introduction to thick-liquid-wall chambers*

Wayne R. MeierLawrence Livermore National Lab

Per PetersonUC Berkeley

Introduction to Thick-Liquid-Wall Chambers*

ARIES MeetingApril 22-23, 2002

* This work was performed under the auspices of the U.S. Department of Energy by the University of California, Lawrence Livermore National Laboratory under Contract No. W-7405-Eng-48.

ARIES HIF Modeling - WRM 4/22/022

Thick-liquid-wall chambers: Key features and issues

HYLIFE-II

• Thick liquid “pocket” shields chamber structures from neutron damage and reduces activation

• Oscillating jets dynamically clear droplets near target

• No blanket replacement required, increases chamber availability

• Suited for indirect-drive targets

Key Issue: Chamber Clearing. Can the liquid pocket and beam port protection jets be made repetitively without interfering with beams? Will vapor condensation, droplet clearing and flow recovery occur fast enough to allow pulse rates of ~ 6 Hz?


Why Thick Liquids?

• Replace fusion materials questions with fluid mechanics questions

– These are questions that can be answered without a $1 billion test facility

• Maximize fusion power density

– Bring final focus/transport elements close to target

– Improve economics

UC Berkeley


Liquid-protection parameter space provides multiple options for target chambers

.

10510.50.1

DynamicClearing

Wetted Wall

Gravity Clearing

MultipleChambers

Thick Liquid

High Yield

No MaterialsTesting

Smaller DriverEnergy

Higher PowerDensity

4

6

10

12

Repetition Rate (Hz)

Mag

net

Sta

ndof

f (m

)

UC Berkeley

107

108

109

1010

1011

1012

1013

1014

1015

10-9 10-8 10-7 10-6 10-5 10-4 10-3 10-2 10-1 100 101

The use of thick-liquid protection reduces the first wallneutron flux as well as the average neutron energy.

Dry Wall

Thick-Liquid Protection

Neu

tro

n F

lux

(n/c

m2 -s

)

Neutron Energy (MeV)

Casen,tot

(n/cm2-s)

En,avg (MeV)

Dry Wall 2.7 1015 3.25

Thick-Liquid

3.9 1014 0.47

Approximately 58 cm of flibe is needed to protect the wall against neutron damage and ensure that it would meet Class C requirements.

10-1

100

101

45 50 55 60 65

WDRWDR goal

Was

te d

isp

osa

l rat

ing

Thickness (cm)

100

101

102

0 10 20 30 40 50 60

dpa goalFirst wall dose (dpa/year)

DP

A/f

ull-

po

wer

-yea

r

Thickness (cm)

55 cm of flibe reduces the first wall damage rate to <3.3 dpa/fpy (100 dpa in 30 fpy).

58 cm of flibe is required to reduce the SS304 first wall waste disposal rating to <1.


Top/Bottom Mid-Plane

Several potential liquid pocket geometries can be assembled from existing single-jet nozzles

High amplitudejet oscillation

Low amplitudejet oscillation

Porous liquid structuresuppresses shock

transmission (> 0.125 secshock transit time)

All porous jets mergeat pocket top

and bottom to fullyenclose target

and shield structures

Use of cylindrical jets for beamgrid allows flow control to

correct pointing errors

Large dimensionpocket opening:• reduces effects of liquid motion on venting,• provides directed debris jet to a separate condenser,• smoothness of oscillating jet surface now less important

Several variants of the HYLIFE-II pocket will be examined.

Asymmetric venting reducespocket symmetry and

debris jetting up beam lines

UC Berkeley


Driver/chamber interface

Credit: K. Springer & R. Holmes, LLNL

Key Issue: Self-consistent design. Can super-conducting final focusing magnet arrays be designed consistent with chamber and target solid angle limits for the required number of beams, standoff distance to the target, magnet dimensions and neutron shielding thickness?


Cut-away view shows beam and target injection paths


Work has progressed to detailed 3D neutronics models - predicting >30 year magnet lifetime

12

34

56 1

23

45

60.00E+00

5.00E+18

1.00E+19

1.50E+19

2.00E+19

2.50E+19

• There is a strong peaking of the fast neutron fluence at the center of the magnet array due to neutron scattering between neighboring penetrations.

• Estimated magnet life is 40-90 years depending on beam-to-structure clearance.

