresearch stage.

3.8. Physics of plasma controland steady state operation

Chapter 8 examines, via integrated modellingcalculations, the ability of proposed ITER controlcapabilities to implement the intended control [93].The emphasis is on adequate control throughout acomplete discharge: from breakdown, which initi-ates plasma formation, and burn-through, whereinthe plasma becomes completely ionized and fullystripped, to kinetic burn control, which attains adesirable and stable level of fusion power, and finallyto non-disruptive shutdown. While this demonstra-tion is based on the specifics of the ITER design, itis exemplary of reactor scale operations. Its successprovides confidence that there are no hidden incon-sistencies in the ITER operational plans.

It is a goal of ITER to demonstrate reactor scalesteady state operation if this proves possible forthe tokamak. General theoretical arguments indicatethat low current, reverse shear, high bootstrap frac-tion discharges are the best approach to steady stateoperation [94]. Figure 13 portrays a representativeequilibrium that is consistent with the ITER poloidalfield coil set and vertical control capabilities [5].

Use of a non-chamber filling discharge permitsconfigurations with high elongation and triangular-ity to investigate the physics of reactor scale steadystate plasmas without a requirement for high-Q.

Table 3 lists properties of this equilibrium. Onenotes that only modest improvements in confinementmultiplier and normalized β are needed to realizethis state. On the other hand, the density needed toattain the prescribed βtor is considerably in excessof the Greenwald value. This is a general propertyof advanced tokamak scenarios where low current isrequired to reach a high bootstrap fraction.

It should be stressed that this equilibrium isunstable in the absence of a nearby conducting shell,and, hence, is subject to the RWM [33] (as dis-cussed in Section 3.3.7). Stabilization by rotation orn = 1 internal feedback coils [34] is required. Experi-ments on DIII-D have found that plasmas in the ‘wallstabilized’ regime spontaneously lose their rotationand are ultimately subject to an external kink mode

Figure 13. Calculated high bootstrap fraction, reverse

shear equilibrium in the ITER FDR plasma chamber.

Table 3. Properties of a steady state discharge

H-mode like

Parameter density profile

Pfusion/PCD (MW) 1500/100

QCD 15

R0/a (m) 8.66/2.32

k95%/δ95% 2.00/0.44

fBS 79%

γCD (×1020 A·W−1·m−2) 0.21

βN/βtoroidal 3.8/3.7%

〈Te〉n/Teo (keV) 12.5/25.9

〈ne〉/neo (×1020 m−3) 1.0/1.1

〈ne〉/nGR 1.4

τE (s) 2.46

τE/τITER93H 1.22

that grows on a wall time-scale [62]. For Table 3,the temperature, density and current density pro-



files are simply assumed, in lieu of using coupled dif-fusion equations based on transport coefficients thatare poorly known in the reverse shear regime andcan often exhibit spontaneous formation of transportbarriers.

Nonetheless, Fig. 13 and Table 3 demonstrate thata reactor scale device can have sufficient flexibil-ity to investigate steady state tokamak operationsas its requirements are presently understood pro-vided provisions are made for internal n = 1 feed-back coils. Sections 3.2.4 and 3.3.7 discuss the RWMand the need for a more quantitative understandingof rotational or active feedback stabilization require-ments. Lastly, we note again that Table 3 calls for adensity considerably in excess of Greenwald, under-scoring the importance of plasma fuelling physics toadvanced tokamak research.

3.9. Summary

This Section has identified a wide range of physicsprocesses occurring in present day tokamak plasmas.It is to the credit of the fusion science effort thatalmost all observations can be understood, at leastconceptually and often quantitatively, in terms ofphysics processes. The next Section will examine howthese processes will be altered and what changes intheir relative importance will occur in the transitionfrom present day devices to reactor scale facilities.

4. Reactor scale experimentalplasma physics

It is illuminating to regard a reactor scale toka-mak as a new scientific facility and ask: what physicsissues can we anticipate investigating that are inac-cessible to the current generation of tokamak facili-ties? Will plasma based design requirements exceedthe capability of engineering and technology torespond? The totality of issues raised by these ques-tions comprises reactor scale plasma physics, whichcan be generally defined as the physics that is domi-nant in reactor scale device, with particular emphasison those issues that cannot be investigated with con-temporary tokamaks (and their evident upgrades).Of course, the proposition that there is importantphysics to be learned only from reactor scale devicesimplies that uncertainties exist in our projections forITER plasma performance. In this way, the sources ofuncertainties in plasma performance projections canbe viewed as opportunities for experimental researchon ITER. In the final analysis, a judicious balance

must be made between projection uncertainty andresearch opportunity in an acceptable and usefuldevice design.

Chapter 9 provides an extensive discussion ofburning plasma physics in terms of opportuni-ties for new experimental physics that an ITERclass machine will enable. This introduction outlinesour principal arguments and illustrates them withselected examples.

It is useful to divide reactor scale plasma physicsinto three elements: (1) energetic particle physics;(2) self-heating and thermal stability; and (3) scaledependent plasma physics. The last arises from thedifference in physics engendered by the fact that areactor must have a substantially larger magneticfield strength and size than present day devices. Letus turn to a brief description of these three elements.

4.1. Energetic particle physics

Superficially, it would appear that the presence ofenergetic particles, specifically 3.5 MeV alpha parti-cles, is a key difference between present day devicesand a burning plasma. These particles have a cen-trally peaked profile, transfer their energy to theambient electrons, and have a characteristic veloc-ity that exceeds the Alfven velocity of the thermalplasma. As such, they are capable of interacting withdiscrete stable MHD modes known as Alfven eigen-modes, destabilizing them through α particle expan-sion free energy. Since present day tokamaks arealso heated by energetic particles, which are createdby fast wave minority ion cyclotron heating or neu-tral beam injection, one must ask: how is it antici-pated that energetic particles physics in reactor scaledevices will differ from that in contemporary toka-maks, when the operational conditions are such thatthe characteristic velocity of the energetic particlesexceeds the Alfven velocity? (Usually this means neg-ative ion beam injection or minority fast wave heat-ing.) First, it is evident that the fundamental drive isweaker in reactor scale devices because the relativefast particle concentration, nf/n, is given by

nf

n=

(3T

Ef

)τf

τE(6)

where τf is the fast particle slowing down time andEf its characteristic energy. Based on values pre-sented in Tables 1 and 2, the relative fast particleenergy density in a reactor scale device is approxi-mately a factor of 10 less than in present day devices.Some dissimilarities arise because the isotropic veloc-ity distribution of the alpha particles differs from the


ITER Physics Basis

anisotropic distribution arising from neutral injec-tion or ion cyclotron sources. Theory also permitsweaker wave particle interactions for v ≈ 0.3 · VAlf ,which can excite Alfven eigenmodes even when theauxiliary heating system fails to produce particleswith v ≥ VAlf . These differences can be accountedfor by theory, which has been a remarkably goodguide [80,81] for linear stability. The key difference ofenergetic particle physics in burning plasma devicesderives from the third element of burning plasmaphysics – that the scale of the device is larger thanthat of present day devices – and not just from thepresence of super Alfvenic particles, which are foundat greater relative densities in current devices. AsChapter 5 explains, the unstable toroidal mode num-bers in reactors are expected to be appreciably higherbecause of the device size, and reactor scale devicesmay exhibit multi-mode Alfven eigenmode turbu-lence.

