Chapter XII Spallation Neutron Sources

alligatorsavoryΠολεοδομικά Έργα

26 Νοε 2013 (πριν από 3 χρόνια και 11 μήνες)

319 εμφανίσεις


Chapter X
II

Spallation Neutron Sources

XII.1

Introduction

A
n intense proton beam
, on hitting
a metal target,
produces
a

copious supply of neutrons

by


spall
ation

, a
word
which

describes
material fragments

being

ejected

from
a
surface
after

impact
.

T
h
is
C
hapter

is mainly concerned with the

high power

proton driver
s

and
targets used to

create

spallation neutrons
, which are
used
for

a variety of

neutron
scattering
experiments
.

In a spallation source, protons are typically accelerated to 1 GeV
in a rapid cycling synchrotron
, cyclotron

or linac
sometimes
followed by a pulse
-
compression accumulator ring. The pulse is then fired at a heavy metal target,
where
each
proton produces 2
0

30 neutrons via primary or secondary reactions. The neutrons are
then slowed down to energies typically in the milli
-
electron
-
volt (meV) range by means
of a “moderator”: a material made of light atoms such as water or liquid hydrogen.


Some of these spal
lation sources, e.g. ISIS, SNS and JPARC produce pulses of
neutrons less than 2 μs in length, others such as the future ESS are designed for 1.5 to 2
ms, while sources based on cyclotrons, like SINQ, produce a continuous stream of
neutrons like a nuclear r
eactor
.(
We shall
describe these different spallation sources later.)
A short pu
ls
e of beam will generate a spectrum of neutrons with different velocity and
wavelength which
arrive
at
a
specimen

downstream at different times.

All of these “proton drivers”
aim for the highest possible beam current
.

Typical
beam power and energy specifications for the examples mentioned above
are: ISIS

(160
-
200 kW, 0.8 GeV), SNS (1.4 MW, 1 GeV, upgrade to 3 MW, 1.3 GeV),

JPARC (0.5
-
1
MW, 3 GeV), SINQ (1.6 MW, 0.59 GeV) and
ESS (5 MW, 2.GeV).

The synchrotrons
use

multi
-
turn, Hˉ injection
build up

intense circulating beams.

In this ingenious
procedure, negatively charged Hˉ ions (protons with two electrons) are deflected by a
magnet to join the circulating orbit, like traffic

entering a roundabout by a slip road.
They are immediately

stripped of
their

two excess electrons as they pass though a thin
foil

to follow the proton traffic around the
roundabout
.
When they return to the magnet
,
after one turn of the roundabout, they
are simple positively charged protons
.

and

the
same
magnet
now
deflects them

in
the opposite sense to help keep them
circulating

together
with the other protons
.
While the beam is at low energy s
crapers and collimators
collect

the halo of the accelerated beam to avoid losses at high energy.

Losses at higher
energy

lead to radioactivity, but also, loss of even a tiny fraction of the beam during acceleration
might easily burn a hole in the beam pipe
,

and even when the beam has be
en extracted
the danger is not passed. Every trick in the accelerator builder’s repertoire must be used
to ensure that the beam does not break into wild instabilities due to its electrical image in
the walls of the vacuum chamber as well as in every compon
ent through which it passes.

There is
inevitably
a powerful thermal shock as megawatts of beam power impinge on
less than a kilogram of metal that forms the spallation targe
t
.

To handle the shock in the
most extreme case

liquid metal

(mercury) targets

are used.

These have the advantage that
they
reform and are their

own cooling fluid
.

The proton beam may provide other particles besides neutrons, such as
radioactive ions and low energy muons and neutrinos,

Such intense, low energy proton
sources may also produce energy by stimulating fission in thorium fueled reactors, or be
used to clean up air, or water, or even eliminate nuclear wastes; these applications are
described later in Chapter XV on Energy and En
vironment.
Accelerators of a similar
power but with even shorter bunches and a somewhat higher energy, can create muons
for a muon collider or for neutrino decay beams, to be sent via the earth to distant
locations.
(
See

Chapter VII for details of these a
pplications.)


XII.2
Neutron Scattering

T
he
technique

of neutron diffraction began in the 1940’s and
19
50’s

at

the

research
reactors
,

where a copious supply

of low energy neutrons

was

available
.


The value of
the

technique
was

recognized
in 1994

when

a
Nobel Prize
was

awarded to Clifford Shull and
Bertram Brockhouse
.

The diffraction pattern of scattered
neutrons from
, for example, a
crystal, depends

on their
wavelength (
remember, particles may be seen to act as waves in the right
conditions!)
.

At suffi
ciently low energy, the
wavelength

of
neutrons is

comparable to
those

of

x
-
rays
,

and they scatter into diffraction patterns

similar to

those seen
at

synchrotron

radiation sources.

Neu
tr
on

wavelength

is linked to its velocity and
is

all
important

for diffraction

studies. The w
avelength
should

be comparable to the

spacing between the nuclei in the
structure of the sample.

