DEALING WITH MEGAWATT BEAMS*†
Fermilab,MS 220, P.O. Box 500
Batavia, IL 60510, U.S.A.
*Work supported by Fermi Research Alliance, LLC under contract No. DE-AC02-07CH11359
with the U.S. Department of Energy.
†Presented at 10thWorkshop on Shielding Aspects of Accelerators, Targets and Irradiation Facilities
(SATIF-10), June 2-4, 2010, CERN, Geneva, Switzerland.
Dealing with MegaWatt Beams
Nikolai V. Mokhov
Fermi National Accelerator Laboratory, Batavia, IL 60510 USA
The next generation of accelerators for MegaWatt proton, electron and heavy-ion beams puts
unprecedented requirements on the accuracy of particle production predictions, the
capability and reliability of the codes used in planning new accelerator facilities and
experiments, the design of machine, target and collimation systems, detectors and radiation
shielding and minimization of their impact on environment. Recent advances in code
developments are described for the critical modules related to these challenges. Examples are
given for the most demanding areas: targets, collimators, beam absorbers, radiation
shielding, induced radioactivity and radiation damage.
The next generation of accelerators for MegaWatt proton, electron and heavy-ion beams
moves us into a completely new domain of extreme specific energies of ~0.1 MJ/g and
specific power up to 1 TW/g in beam interactions with matter. Challenges arise also from
increasing complexity of accelerators and experimental setups, as well as design, engineering
and performance constraints. All these put unprecedented requirements on the accuracy of
particle production predictions, the capability and reliability of the codes used in planning
new accelerator facilities and experiments, the design of machine, target and collimation
systems, detectors, radiation shielding and minimization of their impact on environment. This
leads to research activities involving new materials and technologies, and also code
developments whose predictive power and reliability being absolutely crucial.
A list of high-power proton and heavy-ion accelerators includes those under
Operation: ISIS, PSI, J-PARC and SNS with 0.2 to 1 MW beam power and
upgrade plans up to 1-3 MW.
Construction: CSNS, 0.1-0.2 MW.
Design: FAIR and FRIB with proton to uranium beams up to 0.4 MW, ESS with a
few MW beams, and Project-X with proton beams up to 4 MW.
Consideration: subcritical accelerator-driven systems with proton beams up to 10
Another category is high-energy colliders, operating: proton/antiproton 2 MJ beams of
the Tevatron, and up to 350 MJ proton beams of the LHC, and planned: e+e- (ILC and CLIC,
up to 20 MW) and
+ - with a 4-MW proton source.
1. Critical Areas
Components of the three critical systems of the accelerators listed are especially vulnerable to
the impact of the high-power beams: target stations, beam absorbers and collimators.
The principal issues include: production and collection of maximum numbers of particles
of interest; suppression of background particles in the beamline; target and beam window
operational survival and lifetime (compatibility, fatigue, stress limits, erosion, remote
handling and radiation damage); protection of focusing systems including provision for
superconducting coil quench stability; heat loads, radiation damage and activation of
components; thick shielding and spent beam handling; prompt radiation and ground-water
activation. The most challenging one is a choice of a target technology. Fundamentally, it
depends on a peak power density and power dissipation in the target material as illustrated in
Fig. 1 generated for the FRIB project . It is clear how severe this problem is for intense
heavy-ion beams (up to uranium) of a small spot size impinging on a target or beam window
material. As an example of this extreme, the uranium ion beam power at SIS-100 of the FAIR
project  will be up to 0.1 TW, ion energy range 0.4 to 27 GeV/u, peak specific energy and
power in a lead target 0.1 MJ/g and 1 TW/g, respectively. Rotating multi-slice carbon disks,
liquid lithium or lead targets for heavy ions and open mercury jets for proton beams at
neutrino factories, no-material windows are just several of technologies considered in such
projects. Radiation damage to solid targets, downstream magnets and auxiliary equipment is
identified as one of the key issues in these systems along with reliability and cost of complex
remote handling equipment.
Figure 1: Choice of target technology (Courtesy: W. Mittig)
Absorbers for misbehaved beams along the beamlines, abort beam dumps and those
downstream of the production targets and interaction regions (at linear colliders) are another
challenging systems in the MegaWatt accelerators. These should be able to withstand an
impact of beams of up to full power, say, 0.2 to 20 MW, without destruction over a designed
life-time (at least a few years), fully contain the beam energy, and execute the initial shielding
functions. The absorber technologies for high-intensity beams include:
A laminated graphite core in a cooled aluminum shell. It is proven in more than
20-years of an operational experience at the Tevatron, with the peak instantaneous
temperature rise of T ~1000 oC per pulse. The core is contained in steel
shielding surrounded by concrete. A similar design is used at the LHC with the
beam swept in a spiral during the abort pulse.
A stationary beryllium, aluminum or nickel wall liquid-cooled dump. Thin two-
layer walls are arranged in a V-shape, with a liquid flowing between the layers
and a beam hitting the walls at a small grazing angle. It is successfully used in
several high-power proton machines, but is not feasible for intense heavy-ion
beams: estimated life-time due to radiation damage is only 3 months for 0.4 MW
uranium beam at FRIB .
A water-cooled aluminum-shell rotating drum considered for the FRIB project,
with a life-time estimated as 5 years.
A water-vortex beam absorber considered for an 18-MW electron beam at ILC.
The beam is rastered with dipole coils to avoid water boiling. The entrance beam
window and catalytic recombination are of a serious concern in this design.
Only with a very efficient beam collimation system can one reduce uncontrolled beam
losses in the machine to an allowable level, thus protect personnel and components, maintain
operational reliability over the life of the machine, provide acceptable hands-on maintenance
conditions, and reduce the impact of radiation on environment, both at normal operation and
accidental conditions. Collimators – as the last line of defense in high-power accelerators -
must withstand a predefined fraction of the beam hitting their jaws and - at normal operation -
survive for a time long enough to avoid very costly replacements. Design of the collimation
systems is especially challenging at the high-energy colliders. At the LHC, the overall
collimation efficiency should be better than 99.9% (this value is typical of the Tevatron
collider). The system should manage substantial beam losses of 0.5 MW at normal (slow)
operation, 20 MW in a transient regime of ~1 ms, and up to 5 TW at a beam accident (1 MJ in
0.2 s into 0.2 mm2 area). Novel collimation techniques include crystal channeling and
multiple volume reflection, hollow electron beam scraper, volume reflection radiation,
rotatable/consumable collimators, and marble shells to mitigate hands-on maintenance
2. Modeling Challenges
Particle transport simulation tools and the physics models and calculations required in
developing relevant codes are all driven by these demanding applications. This puts
unprecedented requirements on the accuracy of particle production predictions, the capability
and reliability of the codes used. The challenge is detailed and accurate (to a % level)
modeling of all particle interactions with 3-D system components (up to tens of kilometers of
the lattice in some cases) in energy region spanning up to 15 decades as a basis of accelerator,
detector and shielding designs and their performance evaluation, for both short-term and long-
Five general-purpose, all-particle codes are capable of this - FLUKA, GEANT, MARS,
MCNPX and PHITS - and they are used extensively worldwide for accelerator applications. A
substantial amount of effort (up to several hundreds of man-years) has been put into
development of these codes over the last few decades. The user communities for the codes
reach several thousands of people worldwide. All five codes can handle very complex
geometry, have powerful user-friendly built-in GUI interfaces with magnetic field and tally
viewers, and variance reduction capabilities. Tallies include volume and surface distributions