The 3.5 MeV energy of alpha particles does havethe important consequence of transferring theirenergy directly to electrons, which precludes opera-tional modes with Ti Te, often found with neutralbeam injection heating in current devices.

4.2. Self-heating and thermal stability

Controlled, steady state operation of a fusionplasma implies that transport power loss fromregion I balances the sum of self-heating from alphaparticles and auxiliary heating power. A thermallystable solution further requires that the transportpower losses increase more rapidly with tempera-ture than the sum of the fusion power and auxiliarypower, assuming a feedback loop that decreases aux-iliary power (if any) as α particle power increases.Appendix A derives a formula for the transportpower losses that is equivalent to the ITER98H(y,1)confinement scaling relation Eq. (5) and is

Ploss(MW) = (1.8× 105)n1.6

20 T 2.8610 R0.11

m κ0.8

B3.03T M0.37H2.86

H

×(

aB

I

)2.6(a

R

)1.49

(7)

where n is in units of 1020 m−3, T is the volume aver-age temperature in keV, P is in MW, and I in MA.The strong increase of transport losses with tempera-ture is the factor that permits a stable thermal equi-librium in the presence of a thermonuclear reactionrate that is increasing (but less rapidly) with temper-ature. Since the fusion power increases as n2 while

transport losses increase as n1.6, higher density willlead to higher fusion powers in thermal steady state.For ignited operation with no auxiliary power, therelation is approximately P ∝ n3.2. In actual opera-tion, the constant HH could change abruptly, reflect-ing a change in confinement mode or the creation ofan ITB. In this case, the density will have to respondto maintain a steady fusion power. Clearly, demon-stration of thermal control will be an essential partof burning plasma physics. The characteristic time-scales are the energy confinement time and the cen-tral density buildup time-scale in response to periph-eral plasma fuelling. The latter time-scale is not wellcharacterized by an experimental database and, inall likelihood, depends on the fuelling method.

For the case of steady state operation of atokamak, self-heating generalizes to self-generationof plasma (bootstrap) current via pressure gradi-ents and the degree of self-consistency between therequired current density profile and the bootstrapcurrent density profile. The current density profile,in turn, controls ITBs that close the loop via theirinfluence on pressure gradients and, hence, the boot-strap current. Two time-scales exist in this system,the energy/density scale and the magnetic flux dif-fusion scale. Only with very long pulses appreciablyexceeding magnetic flux diffusion times (∼300 s forITER parameters) will definitive data regarding theprospects for steady state operation of a reactor beavailable.

4.3. Scale dependent plasma physics

The size, plasma current, and magnetic fieldstrength of a tokamak device in which fusion powerbalances transport losses can change the relativeimportance of physics processes and can introducequalitatively new physics, which is negligible inpresent day devices. Chapter 9 discusses the conse-quences of scale in considerable detail. This intro-duction gives five illustrative examples to convey theimportance of size to the investigation of reactorscale plasma physics. We note that we have alreadyargued that plasma scale is the dominant parameterregarding the difference in energetic particle, Alfveneigenmode physics between present day devices anda reactor.

Our first example concerns the scaling of theplasma density relative to the Greenwald value inITER Demonstration Discharges (such as those por-trayed by Fig. 6). The Greenwald normalized density



is defined by

n

nGR=

n20πa2

IMA=

(4π

10

)n20Rq

B(1 + κ2). (8)

The combination of definition (8) with the constantβ and ν∗ scaling of (4) indicates that if a reactoris at the Greenwald density value, as ITER param-eters indicate, then Demonstration Discharges inpresent day experiments at identical β and ν∗ val-ues must have densities appreciably below the Green-wald value. Thus, the integrated system of core con-finement and edge density limit physics can only bedirectly investigated on a reactor scale device.

The second illustration is again an example ofthe coupling between core and edge physics pro-cesses and involves changes in proximity to opera-tional boundaries. Chapter 2 presents empirical scal-ing relations for two powers: the transport powerloss from an H-mode discharge and the power flowthrough the separatrix required to maintain H-modeedge conditions. The scaling of these two powersdiffers [95]. For ITER Demonstration Discharges inpresent day devices, which are prepared to have corevalues of β and ν∗ identical to those anticipated forITER, the transport power loss considerably exceedsthe requisite H-mode power threshold, thereby assur-ing operation in H-mode. For an ITER scale plasma,these two powers are roughly equal and questionsarise as to whether operation near the thresholdpower boundary will realize the full benefits of H-mode confinement. Of course, one can always reduceauxiliary power to operate near threshold in presentday devices, but then the core β and ν∗ values willdiffer from those of ITER, potentially changing coreconfinement.

The general conclusion is that definitive experi-mental investigations of interactions among diverseplasma processes in the core and edge can onlybe achieved in reactor scale devices. Note that thedivertor plasma of the ITER Demonstration Dis-charge in Fig. 6 is attached, not detached. Sec-tion 3.4.1 and Section 3.3 of Chapter 4 remark thatthe present database for core confinement is disjointwith detached divertor database. Thus, integratedexperimental investigations of core–edge compati-bility constitute an important part of reactor scaleplasma physics.

The third illustration concerns the thermalquench phase of full power ITER disruptions, inwhich the thermal energy content of the plasma israpidly deposited onto the plasma facing componentsin the vicinity of the divertor strike points. The mag-

nitude and short duration of the pulse will causevaporization and melting (or sublimation) of divertorstrike point material as well as a portion of the diver-tor chamber wall [47,48]. This regime is not encoun-tered in present day tokamaks, where the thermalpulse associated with the thermal quench can beaccommodated by the heat capacity of solid mate-rial. In the case of ITER, the vaporization occurringat the divertor strike points and surrounding areaswill release carbon and tungsten into the subsequentcurrent quench phase disruption, acting to cool thisplasma and abet the formation of a runaway electronavalanche. In this case, reactor scale plasma physicsincorporates a phenomenology that is unattainablein present day devices. Investigation of such plasmasis also part of burning plasma physics.

For our fourth example, we turn to the avalanchetheory of runaway electron generation [49] presentedin Section 4.4 of Chapter 3. The key result is that thenumber of avalanche e foldings in runaway electrondensity is proportional to the plasma current andis of order unity in present day experiments. How-ever, in an ITER class device, the number of e fold-ings is large, making runaways generally negligiblein present day machines but of clear importance toreactors.

Lastly, we note that many tokamaks with ITBshave plasmas with Ti Te, which is known toreduce microinstability transport [96,97]. This condi-tion arises because, in present day experiments with40–100 keV beams, the neutral beam auxiliary heat-ing power is transferred primarily to the ions. How-ever, in reactor class devices, the beam energy mustexceed 1 MeV to assure adequate penetration. Con-sequently, both auxiliary and thermonuclear α parti-cle power flows principally to electrons. In any event,for reactor plasmas, the electron–ion equilibrationtime will be short compared to the energy confine-ment time, assuring Te ≈ Ti. Direct, central fuellingby neutral beams is also negligible in a reactor butoften important in providing peaked density pro-files for ITB discharges in present day experiments.Experimental investigations of ITBs and their powerthreshold scaling relevant to burning plasmas shouldfocus on plasmas with negligible beam fuelling, withelectron heating, and with Te ≈ Ti.