3D Tart model for HYLIFE-IIFast neutron flux for 36 magnet array


IFE system phenomena cluster into distinct time scales

• Nanosecond IFE Phenomena– Driver energy deposition and capsule drive (~30 ns)– Target x-ray/debris/neutron emission/deposition (~100 ns)

• Microsecond IFE Phenomena– X-ray ablation and impulse loading (~1 s)– Debris venting and impulse loading (~100 s)– Isochoric-heating pressure relaxation in liquid (~30 s)

• Millisecond IFE Phenomena– Liquid shock propagation and momentum redistribution (~50 ms)– Pocket regeneration and droplet clearing (~100 ms)– Debris condensation on droplet sprays (~100 ms)

• Quasi-steady IFE Phenomena– Structure response to startup heating (~1 to 104 s)– Chemistry-tritium control/target fabrication/safety (103-109 s)– Corrosion/erosion of chamber structures (108 sec)

Pri

nci

pal

foc

us

for

IFE

Tec

hn

olog

y R

&D

...

UC Berkeley


All IFE scientific topics can be identified and characterized by time scale and spatial location

Time Scale (Phenomena Duration)Spatial Volume Nanosecond

(Target Gain)Microsecond Millisecond

(Rep. Rate)Quasi-Steady

(Safety/Reliab.)

Capsule Neutron/ion/x-ray emission

—

Hohlraum (if used) X-ray and debrisemission

—

Driver energy transport paths Beam transportand focusing

Debrisaccumulation

Pocket Void/Ve nt Paths — —

External Condensing Region —

Target debrisexpansion/

interaction withablation debris,venting, impulse

Debriscondensation

—

Target-facing Surface Layers X-ray deposition Ablation/impulse

Blanket (liquid/solid) Neutron heatingrelaxation

Liq. hydraulics/solid thermal

mechanics

Activation, neutrondamage (solids),

safety

Final focus elements — — Damage rate

Chamber structures

Neutron andgamma

deposition

— —

Coolant recirc./heat recoveryloop

— — —

Safety, tritium,activation,corrosion

Accelerator/laser systems Driver physics — Driver rep. rate and reliability

Target injection — — Accel./heating —

Target fabrication — — — Safety/reliability

Balance of plant — — — Safety/reliability

UC Berkeley


Millisecond Chamber Phenomena

• Liquid pocket disruption and regeneration

– Pressure waves travel large distances over millisecond time scales, so liquid flow is incompressible

– Major liquid phenomena can be reproduced in scaled water experiments

• Ablation and target debris condensation

– Occurs on droplet sprays physically isolated from liquid pocket

– Condenser region baffling optimized for recovery and concentration of volatiles (He, DT, Hg, etc.)

– Experiments can used pulsed power to generate vapor/plasma from prototypical chamber materials (UCLA work)

UC Berkeley


The TSUNAMI code predicts microsecond venting phenomena

UC Berkeley


Porous Liquid Can Attenuate Shocks

• Impulse from x-ray ablation and pocket pressurization generates shock

• Simple “snowplow” model gives shock transit time as function of liquid void fraction :

• Shocks require > 100 ms to arrive at outside of porous pocket liquid

• Caveat: pocket openings may collimate high-velocity liquid

xxs

vs

xi

t = t1vl

t = 0

0

I

PjDj

t 1 xs

2

2I

Tim

e (s

ec)

0

0.2

0.4

0.6

0 0.5 1 1.5 2Shock Position (m)

I = 1000 Pa sec

- 50 cm line density

= 0.5

0.7

0.3

0.1

UC Berkeley


Navier-Stokes governs liquid hydraulics phenomena

Mass and momentum conservation

v 0 v

t v v

p 2v g

Free surface pressure boundary condition with impulse load I

p pv 1

r1

1

r2

where pv ,ave

IU

L

U

Lpv dt

0

L U

Nondimensionalize with appropriate scaling parameters:

v* v/ U * L p* p/ U 2 t* f t r* r / L pv*

pv L

IUGiving governing equations:

* v* 0 Stv*

t* v* *v* *p*

1

Re*2

v* 1

Fr

g

g

p* I*pv*

1

We

1

r1*

1

r2*

Major simplifications: No EOS, No energy equation No MHD

A scaled system behaves identicallyif initial conditions and St, Re, Fr,I*, and We are matched...