S
electing neutrons of the correct velocity,
or tagging neutrons by

their velocity, is
therefore
critical.

A research reactor is
at a disadvantage in
this respect
in
comparison with an accelerator, as the
accelerator is

able to produce very short, intense
pulses of neutrons.


In

a

reactor
,

neutrons of many velocities emerge continuously
from
a port in the shield wall.
Selecting
velocities is di
ffi
cult

and although possible with
mechanical choppers the
e
ffi
ciency of neutron use
is very low. O
n the
other

hand, a
pulse

of protons from an accelerator can be as short as

a microsecond or less
.
When this hits a
target the neutrons it pr
oduces
,

all leave at about the same time.
As the speed of

very
low
-
energy neutrons is only a few 100 meters per second, it takes many milliseconds for
the
m to travel to a

target station, typically a few 10’s of meters

away
.
By measuring
the

time of flight

from
the source

to a

de
tector

one may tag every neutron with
its

wavelength
.

T
he net e
ffi
ciency

of neutron use is
therefore
much higher than with
neutrons

from a
reactor.
In
fact, the
productivity of a 1 MW
,

short pulse
spallation source
is comparable to t
hat of a 60 MW

research reactor.


Neutrons scatter off nuclei of target atoms and not off the electron clouds that
surround them. One consequence of this is that neutrons are more sensitive to elements
with light nuclei than are x
-
ray sources. For example neutron diffraction is able to
dis
tinguish hydrogen and deuterium in the presence of heavier elements, a feat, only just
possible with x
-
rays. Spallation neutron and synchrotron radiation sources are therefore
complementary, and serve the same communities of material scientists in differen
t ways.



Figure
12
.1
is an example of a complex

structure mapped with neutron

beams.
Other applications of neutron diffraction are the

study of
high te
mperature
superconductivity,
the
magnetic properties of materials,
the
quality of welds and strains
in critical structures such as aircraft wings
,

and for research in earth sciences, biology,
and archeology. Neutron scattering has been used to study the st
orage of hydrogen,
to
weld difficult materials,
studies of
magnetism, peering into polymers, cutting and pasting
enzymes, investigating flow patterns around an airplane engine, studying various forms
of ice, investigating superconductors, and on and on.

XII.3

B
rief
History of spallation sources


Short

pulse
spallation sources
were first proposed

by J Carpenter

at ANL

in 1972
, using
a
rapid cycling
synchrotron (RCS) as the proton driver
.

A 450 MeV, 6.4 kW, 30 Hz RCS

was built
for the
ir

IPNS source
, and it worked
successfully for

over 26 years
.
The KENS
source at KEK, Tsukuba, Japan
soon
follow
ed
, using protons from a 20 Hz, 500 MeV,
3.5 kW bo
oster RCS

to feed a new spallation target.


Though ANL had plans for a
more

intense

source, the next

advance was

the
building of

a

160 kW RCS (a factor 25 increase in beam power) for
the ISIS source at
RAL (1979
-
85)
.

Shortly afterwards

(1981
-
86)
,

a 12 Hz, 797 MeV, 80

kW
compress
or
ring (
the
PSR)

was built
at the end
of
the LANL
linac
for the
LANSCE

(Los Alamos
Nuclear Science Center)

source
.
The linac

(
1965
-
72)

was the far reaching concept of Luis
Rosen to build the world’s largest and most powerful proton linac for a
π
-
meson factory.
The

beam energy
, current and pulse rate are 800 MeV,

1 mA and

120 Hz
, respecti
vely.

(See
Fig
.

12
.2
)
.

Now this
is very much a spallation source with many experiments
performed at the facility
.

.


XII.4
ISIS and SINQ


ISIS

is a name used locally for the River Thames, and also that of an Egyptian goddess
(who could restore life to the dead).
This is

perhaps appropriate

since

the 50

Hz
, 800
MeV RCS for the ISIS source
was

partially

made of equipmen
t used in two earlier
accele
rators. It opened for users in 1985, and now

typically serv
es 1,600 users
,

doing 600
experiments each year.
(S
ee sidebar for G. Rees
who was its principal designer).
Not only
does the facility produce neutrons
,

but it also has a
n

active program in muon

phy
sics. In
fact
, a

cooling experiment
, MICE,

for a neutrino
factory or a muon collider
, is stationed
on a new muon beam line, as

described in Chapter VII.

The success of the ISIS science
program prompted the development of Megawatt sources
.


For the SINQ sou
rce, a
t the Paul Schering Institute (PSI)

in Switzerland,

a > 2
MW target station was built and
the 590 MeV cyclotron

beam power

was raised

to

1.8
MW, giving a flux of

14
10
neutrons/cm
2

sec. (See Fig
.
12
.3).

The beam is
continuous
and not

pulsed
.

Hence the
efficiency
is low and more comparable to that

of a research
reactor.