Taken collectively, these examples illustrate thatreactor scale plasma physics has fundamental dif-ferences from the plasma physics of present daymachines. Not all plasma regimes found on presentday tokamaks are accessible to reactor class dis-charges. Reactor class discharges can be affected by


ITER Physics Basis

processes negligible in present day devices. Undis-ciplined translation of plasma performance frompresent day devices to reactor scale facilities can bequite misleading. One can conclude that, in addi-tion to achieving an understanding of individualphysics processes, a reliable reactor physics basismust address issues of reactor relevant integrationof individual processes inaccessible to present dayexperiments. Reactor scale experimental physics hasmuch to teach us.

5. Projecting ITER operations

Fundamentally, the ITER Physics Basis is a set ofprojection methodologies with which to extrapolatefrom data provided by the present generation of toka-maks and guide the design of a reactor scale device,as exemplified by the ITER parameters of Table 1.Extrapolation methodologies are needed for codesand theoretical models as well. The goal is that theresearch facility that results should provide uniqueand essential information needed for the design of asubsequent fusion power station. This goal will bemet if the methodologies are sufficiently accurateso that tokamak discharges can firstly be reliablyproduced, controlled and diagnosed, and secondlyhave appreciable thermonuclear power and param-eters close to those proposed for ITER.

ITER projection methodologies can be organizedinto three classes:

1. those which bear on the issue of whether a sin-gle pulse can be reliably produced;

2. classic plasma physics performance projectionsfor reactor scale devices; and

3. multiple pulse, plasma wall erosion and tritiumretention issues.

In a number of areas, our predictive ability doesnot suffice to permit machine operation at fullparameters immediately. The usual course of gradu-ally increasing experimental parameters and observ-ing plasma physics and machine response should bethe norm. The assessment of projection methodolo-gies then should be based on their ability to guidesuch an experimental campaign and to benefit fromthe data returned by the observations.

5.1. Single pulse issues

The starting point for an assessment of projectionsregarding single discharge issues is an assumption

that the nominal plasma performance objectives willeventually be reached through a series of shots withgradually increasing parameters. The principal singleshot issues are plasma initiation, control, disruptions,thermal stability, and heat load on the divertors. Thebasic methodology is to use physics information todevelop conservative design requirements that willassure that machine components will survive andfunction, even at extreme limits of projected plasmaperformance. For disruption physics, quantitativestatements rest on databases derived from presentday devices, extrapolated to a reactor via straightfor-ward arguments. In the case of toroidally asymmet-ric halo currents, new and definitive databases havebeen created that adequately bound the problem.To evaluate divertor heat fluxes, results from twodimensional modelling codes, which replicate presentday divertor observations, are invoked. These codesfind the desired partly detached divertor solutionsfor ITER like plasmas.

Initially, discharges will be in proton plasmas,with limited auxiliary heating power (100 MW orless). Breakdown and burnthrough will be fundamen-tally the same in reactor scale plasmas as in presentday tokamaks, with the additional help of electroncyclotron heating to assure breakdown and acceler-ate burnthrough. As has been the case with presentday tokamaks, an experimental period will likely beneeded to understand the specifics of the poloidalfield null formation and vertical field ramp up asso-ciated with bringing the current up to and past the1 MA level.

Accurate estimates of power dispersal and diver-tor plate heat fluxes will rest on two dimensionaldivertor modelling codes. Based on the impressivesuccesses of divertor modelling codes in presentday plasmas, initial low to moderate power protonplasma discharges should suffice to calibrate thesecodes and confirm, via density and power scans, theirability to predict detached divertor status for fullparameter shots. The key calibration issue is thescaling of the turbulent cross field diffusivity, whichhas been taken as an adjustable input in simulatingpresent day experiments. The results are sensitiveto the adopted value [72]. Since proton plasmas willlikely be in L-mode, there will be no need to reacha prescribed power through the separatrix in cali-brating the divertor modelling codes. With a widerange of pellet and gas puff fuelling, impurity injec-tion options, and pumping speeds, a reactor scalefacility will have considerable flexibility to optimizepower dispersal and divertor heat flux solutions. The



methodology to assure adequate power dispersal ona single shot basis is sound.

Proton plasma discharges of increasing currentand energy content will also calibrate our ability toevaluate disruption consequences at full parameters.Our database analyses and modelling indicate thatdisruptions of low energy content discharges withq ≈ 6 and 100 MJ of thermal energy should be easilyaccommodated by designs such as that presented inthe ITER Final Design Report [4]. A calibration ofour estimates of disruption consequences by observa-tions of such discharges should confirm our predic-tions that the FDR design can withstand full param-eter disruptions. Similar remarks apply to verticalposition control, where q ≈ 6 discharges will requiresmaller control voltages than full parameter q ≈ 3shots.

Combining our present projection methodologies– which indicate no problems on a single shot basisfor low current, low energy content plasmas – withdata from discharges of increasing performance gen-erates confidence that a reactor class device can beoperated at full parameters with no adverse conse-quences on a single shot basis.

5.2. Physics performance projections

Section 3 has made it clear that there are manyplasma processes that require extrapolation fromcontemporary tokamak plasmas to a reactor scaledevice. Our assessment is that, collectively, theextrapolations give sufficiently accurate guidancethat the major parameters of a reactor scale experi-ment can be chosen with confidence, that ELMy H-mode plasma performance will lie close to the nom-inal projections, and that tradeoffs made in reach-ing the major design parameters are indeed mean-ingful. Arguably, the major weakness is the fact thata reactor scale tokamak operates close to the Green-wald density value and the H-mode power threshold.But confinement degradation near these boundariesis just the issue that Section 4 concludes needs areactor scale device for resolution.

Let us briefly discuss the major extrapolationmethodologies for plasma performance.

1. The global confinement scaling relation col-lapses data from a quite diverse range of mag-netic field strengths and sizes to a relation thathas only 15% RSME. A very careful discussionof attendant uncertainties is presented.

2. Neoclassical tearing modes limit the long pulse

β values and the observed magnetic fluctua-tions are in accord with theory. Because satu-rated islands depend only on β and ν∗, whichare similar in present day discharges and ITER,the values of βN ≥ 2.2 found in present dis-charges should be attainable in ITER as well.The neoclassical tearing mode β limit is not‘hard’, and theoretical prospects for stabiliza-tion of neoclassical tearing modes are just nowbeing tested experimentally.

3. The physics of the density limit is not clear,but values in excess of the Greenwald value aredifficult to achieve with gas puffing. Neverthe-less, pellet injection, particularly inside pelletlaunch [64], has led to discharges with core den-sities exceeding the Greenwald value. Reactorscale devices operate near the Greenwald den-sity value and would benefit from even higherdensities.

4. Experiments have demonstrated, and codeshave predicted, that divertor detachment canbe attained in present day devices. The samecodes give acceptable power dispersal for reac-tor scale experiments. Further code validationin initial proton plasmas is planned.

5. Disruption databases have led to simple ex-trapolation principles. The ITER Final DesignReport [4] attests to the fact that a reactorscale device is consistent with engineering solu-tions that withstand disruption effects, apartfrom a gradual erosion in the divertor chamber.Replacement of divertor cassettes is a designfeature.