UC Berkeley


Single-jet experiments provide jet geometries for constructing integrated pockets

UCB Stationary Jets (1.6 cm x 8.0 cm,view from flat side, Re = 160,000, We = 29,000)

Bad:Breaks up Better: No Droplets

Stationary Oscillating

UC Berkeley


Recent experiments show that cylindrical jets can be sufficiently smooth for beam-line protection

honeycombRe = 100,000

screen/nozzle

cutter blade Re = 70,000 (no conditioning)

Re = 186,000

Jet with 1.5 : 1.0 nozzlecontraction ratio

Flow Conditioning

UC Berkeley


IFE thick liquid experiment scaling

Partial Pocket, HITF, and ETF scaling all preserve impulse effects

Single/Multiple Jet Integral Systems

Phase II Phase III DEMOHYLIFE Single Jet Partial

PocketHITF ETF HYLIFE

Geometric Scale 1 0.24 0.24 0.33 0.42 1Target Yield (MJ) — — — — 30 350

Volumetric Flow (m3/s) 0.84 0.01 0.16 4.76 8.58 75.00Oscillation Frequency (Hz) 6.0 27.1 12.2 9.1 9.5 6.0Nozzle Velocity U (m/s) 12.0 13.0 5.9 6.9 7.8 12.0Number of Jets 1 1 10 89 89 89Jet Dimensions D (cm) 7 1.68 1.68 various various variousJet Dimensions W (cm) 100 8.1 16 various various variousPumping Power (kW) 356 2 21 907 1,530 31,800

Storage Tank Size (m3) N/A 4 4 300 N/A N/AJet Reynolds Number Re 160,000 160,000 99,000 160,000 43,700 160,000Jet Weber Number We 103,000 21,000 7,900 15,200 18,100 103,000Froude Number Fr 7.3 19.4 7.3 7.3 7.3 7.3Working Fluid Flibe Water Water Water Flibe Flibe

Phase I

UC Berkeley


Computational tools have provided new insights

Computation plays thekey role in predicting impulse

loads to jets

CFD provides importantinsights for jet response

Droplet formation

UCLA

Droplet ejectionfrom cylindrical

jet surface

Tsunami simulation of vapor venting through jet

array

Code/experimentcomparison for

shock propagationover tube array

U. Wisc.UCB

Regions flattened by interaction with neighboring jet

Simulations from UCLA

Flow Direction

CollidingHYLIFEslab jets


The electro-thermal plasma source: a powerful and cost effective solution for pulsed vapor generation

• Based on existing knowledge from other experiments (NCSU)

• Capable of generating prototypical vapor density of flibe in a practical size chamber

• Discharge characteristics (fast rise time, short period) adequate to simulate IFE post-shot event

New plasma gun is being developed for liquid flibe high-T environment:

• ceramic insulator instead of plastic

• gun entirely inside the vacuum chamber

Technical issues:

• achieve unaided breakdown at 550 C flibe vapor pressure (0.2 mTorr)

• avoid chemical contamination from ablation of insulating materials and secondary discharges


A number of alternatives have been considered for thick liquid concepts

• We have evaluated flibe, flinabe, LiPb, Li and LiSn for pumping power requirements and TBR

• Calculated thickness of the liquid pocket is such that FW damage is limited to 100 dpa after 30 FPY operation

• Pumping power considers velocity head, friction loses and lift power

• LiPb and LiSn pumping power requirements are excessive

• Li has a large tritium inventory and poses fire hazards

• Only flibe and flinabe stand as reasonable options

Liquid Composition Thickness required (m)

Pumping power (MW)

TBR (1-D)

Flibe BeF2 (34%) LiF (66%) 0.56 48.46 1.25

Flinabe1 BeF2 (33.4%) LiF (33.3%) NaF (33.3%) 0.62 55.26 1.07

Flinabe2 BeF2 (37.5%) LiF (31.5%) NaF (31%) 0.62 63.23 1.07

LiPb Li (17%) Pb (83%) 1.03 681.76 1.61

Li Li (100%) 1.25 65.01 1.80

LiSn Li(50%) Sn (50%) 0.59 158.91 1.15


Some possible areas to for ARIES to study

• Design space of blanket thickness/wall radius/radiation damage limits for different first wall structural materials

– Possible higher damage limit thinner blanker or smaller chamber reduced pumping power and/or closer final focus magnets

• Alternate structural material (ferritic, SiC, C/C?) and compatibility with flibe and hohlraum materials (D-K Sze?)