XII.5
The SNS and J
-
PARC

The first short pulse spallation source to reach 1 MW is

the SNS,
at Oak Ridge, USA, and
it has operated at this level for a few years
(
S
ee Fig.
12
.4
). Its specifications are an
average p
roton beam current of 1.4
mA at

1 GeV
,

so it

will

deposit 1.4 MW of power in
the target and its surroundings. Each proton pulse on target is 0.7 microseconds long,
short enough to provide excellent wavelength tagging

of each neutron,
and
with enough
intensity to produce excellent images of the most complex material stru
c
tures.
The
injector for the

accumulator ring embodi
es all the latest technology. A
n



ion source is
followed by an RFQ
, a drift tube linac, a cavity
coupled linac and

a
331 m long
superconducting linac.

More than 1000 turns of

Hˉ beam are fed
to

an accumulator

ring,
of 248 m circumference,

and then

extracted and

shot at a

liquid

mercury target.

There are
plans to increase the source power to 3 MW

at 1.3 GeV

in future years
.


The
1.5 B$
,

SNS
c
ost
is comparable with
that

of large

particle physics
’ machines.
Of the various

aspects of accelerator physics and engineering involved in constructing
and operating such a major fa
cility, the complexity of ac
tive

handling for maintenance
and

the
beam loss

suppression

are
the most demanding. If the co
mplex is not to become
too radioactive

to handle,
uncontrolled losses must be

kept

below
1

watt per meter length.


It is surprising to find that the SNS is the first US accelerator

to be a joint e
ff
ort of
man
y
of the major national laboratories under the Department of Energy
.

Berkeley
produced
the ion source and
RFQ
, Los Alamos,
the normal
-
conducting Alvarez and

side
coupled linacs, Je
ff
erson,
the superconducting linac, Brookhaven
,
the accumulator ring,
Oak Ridge
,

the target,

and Argonne

coordinated the

neutron
-
scattering instrumentation.

The J
-
PARC facility at JAERI, Tokay, Japan, is unique in that it serves many
d
ifferent sciences at one site. The

large accelerator complex consists of
a
n
ion source,
RFQ,
linac (
200 (400) MeV), RCS (3 GeV) and a slower cycling main ring synchrotron
(50 GeV, 1,5
67 m circumference).
Parts of the J
-
PARC

facility are shown in
Fig
s
.

12
.5
and
12
.6
.

The 50 GeV ring caters for n
uclear physics

and for hypernuclei, neutrino
oscillation and rare decay experiments.
Neutrino

beam
s

have already been directed

to

far
-
away Super Kamiokande
.

The 3 GeV RCS serves

both

a muon beam research facility and
a spallation neutron

facility.

A liquid mercury target is used to produce

the spallation
neutrons,
and the

eventual goal for the RCS is to provide

1 MW

of beam power
.



XII.6

Spallation Source Target Stations


A major engineering effort is involved in building the target stations and neutron beam
lines, of which there may be up to twenty. Extensive concrete
-
steel shielding surrounds
the target and its associated moderators, r
eflectors and cooling circuits, all of which are
installed and serviced by remote handling (whereas the accelerators use only active
handling techniques). There are small shielding gaps, adjacent to moderators, for the
emergent neutron beams. Early targets

used a water cooled heavy metal, such as lead,
tungsten, tantalum, or depleted uranium, often clad with zircalloy. Typical diameters
were 0.1 to 0.2 m and typical lengths to contain the proton’s nuclear cascade were 0.5 m.


Tritium produced in cooling

water circuits had to have a controlled release. The targets
have to withstand shock waves due to incident protons, especially when pulses are short,
and also the damage due to radiation and corrosion. New designs have been needed for
the Megawatt sources
, where liquid metals such as mercury and lead
-
bismuth are used.


XII.7
The Future


The

next generation spallation source has been
discussed

in Europe for many years

and
it
has

finally

been decide
d t
o build one at

Lund

in

southern

Sweden.
After
the

initial study,
there were
plans

to build a
10
MW,

1.
334

GeV

Hˉ linac, with 5 MW

for
a long pulse
target (2.5 ms, 16.666 Hz) and 5 MW for a short pulse target (< 1 μs, 50.0 Hz, via

two
accumulator rings.) After ten more years of debate, the design has been

scaled down to a
5 MW,
with a
2.5 GeV proton linac, for a (1.5 ms, 20 Hz), long pulse source. The target
plans to use a slowly rotating structure, with tungsten target elements. Long pulse
sources, though not so efficient as short pulse sources for epithe
rmal neutrons are
competitive for thermal and slow neutrons.

China is

building

its CSNS spallation source. It is similar to that of a 1994, Austrian
source,
AUSTRO
N
,

planned for a
scientific research

center

for

central Europe
.
A 25 Hz,
100 kW, 1.6 GeV
proton synchrotron with a 70 MeV, Hˉ injector linac is under design at
IHEP, Beijing, for later construction in Dongguan, south China. A subsequent linac
energy of 130 MeV will allow the output beam power to be increased to 200 kW.