6. Potential instabilities resulting from energeticalpha particles can be stabilized by modest pro-file readjustments.

7. Subject to some uncertainty, transport lossesfrom reactor scale experiments will exceed theH-mode power threshold, thereby establishingELMy H-mode operations.

All of the extrapolations suggest that an ITERlike experiment will produce a burning plasma andwill constitute a facility to carry out the researchneeded to support the design of an experimentalpower station. The ITER design has considerableflexibility to respond to research results and to opti-mize ELMy H-mode and advanced operations at thereactor scale. A major revision of this conclusion is


ITER Physics Basis

not foreseen as a result of future research in presentday devices.

5.3. Multiple pulse and erosion issues

A reactor scale plasma opens a new design issuein which consideration of the erosion of plasma fac-ing components over many pulses is a key elementin materials selection and the design of plasma fac-ing components for the first wall and divertor cham-ber [98]. Erosion can take place by several pro-cesses: physical and chemical sputtering in steadystate divertor operation; an increase of these pro-cesses under ELM heat loads; and vaporization andmelting of divertor and baffle wall as a consequence ofthe heat pulse associated with a disruption thermalquench. Our projection methodologies are again twodimensional divertor codes that build in laboratorymeasurements of sputtering processes. Disruptionthermal quench codes are also under development.On the basis of their predictions, calibrated againstDIMES [99] and other tokamak measurements [100],an initial selection of plasma facing materials hasbeen made for ITER, and is documented in theFinal Design Report [3–5] and Fig. 12. The principalmethodologies for optimizing this choice of plasmafacing materials are associated with the fundamen-tal device design, which has the flexibility to replacethe entire divertor chamber, and in experimentation,especially during the proton plasma stage when thedivertor chamber will be accessible, which will permitcalibration of divertor erosion and disruption meltloss codes as well as assessment of the core concen-tration of impurities. Our present projections sufficeto bound these issues.

Moreover, in order that the total in-vessel tritiuminventory not exceed 1 kg for safety considerations(using the ITER design as a guide), tritium should berecovered from tritium that has been implanted intoplasma facing components or that resides in code-posited layers along with carbon sputtered or sub-limated from the divertor strike plates [101]. BothTFTR and JET found a secular accumulation of tri-tium during their DT operations [11, 78]. Tritiumrecovery is addressed by design, which provides forhot surfaces in the divertor chamber area, and byexperimentation with recovery methods using basiclaboratory studies, present day tokamaks, and theproton plasma phase of a reactor scale tokamak [79].Discharge cleaning techniques involving oxygen havebeen proposed. Because it is difficult to extrapolatea complex surface chemistry from present day toka-

maks to an ITER like device with its very long pulse(which could self-clean plasma facing components),the precise nature of the tritium retention issue andits possible resolution must await reactor scale longpulse experiments. A flexible device design, incorpo-rating elements such as the ITER divertor cassettes,divertor chamber access to measure hydrogen depo-sition patterns during proton plasmas, and an abilityto carry out various plasma cleaning techniques withoxygen constitute the elements needed to resolve thetritium retention issue.

6. Concluding remarks

Section 4 argues that there are essential physicsdifferences between reactor scale devices and con-temporary tokamak research facilities. It follows thatexperiments on reactor scale devices are requiredto provide data to support design of an experimen-tal fusion power station. For the next step device,a judicious balance must be struck between min-imizing project costs, uncertainties in performanceprojections, entering regimes of new plasma physics,and minimal extrapolations to a reactor. One mustkeep in mind that this balance will depend on theoptimization goals and constraints. For example, thedesign that minimizes cost to attain a Q ≥ 10 burn-ing plasma experiment will differ from a design thatminimizes the cost per unit net power output. Acommon key issue is: can we ascertain whether aproposed design will attain an approximate balancebetween heating by thermonuclear alpha particlesand transport power losses, thus enabling a burn-ing plasma experiment? Based on the arguments pre-sented in this introduction and the material in thebulk of the article, we conclude that the extrap-olation principles now at hand, and set forth inthis article, provide guidance for a design operat-ing with ELMy H-mode plasmas. This guidance isas close to reliable as current facilities permit, andkey boundaries, such as the β limit, are describedaccurately enough. With regard to core confinementdegradation near the Greenwald density and H-modepower threshold limits, scaling arguments presentedin Section 4 indicate that direct examination of theseissues is possible only in a reactor scale device. Wenote that, in addition to core collisionality, plasmafuelling, divertor baffling, and edge shape (elongationand triangularity) for reactor scale facilities differappreciably from present day devices. Establishing acommon physics of inside pellet launch across a spec-trum of tokamak facilities should serve to support



planned operating densities. Turning to power dis-persal and particle control issues, agreement betweenavailable two dimensional codes and experiment issufficiently high so that these codes can determinethe relative variation in divertor performance withrespect to divertor leg length as a function of parame-ters such as the SOL density. Nonetheless, continuedresearch on present day devices is needed to con-firm and strengthen our confidence that the picturepresented by the collection of methodologies is accu-rate, as well as to reduce the uncertainties and toexplore prospects for active control measures such asstabilizing neoclassical modes and increasing densityabove the Greenwald value via inside pellet launch.Certain key theoretical and computational advanceswould be highly desirable to narrow uncertainties,for example a predictive theory for the L→H powerthreshold, a physics mechanism for the Greenwalddensity limit, and a predictive physics model for theeffect of divertor detachment and core density on coreconfinement.

With regard to advanced and steady state opera-tional modes, there does not appear to be any crucialconflict regarding device designs based on ELMy H-mode physics and the flexibility foreseen as necessaryto exploit whatever advanced modes future researchuncovers. But the data in hand at the time of writingdo not possess the commonality across tokamaks andperformance duration to make advanced operationsthe design basis for nominal plasma performance.

Overall, physics research during the ITER/EDAproject, which is summarized in this article, has pro-vided projection methodologies that permit mean-ingful performance assessments as well as studiesof cost–performance tradeoffs of candidate designsfor a burning plasma facility based on ELMy H-mode physics. It is understood that such a facilitywould return data on both inductive ELMy H-modeand advanced plasma operational scenarios that areessential to the design of a commercial fusion powerstation and cannot be obtained by contemporaryresearch facilities.

Appendix A

Although tokamaks are complex systems, thebasic parameters are determined by simple criteria[102]. These criteria are the requirements for ade-quate energy confinement, for sufficient MHD stabil-ity and plasma control to avoid frequent disruptions,for adequate shielding to protect the superconduct-ing coils from excessive nuclear heating and insulator

damage, and for acceptable stresses in the toroidalfield coils.