• Mechanical design of oscillating nozzle and flow conditioning system

• Chamber/driver interface design issues/options


More slides on thick liquid wall chambers

Per F. PetersonDepartment of Nuclear EngineeringUniversity of California, Berkeley

April 17, 2002

Design Methods for Thick Liquid Protection of IFE Target Chambers

IFE Tutorial: http://www.nuc.berkeley.edu/thyd/icf/IFE.html

• Introduction: Chamber concepts, and the thick-liquid option

• Scaling review: Importance of time/spatial scales and phenomena coupling

• Bottom up: Understanding and modeling specific chamber phenomena and their coupling

– Nanosecond phenomena– Microsecond phenomena– Millisecond phenomena <--- Liquid hydraulics– Quasi-steady phenomena


IFE target chamber must meet four requirements

• Regenerate chamber conditions for target injection, driver beam propagation, and ignition at sufficiently high rates (i.e. 3 - 6 Hz)

• Protect chamber structures for several to many years or allow easy replacement of inexpensive modular components

• Extract fusion energy in high-temperature coolant, regenerate tritium

• Reduce radioactive waste generation, inventory, and possible release fractions low enough to meet no-public-evacuation standards.

Chamber will be 9-15% of total capital costDesign, not chamber cost, is the most important issue

UC Berkeley


Experiments can take advantage of recent scaling advances

• In IFE strong phenomena decoupling occurs in both time and space– Spatial decoupling boundaries

• small or unidirectional mass and energy fluxes• large time scale differences—slow side sees integral effect of fast

– Temporal decoupling boundaries• large time scale differences —slower phenomena sees integral effect of

fast– Inside these boundaries, phenomena

interactions must be considered• Phenomena change differently with reduced

geometric scale, time scale ratios for important coupled phenomena must be preserved to study interactions

S. Levy, 1999

Liquid pocket formation and hydraulic response can be studiedseparately from ablation, venting and condensation, using asimulant fluid (water) at reduced geometric scale.Reduces experiment cost by factor of ~50 to not use molten salt

UC Berkeley


Nanosecond phenomena control scientific viability




(Safety/Reliab.)


—


—


Debrisaccumulation





Debriscondensation

—




mechanics


safety


Chamber structures

Neutron andgamma

deposition

— —


— — —






Nanosecond phenomena:

• Target gain > Must be understood to judge the scientific viability of IFE

• Target output > Must be understood to predict chamber response

UC Berkeley


Millisecond phenomena control repetition rate




(Safety/Reliab.)


—


—


Debrisaccumulation





Debriscondensation

—




mechanics


safety


Chamber structures

Neutron andgamma

deposition

— —


— — —






Millisecond phenomena:

• Control the repetition rate > Must be understood to judge the engineering viability of IFE

• Initial conditions > Created by nanosecond and microsecond phenomena

UC Berkeley


Quasi-steady phenomena control safety and reliability




(Safety/Reliab.)


—


—


Debrisaccumulation





Debriscondensation

—




mechanics


safety


Chamber structures

Neutron andgamma

deposition

— —


— — —






Quasi-steady phenomena:

• Control safety > Must be understood to judge the engineering viability of IFE and of experimental facilities

• Control reliability > Must be understood to judge the attractiveness of IFE

UC Berkeley


Nanosecond Chamber Phenomena

• Driver energy transport– Shielding material standoff and gas density distribution– IRE will provide primary experimental test capability

• Target x-ray/debris/neutron emission– The most important questions are:

• partitioning of energy between x-rays, debris, and neutrons• effective x-ray black body temperature(s)• directional characteristics of x-rays/debris• control of emission by mass addition outside hohlraum

– High energy density/radiation dominates energy transport– Target design codes can model– Multidimensional effects likely important

in partition of energy between x-rays and debris kinetic energy

• Neutron shielding/energy deposition– 3-D codes (e.g. TART) can model

UC Berkeley


Microsecond Chamber Phenomena

• X-ray ablation, debris, venting, impulse loading (Chamber dynamics)– Experiments

• Z is currently the highest energy x-ray source available, has extensive DP diagnostics for x-ray ablation

• X-ray ablation - most important impulse source

• Reproduce 3-D gas dynamics/radiationtransport/reradiation/pocket pressurehistory