Let us turn first to energy confinement anddevelop a simple ignition criterion. The startingpoint is the assumption of ELMy H-mode opera-tion and an energy confinement time scaling relation,IPB98(y,1), which places data from a wide range oftokamaks onto a common curve with just 15% rmsdeviations. The expression for the energy confine-ment time is

τE = (0.0503 s)HHI0.91MA B0.15

T n0.4419 P−0.65

MW

×R2.05m κ0.72M0.13

(a

R

)0.57

(A1)

where the elongation is defined as κ = S0/(πa2)with S0 being the plasma cross-sectional area. HH

denotes a confinement multiplier, which is intro-duced to assess the sensitivity of results to varia-tion in confinement. This expression can be recastto express transport losses as a function of tokamakparameters

Ploss(MW) = (1.8× 105)n1.6

20 T 2.8610 R0.11

m κ0.8

B3.03T M0.37H2.86

H

×(

aB

I

)2.6(a

R

)1.49

(A2)

where T10 denotes the volume average temperaturein units of 10 keV. Plasma heating by fusion alphaparticles can be represented by

Pfusion = (2.8 MW)(nDT,20)2〈σv〉22Ra2κ (A3)

where 〈σv〉22 denotes the volume average fusionreactivity, including a negative contribution frombremsstrahlung losses, in units of 10−22 m3· s−1. Itis understood that 〈σv〉22 is a function of volumeaverage temperature. For the purposes of elemen-tary estimates, we neglect fuel dilution by impurityatoms and helium ash, and equate nDT = n. Ignitionrequires that

n0.420

(I

aB

)2.6

R2.89m

(a

R

)0.51

κ0.2B3.03M0.37H2.86H

×max 〈σv〉22

T 2.8610

> 6.4× 104. (A4)

The indicated maximization over temperature de-fines the most favourable temperature for ignition atfixed density. The optimized temperature is some-what below 10 keV. For the ITER FDR parameters


ITER Physics Basis

of Table 1, the left hand side achieves a value of14 × 104, thus assuring ignition with some marginfor fuel dilution.

Fuel dilution and impurity bremsstrahlung doplay a quantitative role and can be expressed byreplacing 〈σv〉22 by an effective reactivity 〈σv〉e,22

in Eqns (A4) and (A8). Figure 14 presents graphs ofthe combinations entering Eqns (A4) and (A8) for arepresentative fuel dilution and Zeff . Here, 〈σv〉e,22

is defined via

〈σv〉e,22 = η2〈σv〉α − (0.12)Zeff

√T10

where 〈σv〉a denotes the thermonuclear contributionto the reactivity, η = nDT /ne, and the second termgives bremsstrahlung losses.

0

0.5

1

1.5

2

1 10 100

n=1, Zeff=1n=0.925, Zeff=1.5

T (keV)

T (keV)

<σv

> eff

T–2.

86

0

0.5

1

1.5

2

2.5

3

1 10 100

n=1, Zeff=1n=0.925, Zeff=1.5

<σ

v>

eff

T–

3 .2

6

Figure 14. Plots of the combinations entering Eq. (A4)

and Eq. (A8) for a pure DT plasma and for fuel dilution

nDT /ne = 0.925 and Zeff = 1.5.

Criterion (A4) supposes the density to be fixed,but indicates that the higher the density, themore readily ignition is attained. There are twolimits on density: the Greenwald limit and thebeta limit. If one assumes the Greenwald limit,the ignition criterion can be expressed entirelyin terms of machine parameters and the non-dimensional plasma performance figure of merit

MGR = (n/nGR)0.4H2.86H (I/aB)3.0:

MGRB3.43R2.49m

(a

R

)0.11

κ0.2M0.37

×max 〈σv〉22

T 2.8610

> 10.1× 104. (A5)

The benefits of plasma shaping and elongation arecontained almost entirely in the normalized currentIMA/aBT. An analysis of the H-mode confinementdatabase has shown no explicit dependence of con-finement on triangularity beyond that implicit in thenormalized current. A simple version of ignition cri-terion (A5) is

(IMA

R

a

)>

(46HH

)

×

H0.05H R0.21

B0.14a0.04κ0.07M0.12(n/nGR)0.13

×[max(〈σv〉22/T 2.8610 )]0.33

(A6)

where the · factor on the right hand side variesvery slowly and has a value very close to unity. Thus,HHIMAR/a is a good figure of merit for ignition inITER class tokamaks. The value for the FDR designis HHIMAR/a = 60, which reflects fuel dilution, afinite operating space, and modest margin. For ana-lytic estimates, the formula(

HHIMAR/a

50

)3

=Q

Q + 5(A7)

should provide a relatively transparent way of esti-mating fusion performance consequences. The valueof IMAR/a = 50 is an ad hoc increase over Eq. (A6)and provides for fuel dilution, etc. When the lefthand side of Eq. (A7) exceeds unity, the plasma isignited and thermal balance occurs by decreasingdensity below the Greenwald value or by increasingthe temperature beyond the optimum value, therebydecreasing 〈σv〉T−2.86.

In circumstances where the density is limited byβN = β%(aB/I), the ignition criterion reads

MβB3.83R2.89

(a

R

)0.51

κ0.2M0.37

×max 〈σv〉22

T 3.2610

> 28× 104 (A8)

where the plasma performance figure of merit isdefined by Mβ = β0.4

N H2.86H (I/aB)3.0. It is notewor-

thy that confinement performance, as measured by



HH , is more important than either the βN limit orGreenwald normalized density in attaining ignition,according to (A8) and (A5), respectively

The purpose of this Appendix is to estimate theoverall physics parameters of density, temperature,magnetic field strength, and radial build requiredfor a tokamak to attain ignition. Radial build com-prises the dimensions of the major tokamak compo-nents, central solenoid, toroidal field coils, vacuumvessel, blanket and shield, and plasma in the equa-torial plane. Figure 15 illustrates the radial build ofa generic tokamak reactor.

OH Coil

TF Coil Shield Plasma TF

CoilShield

Cen

terli

ne

Bcoil

Schematic Radial Build of a Tokamak

Rcoil δBS a

R

Bcoil

Figure 15. Schematic radial build of a tokamak.

A combination of technological constraints arisingfrom the maximum field permitted for superconduct-ing magnets and the shield thickness needed to pro-tect the magnets with the physics requirements ofsufficient energy confinement and macroscopic sta-bility determines the radial build. It is fortunate forthe prospects of magnetic fusion energy that the sizeneeded to attain ignition is, in a broad sense, justthe size needed to accommodate the required shield.

A simple equation can be derived for the size of atokamak based on the build portrayed in Fig. 15, thedefinition of the MHD safety factor q95, the depen-dence on the plasma shape f , the aspect ratio A

(A = R/a, where R is the major radius and a isthe minor radius), the distance δBS between the coiland the plasma inner radius including blanket andshield, and the maximum field at the inner portionof the toroidal field coil, Bc, together with the 1/R

fall-off of the toroidal field (B = BcRc/R), where Rc

denotes the radius of the inner portion of the toroidalfield coil. From Fig. 15, the plasma major radius is

R = Rc + δBS + a. (A9)

A reasonably good fit for q95 for an elongated plasmais [103]

q95 =5a2BT

RIMAf (A10)

where f describes the role of plasma shape through

elongation κ, aspect ratio A, and triangularity δ:

f =1 + κ2(1 + 2δ2 − 1.2δ3)

2(1.17− 0.65A−1)

(1−A−2)2.

(A11)

More generally, Eq. (A10) constitutes a definition forf in terms of an equilibrium solution to the Grad–Shafranov equation, with the other parameters spec-ifying the solution.

Tokamak confinement performance is generallyenhanced by lowering q95 and increasing the shap-ing function f through elongation. Both these stepsserve to increase the plasma current. However, stabil-ity limits the extent to which q95 and κ can be varied.If q95 < 2.5, the disruptions become more frequentand confinement performance degrades relative toscaling expression (A1). Increases in elongation leadto vertical positional instabilities that are difficult tocontrol if κ > 1.7 and the poloidal field control coilslie outside the toroidal field coils as Fig. 1 portrays.Prudent and reliable operation of a tokamak reac-tor suggests the values q95 ≈ 3 and f ≈ 2.3. Thereare also limits on the maximum magnetic field at thecoil resulting from limiting fields for superconductiv-ity as well as mechanical stress. For Nb3Sn, we takeBc = 12 T; for NbTi a limit of Bc = 10 T is appro-priate. A length of δBS = 1.3 m is required to shieldthe superconducting magnets from radiation.

Let us introduce a characteristic length a0

a0 =(IMA[R/a])q95

5Bcf≈ 1.3 m (A12)

where the numerical value comes from the FDR igni-tion condition, IMAR/a = 60, and the limits dis-cussed above. Then, using Eqns (A9) to (A12), onecan calculate the size and aspect ratio of a tokamakthat just fulfils the ignition condition via

R

a=

Rc

a0= A, a =

(Rc + δBS

Rc − a0

)a0. (A13)

For an aspect ratio R/a = 3 design, the solution is

a = 2.6 m, Rc = 3.9 m,

δBS = 1.3 m, R = 7.8 m

(A14)

which lies close to the parameters of Table 1.Thus, the size and volume of ITER, or any elon-

gated tokamak, is determined by the six parameters:plasma current, Ip, and f , q95, A, Bc and δBS . Thefirst five of these enter through the combination a0,which governs the size of the device in terms of its


ITER Physics Basis

performance goals, expressed by IMAR/a, and itsshaping capability, expressed by f . These parame-ters are chosen to reflect the goals of ITER, and areinfluenced by physics and engineering constraints.The combination of aspect ratio and plasma cur-rent is largely determined by energy confinementrequirements according to Eq. (A6); MHD stabilityand energy confinement requirements determine theedge safety factor, q; the stress limits and thermalstability margin for the toroidal field coils lead toa limit on the maximum field at the coil, Bc; theneutron shielding requirements specify the thicknessof the shield and blanket, δBS . Physics choices foreach of these parameters have been made by theITER design team based on the information collectedand assessed by the international fusion communitythrough their representatives on the ITER ExpertGroups as described in this article.

Appendix B. Article 1, ITER EDAAgreement

(1) In accordance with this Agreement, its An-nexes and Protocols, the Parties, subject to theirlaws and regulations, shall conduct jointly theEngineering Design Activities (EDA) to produce adetailed, complete, and fully integrated engineeringdesign of ITER and all technical data necessary forfuture decisions on the construction of ITER. Suchdesign and technical data shall then be available foreach of the Parties to use either as part of an interna-tional collaborative programme or in its own domes-tic programme.

(2) The overall programmatic objective of ITER,which shall guide the EDA, is to demonstrate the sci-entific and technological feasibility of fusion energyfor peaceful purposes. ITER would accomplish thisobjective by demonstrating controlled ignition andextended burn of deuterium–tritium plasmas, withsteady state as an ultimate goal, by demonstratingtechnologies essential to a reactor in an integratedsystem, and by performing integrated testing of thehigh heat flux and nuclear components required toutilize fusion energy for practical purposes.

Appendix C. ITER Special WorkingGroup 1 – Review Report

Preamble

• In accordance with Article 10 of the ITER EDAAgreement,

• with reference to Sections 1 and 2 of Protocol 1,

• in the light of the Guidelines for SWG1imposed by the 1st ITER Council Meeting(Attachment 1),

• on the basis of the ITER Conceptual DesignActivities Final Report, ITER DocumentationSeries No 16, and the document referred totherein,

the Special Working Group has agreed as follows.

General constraints

The ITER detailed technical objectives and tech-nical approaches, including appropriate margins,should be compatible with the aim of maintainingthe cost of the device within the limits comparableto those indicated in the final report of the ITERCDA as well as keeping its impact in the long rangefusion programme.

ITER should be designed to operate safely and todemonstrate the safety and environmental potentialof fusion power.

Performance and testing

ITER should have a confinement capability toreach controlled ignition. The estimates of confine-ment capability of ITER should be based, as in theCDA procedure, on established favourable modes ofoperation.

Plasma performance

• ITER should demonstrate controlled ignitionand extended burn for a duration sufficient toachieve stationary conditions on all time-scalescharacteristic of plasma process and plasmawall interactions, and sufficient for achievingstationary conditions for nuclear testing ofblanket components. This can be fulfilled bypulses with flat top duration in the range of1000 s. For testing particular blanket designs,pulses of approximately 2000 s are desirable.

• ITER should also demonstrate steady stateoperation using non-inductive current drive inreactor relevant plasmas.

Engineering performance and testing

ITER should



• demonstrate the availability of technologiesessential for a fusion reactor (such as super-conducting magnets and remote maintenance);

• test components for a reactor (such as sys-tems to exhaust power and particles from theplasma);

• test design concepts of tritium breeding blan-kets relevant to a reactor. The tests foreseen onmodules include the demonstration of a breed-ing capability that would lead to tritium self-sufficiency in a reactor, the extraction of highgrade heat, and electricity generations.

Design requirements

The choice of parameters of the basic deviceshould be consistent with margins that give confi-dence in achieving the required plasma and engi-neering performance. The design should be suffi-ciently flexible to provide access for the introduc-tion of advanced features and new capabilities, andto allow for optimizing plasma performance duringoperation. The design should be confirmed by thescientific and technological database available at theend of the EDA.

An inductive pulse flat top capability, underignited conditions, of approximately 1000 s should beprovided. In view of the ultimate goal of steady stateoperation, ITER should be designed to be compat-ible with non-inductive current drive, and the heat-ing system required for ignition in the first phase ofoperation should have current drive capability.

To carry out nuclear and high heat flux compo-nent testing at conditions relevant to a fusion powerreactor:

• the average neutron wall loading should beabout 1 MW·m−2;

• the machine should be designed to be capableof at least 1 MW· a·m−2 to carry out longertime integral and materials tests.

It is desirable to operate at higher flux and fluencelevels. Within the engineering margins the ITERdesigners should examine the implications and pos-sibilities of exploiting a wider range of operationalregimes. The design of the permanent componentsof the machine should not preclude achieving fluencelevels up to 3 MW· a·m−2. For the second phase ofoperation, the design should include the capabilityof replacing the shield with a breeding blanket.

Operation requirements

The ITER operation should be divided into twophases:

• The first phase, the Basic Performance Phase,is expected to last a decade including a fewthousand hours of full DT operation. Thisphase should address the issues of controlledignition, extended burn, steady state opera-tion, and the testing of blanket modules. Itis assumed that for this phase there will bean adequate supply of tritium from externalsources.

– Controlled ignition experiments in ITERwill address confinement, stability andimpurity control in alpha particle heatedplasmas. Extended burn experiments willaddress, in addition, the control of fusionpower production and plasma profiles,and the exhaust of helium ash.

– The aim of current drive experiments inthis phase should be the demonstration ofsteady state operation in plasmas havingalpha particle heating power at least com-parable to the externally applied power.Using the heating systems in their cur-rent drive mode, non-inductive currentdrive should be implemented for profileand burn control, for achieving modes ofimproved confinement, and for assessingthe conditions and power requirements forthe above type of steady state operation.Depending on the outcome of these exper-iments, additional current drive powermay have to be installed.