– Numerical modeling• Equations of state must

include vaporization, dissociation, ionization

• Radiation transport isimportant first 10’s ofmicroseconds

• Existing codes (2-D w/ TSUNAMI, 1-D w/ BUCKY)

Inserting wire array in Z

TSUNAMI results

UC Berkeley


Flibe x-ray ablation experiments on “Z” can be compared to simpler materials with known EOS’s (LiF, Li metal)

• LiF has been used as a non-toxic, well-characterized surrogate for flibe in recent experiments– Experiments at 41 J/cm2 match expected

wetted wall fluences• Koyo (laser chamber)• Osiris (heavy-ion chamber)

– Sesame EOS is available for LiF• Gives impulse prediction 10% less than ideal-gas

EOS– UCB/LANL predicted 2.8 m ablation

matches 3 - 4 m measured with LiF• Greater ablation, 4.2 m, is predicted for flibe;

will be confirmed in upcoming tests• Li metal has been tested at higher fluences

(~1000 J/cm2) under DP programs– Time-resolved diagnostics required due to

sample destruction– IFE samples can be treated with same

approach as DP effects testing work Cast and diced Flibe diskbeing handled in glovebox

14 mm

5 mm

LiF sample exposed to 41 J/cm2

shows clear ablation step

0.4 mm


Liquid jets can be optimized with single-jet experiments

• Design issues:

– Inlet plenum provides turbulent flow

– Flow calming section reduces core turbulent eddy energy and size

• perforated plates

• honeycomb

• screens

– Converging section

• further suppresses turbulence

• increases core flow mean velocity uniformity

• thins wall boundary layer

• contraction ratio sets jet packing density

– Cutter (optional) removes boundary layer

– Residual nonuniformity in exit velocity generates jet surface roughness

Velocity nonuniformity providesexcess kinetic energy, aftervelocity profile relaxes

u r 3 dA

0

A

U 3A

BoundaryLayerCutter

Screen(Typical)

Honey-comb

PerforatedPlate

rJ

rO

rN

ConvergingNozzleSection

GlobeValve

InletDistrib.Plenum

O-Ring

UC Berkeley


Honeycomb and screens can reduce core turbulence in flow calming section

• Honeycomb can greatly reduce transverse turbulence and secondary flow amplitudes

• Jets exit from each honeycomb cell

• Breakup of jet kinetic energy into isotropic turbulent kinetic energy occurs downstream

• A screen at the honeycomb exit can trip smaller instability modes and cause more rapid turbulence decay

0

4

8

2

6

10

u (m

/s)Average Longitudinal Velocity u

Downstream of a Cell Centerline

0

0.1

0.2

0.3

0.4

0 4 6 8 10 12

u' /

u

Distance From Honeycomb Exit (cm)

Fluctuating Velocity Downstreamof a Cell Centerline

Fluctuating Velocity WithExit Screen

UC Berkeley


Neutron shielding requires significant standoff of beam-line shielding nozzles

Side View End View

Jets closer to target require longer stand-off distance L, and largerjet L/D degrades jet smoothness

NozzleBlock

# L

NozzleBlock# R

r

rpi

rpo

zji

a

zjo

Bank "L" Bank "R"

OscillatingPocket Liquid

Crossed JetArrays

L1L

NozzleBlock

# L

NozzleBlock# R

z

UC Berkeley


Beam standoff sets liquid envelope for jet grid

• The volume available for the liquid jet envelope depends on the required standoff angle from the beams, Sn

• The fraction of the liquid envelope that can be filled with liquid depends on:

– Surface roughness • jet L/D• area contraction ratio• boundary layer trimming

– Pointing error– Velocity error

• gravity deflection• dilation

– Flow control to cylindrical jets can partially correct pointing errors

a

ny

Rn

a

nx

Liquid-JetEnvelope

j = 0

j = 1

j = Na

i = 0i = -Na

rEnj

Beam ShieldingStandoff Envelope

LiquidJet

Gravity Deflection

UC Berkeley


Cylindrical jets can be arrayed for beam-line shielding

• Staggered geometry reduces collimation of liquid droplets and slugs down beam lines