– Functional tests of blanket modules inthis phase should consist of a few thou-sand hours on integral burn time, in par-allel with the physics programme, includ-ing continuous test campaigns of 3–6days at a neutron wall loading of about1 MW·m−2.

• The second phase, Enhanced PerformancePhase, is also expected to last a decade, withemphasis placed on improving overall perfor-mance and carrying out a higher fluence com-ponent and materials testing programme. Thisphase should address high availability oper-ation and advanced modes of plasma opera-tion, and may address reactor relevant blanket


ITER Physics Basis

segment demonstration. Operation during thisphase should include continuous testing cam-paigns lasting 1–2 weeks, and should accumu-late a fluence of at least 1 MW· a·m−2.

A decision on incorporating breeding for thisphase should be decided on the basis of theavailability of tritium from external sources,the results of breeder blanket testing, and expe-rience with plasma and machine performance.

The implementation of the Enhanced Perfor-mance phase should be made following a review ofthe results from the Basic Performance Phase andan assessment of the relative value of the proposedelements of the programme.

Final recommendation

That the above objectives can be achieved andthat the ‘Guideline for SWG1’ provided by the ITERCouncil at its first meeting will be complied withshould be confirmed by the Director in the outline ofthe design referred to in that Guideline.

Attachment 1 – Guideline for SWG1

The IC recommends as a general guideline forSWG1 that detailed technical objects and techni-cal approaches including appropriate safety margins,should be compatible with the aim of maintainingthe cost of the device within the limits comparableto those indicated in the final report of the ITERCDA as well as keeping its impact in the long-rangefusion programme.

The IC asks the Director to present an outlineof the design within about 10 months, at the timewhen a draft agreement of Protocol 2 should havebeen prepared by SWG-2.

References

[1] Holdren, J.P., Ann. Rev. Energy Environ. 16

(1991) 235.

[2] Raeder, J., et al., in Safety and Environmental

Assessment of Fusion Power (SEAFP), European

Commission, Brussels (1995).

[3] Final Design Report, ITER EDA Documentation

Series, IAEA, Vienna (1998) (to be published).

[4] Technical Basis for the Final Design Report, ITER

EDA Documentation Series, IAEA, Vienna (1998)

(to be published).

[5] ITER EDA Physics Design Description Documents

(available according to the intellectual property

provisions of the ITER Agreement), IAEA, Vienna

(1998).

[6] Post, D.E., et al., ITER Physics. ITER Documen-

tation Series, No. 21, IAEA, Vienna (1991).

[7] ITER EDA Agreement and Protocol 1, ITER EDA

Documentation Series No. 1, IAEA, Vienna (1992).

[8] JET Team (presented by M. Keilhacker), Plasma

Phys. Control. Fusion 39 (1997) B1.

[9] Start, D.F.H., et al., Phys. Rev. Lett. 80 (1998)

4681.

[10] Strachan, J.D., et al., Phys. Rev. Lett. 72 (1994)

3526.

[11] JET Team (presented by A. Gibson), Phys. Plas-

mas 5 (1998) 1839.

[12] Hawryluk, R.J., Rev. Mod. Phys. 70 (1998) 537.

[13] Wesson, J., Tokamaks, Clarendon Press, Oxford

(1997) 109.

[14] Kadomtsev, B.B., Sov. J. Plasma Phys. 1 (1975)

295.

[15] Porcelli, F., Rosenbluth, M.N., Plasma Phys. Con-

trol. Fusion 40 (1998) 481.

[16] H-Mode Database Working Group (presented by

F. Ryter), Plasma Phys. Control. Fusion 38 (1996)

1279.

[17] Fukuda, T., et al., Nuclear Fusion 37 (1997) 1199.

[18] Righi, E., et al., Plasma Phys. Control Fusion 40

(1998) 857.

[19] Burrell, K.H., Phys. Plasmas 4 (1997) 1499.

[20] Zeiler, A., Drake, J.F., Rogers, B., Phys. Plasmas

4 (1997) 2134.

[21] Scott, B., et al., in Fusion Energy 1998 (Proc.

17th Fusion Energy Conf. Yokohama, 1999),

IAEA, Vienna (1999). Paper IAEA-F1-CN-69

Th1-05.

[22] Igitkhanov, Y., Janeschitz, G., Pacher, G.W., et

al., Plasma Phys. Control. Fusion 40 (1998) 845.

[23] Cordey, J.G., in Fusion Energy 1998 (Proc. 17th

Fusion Energy Conf. Yokohama, 1999), IAEA,

Vienna (1999). Paper IAEA-F1-CN-69 EX7-1.

[24] Boucher, D., in Proceedings IAEA Technical Com-

mittee Meeting on Advances in Simulations and

Modeling of Thermonuclear Plasmas June 1992

(Montreal, Canada) 142.

[25] Connor, J.W., et al., in Fusion Energy 1996 (Proc

16th Int. Conf. Montreal, 1997) Vol. 2, IAEA,

Vienna (1997) 935.

[26] Kinsey, J.E., Bateman, G., Phys. Plasmas 3 (1997)

3344.

[27] Dimits, A.M., Phys. Rev. Lett. 77 (1996) 71.

[28] Beer, M.A., et al., Phys. Plasmas 4 (1997) 1792.

[29] Sydora, R.D., et al., Plasma Phys. Control. Fusion

38 (1996) A281.

[30] Vandenplas, P.E., et al., J. Plasma Physics 59

(1998) 587.



[31] JT-60U Team (presented by H. Shirai), Phys. Plas-

mas 5 (1998) 1712.

[32] Rice, B.W., et al., Plasma Phys. Control. Fusion

38 (1996) 869.

[33] Ward, D.J., Bondeson, A., Phys. Plasmas 2 (1995)

1570.

[34] Navratril, G.A., et al., Phys. Plasmas 5 (1998)

1855.

[35] Sauter, O., et al., Phys. Plasmas 4 (1997) 1654.

[36] La Haye, R.J., Sauter, O., Nuclear Fusion 38

(1998) 987.

[37] La Haye, R.J., et al., Nuclear Fusion 37 (1997) 397.

[38] Gunter, S., et al., Nuclear Fusion 38 (1998) 1431.

[39] Huysmans, G.T.A., et al., in Controlled Fusion

and Plasma Physics 1997 (Proc. 24th Eur. Conf.

Berchtesgaden, 1997), Vol. 21A, Part IV, European

Physical Society, Geneva (1997) 857.

[40] Zohm, H., Phys. Plasmas 4 (1997) 3433.

[41] Perkins, F.W., et al., in Controlled Fusion and

Plasma Physics 1997 (Proc. 24th Eur. Conf.

Berchtesgaden, 1997), Vol. 21A, Part III, Euro-

pean Physical Society, Geneva (1997) 1017.

[42] Zohm, H., et al., in Fusion Energy 1998 (Proc. 17th

Int. Conf. Yokohama, 1999), IAEA, Vienna (1999).

Paper IAEA-F1-CN-69.

[43] La Haye, R.J., et al., Nuclear Fusion 32 (1992)

2119.

[44] Schuller, F.C., Plasma Phys. Control. Fusion 37

(1995) A135.

[45] Wesson, J. A., et al., Nuclear Fusion 29 (1989) 641.

[46] Wesson, J.A., et al., Phys. Rev. Lett. 79 (1997)

5018.

[47] Pestchanyi, S., et al., in Controlled Fusion

and Plasma Physics 1997 (Proc. 24th Eur.

Conf. Berchtesgaden, 1997), Vol. 21A, Part III,

European Physical Society, Geneva (1997)

981.

[48] Hassanein, A., et al., J. Nucl. Mater. 241–243

(1997) 288.

[49] Rosenbluth, M.N., Putvinski, S.V., Nuclear Fusion

37 (1997) 1355.

[50] Pautasso, G., et al., in Controlled Fusion and


Prague, 1998), European Physical Society, Geneva

(1998). B169PR.

[51] Yoshino, R., et al., Nuclear Fusion 37 (1997) 1161.

[52] Zohm, H., Plasma Phys. Control. Fusion 38 (1996)

105.

[53] Connor, J.W., Plasma Phys. Control. Fusion 40

(1998) 191.

[54] Connor, J.W., Plasma Phys. Control. Fusion 40

(1998) 531.

[55] Osborne, T.H., et al., in Controlled Fusion and


Berchtesgaden, 1997), Vol. 21A, Part III, Euro-

pean Physical Society, Geneva (1997) 1101.

[56] Takase, Y., et al., Phys. Plasmas 4 (1997) 1647.

[57] Rettig, C.L., et al., Phys. Plasmas 5 (1998) 1727.

[58] Voitsekhovitch, I., et al., Nuclear Fusion 37 (1997)

1715.

[59] Fujita, T., et al., Nuclear Fusion 38 (1998) 207.

[60] Gruber, O., et al., in Fusion Energy 1998 (Proc.

17th Int. Conf. Yokohama, 1999), IAEA, Vienna

(1999). Paper IAEA-CN-69/OV4/3.

[61] Ide, S., et al., in Fusion Energy 1998 (Proc. 17th

Int. Conf. Yokohama, 1999), IAEA, Vienna (1999).

Paper IAEA-CN-69/CD1/4.

[62] Garofalo, A.M., et al., in Controlled Fusion and


Prague, 1997), European Physical Society, Geneva

(1997). B094PR.

[63] Greenwald, M., et al., Nuclear Fusion 28 (1988)

2199.

[64] Lang, P.T., et al., Phys. Rev. Lett. 79 (1997) 1487.

[65] Maingi, R., et al., in Fusion Energy 1998 (Proc.


(1999). Paper IAEA-CN-69/EXP2/10.

[66] Janeschitz, G., et al., Fusion Energy 1996, IAEA,

Vienna 2 (1997) 755.

[67] Mertens, V., et al., Fusion Energy 1996, IAEA,

Vienna 1 (1997) 413.

[68] Loarte, A., et al., Nuclear Fusion 38 (1998) 331.

[69] Schneider, R., et al., in Fusion Energy 1998 (Proc.


(1999). Paper IAEA-CN-69/THP2/05.

[70] Mahdavi, M.A., et al., Fusion Energy 1996, IAEA,

Vienna 1 (1997) 397.

[71] Kukushkin, A., et al., Fusion Energy 1996, IAEA,

Vienna 2 (1997) 987.

[72] Kukushkin, A.S., et al., in Fusion Energy 1998

(Proc. 17th Int. Conf. Yokohama, 1999), IAEA

Vienna (1999). Paper IAEA-CN-69/ITERP1/13.

[73] Hubbard, A.E., et al., Phys. Plasmas 5 (1998)

1744.

[74] Jacquinot, J., et al., Nuclear Fusion 39 (1999) 235.

[75] Miller, R.L., et al., Plasma Phys. Control. Fusion

40 (1998) 753.

[76] Kaufmann, M., et al., Fusion Energy 1996, IAEA,

Vienna 1 (1997) 79.

[77] Chiocchio, S., et al., in Fusion Engineering 1997

(Proc. 17th IEEE/NPSS Symp. San Diego, 1997),

Vol. 1 (1997) 331.

[78] Mueller, D., et al., in Fusion Engineering 1997

(Proc. 17th IEEE/NPSS Symp. San Diego, 1997),

Vol. 1 (1997) 279.

[79] Federici, G., Anderl, R.A., Andrew, P., Brooks,

J.N., Causey, R.A., et al., J. Nucl. Mater. 266–

269 (1999) 14.

[80] Nazikian, R., Fu, G.Y., et al., Phys. Rev. Lett. 78

(1997) 2976.

[81] Kramer, G.J., Saigusa, M., et al., Phys. Rev. Lett.

80 (1998) 2594.


ITER Physics Basis

[82] Cheng, C.Z., et al., Fusion Energy 1996, IAEA,

Vienna 2 (1997) 953.

[83] Candy, J., et al., Phys. Plasmas 6 (1999) 1822.

[84] Itoh, S., et al., Fusion Energy 1996, IAEA, Vienna

3 (1997) 351.

[85] Waltz, R.E., et al., Phys. Plasmas 5 (1998) 1784.

[86] JT-60 Team (presented by T. Fujita), Plasma

Phys. Control Fusion 39 (1997) B75.

[87] Rogers, J.H., et al., in Radiofrequency Power in

Plasmas 1997 (Proc. 12th Topical Conf. Savannah,

1997), AIP, New York (1997) 13.

[88] Eriksson, L.-G., et al., Nuclear Fusion 33 (1993)

1037.

[89] Hemsworth, R., et al., in Fusion Energy 1996,

IAEA, Vienna 2 (1997) 927.

[90] Erckmann, V., Gasparino, U.V., Plasma Phys.

Control. Fusion 36 (1994) 1869.

[91] Rey, G., et al., in Fusion Engineering 1997 (Proc.

17th IEEE/NPSS Symp. San Diego, 1997), Vol. 1

(1997) 433.

[92] Costley, A.E., et al., Diagnostics for Experimen-

tal Thermonuclear Fusion Reactors, Plenum Press,

New York (1998) 41.

[93] Wesley, J., et al., Fusion Technol. 32 (1997) 495.

[94] Boucher, D., et al., Fusion Energy 1996, IAEA,

Vienna 2 (1997) 945.

[95] Petty, C.C., Luce, T.C., Nuclear Fusion 37

(1997) 1.

[96] Kotschenreuther, M., et al., Phys. Plasmas 2

(1995) 2381.

[97] Petty, C.C., et al., Phys. Rev. Lett. 83 (1999) 3661.

[98] Pacher, H.D., et al., J. Nucl. Mater. 241–243

(1997) 255.

[99] Whyte, D.G., et al., J. Nucl. Mater. 241–243

(1997) 660.

[100] Guo, H.Y., et al., J. Nucl. Mater. 241–243 (1997)

385.

[101] Brooks, J.N., et al., J. Nucl. Mater. 241–243

(1997) 294.

[102] Post, D.E., Uckan, N.A., Fusion Technol. 21 (1992)

427.

[103] Uckan, N.A., et al., ITER Physics Design Guide-

lines, IAEA, Vienna (1990) 10.

(Manuscript received 8 September 1998Final manuscript accepted 25 May 1999)

E-mail address of F. Perkins:[email protected]

Subject classification: M0, P0