• Pitch to diameter ratio Pn/2rJn will be between 1.6 and 2.5

rJnj

rEnj

rNnj

n

n-1/2

n-1

Pn

Staggered Jet Array Cross Sections Nozzle Cross Section

Cutter DischargeCollection Tray

Cutter BladesAttach to Bottomof Nozzle Block

Beam StandoffEnvelope

UC Berkeley


UCB vortex test stand is now studying vortex injection and extraction methods for beam-tube protection

Side view showing operation at30° angle (extraction nozzle usedon right, vortex fan on left)

End viewsInjection

nozzle

UC Berkeley


Partial pocket experiments allow study of disruption

• 1/4-scale partial pocket multiple-jet experiments to study:

– Jet (various configurations) disruption by scaled propellant detonation

– Shock propagation and droplet/slug generation from multiple colliding jets

– Forced clearing of droplets confirmed by scattered light from laser-beam

SIDE VIEW TOP VIEWNozzle Block

Lexan SideWalls

Laser BeamConfirmsClearing

ScaledChemical

Detonation

PARTIAL POCKET SCHEMATIC

UC Berkeley


Chemical propellants can generate scaled impulse loads and disrupt thick liquid jets

• Numerical simulation allows comparison of scaled impulse for 1/4-scale jet disruption experiments

• Chemicals deliver impulse over longer time scale, but still rapid compared to > millisecond liquid response

.5m

r-axis

zz-axis

Chemical propellant jet IFE chamber

40sec 2sec

.125m

80sec 4sec4-cartridge firing deviceand impulse calibration

disk

UC Berkeley


Vacuum Hydraulics Experiment (VHEX) studies IFE jet disruption and regeneration

• Create hydrodynamically similar single jets and several jet arrays

• Transient flow into large vacuum vessel—water simulates flibe

t = 0 t = 0.8 ms(muzzle flash)

t = 1.6 ms(plume has hit)

t = 32 ms(peak deflection)

Impulse loadcalibration underway

UCB

UC Berkeley


Cartridges can provide required impulse loading

Single-jet disruption at 10.3 Hz

UC Berkeley


UCB disruption experiments are studying response of a 96-jet array to scaled impulse loading

- 6 msec

25 cm

2 msec 10 msec

“New” liquid interface Impulse-affectedregion- note divot

18 msec 26 msec

96-jet nozzleassembly in operation

Numerically-machined 96-jet nozzle

UC Berkeley


UCB has improved flibe vapor pressure predictions and identified a new salt composition allowing lower pressures

• Detailed activity coefficient data has allowed the vapor pressure of flibe to be accurately predicted at lower temperatures

• Ternary salt systems (“Flinabe,” LiF/NaF/BeF2) have been identified with very low melting temperatures (320°C)

– In beam tubes this low temperature molten salt creates a large reduction in the equilibrium vapor pressure (109/cc at 400°C)

Cooler

~350°C

RegenerativeHeat Exch.

Pump

ChamberFlibe

Pump VacuumDisengager

~600°C

Vortex Tube

.

1.00E+10

1.00E+11

1.00E+12

1.00E+13

1.00E+14

1.00E+15

460 500 540 580 620 660 700

temperature (C)

m o l e c u l e s / c ctheoretical modelORNL extrapolation

Recent flibe vapor pressureprediction

A degassing system may permitflinabe to be used for He/H2 pumping

UC Berkeley


Conclusions

• IFE has strong temporal and spatial phenomena decoupling– Pulsed complex systems: sequence from fast to slow phenomena

• Fast phenomena provide initial conditions for slower phenomena• First-principles modeling appears possible• Large temporal and spatial decoupling of subgroups of phenomena

simplifies experiments– Temporal decoupling: nanosecond/microsecond/millisecond/quasi-

steady– Spatial decoupling: driver/final focus/pocket/condensers/balance of

plant• Current status of liquid hydraulics research

– Single-jet nozzle designs are now available for constructing pockets• Reliability needs to be confirmed

– nozzle optimization studies to increase strength of nozzle components– single-jet molten salt experiments

• Liquid vortexes are still needed– Multiple jet interactions and pocket disruption/clearing now need study

UC Berkeley


A head recovery system was designed to minimize pumping power

Downward flow redirected by vanes to pressurize exit pipesRef. P. House

introduction to thick-liquid-wall chambers*

Documents

neutron damage

effects of liquid motion

thick liquids

wall chambers

magnet arrays

magnet dimensions

chamber clearing

beam port protection