Download Cosmic Rays and Humans in Space

Cosmic Rays and
Humans in Space
Cary Zeitlin
Southwest Research Institute
Boulder, CO
100 Years of Cosmic Ray Physics
• Earliest experiments by Wulf, Hess circa
1912.
• High-energy physics originated in studies
of cosmic rays.
• Still a unique window for observing the
universe at the highest energy scales.
Applications of Cosmic Rays
• An important application of cosmic rays is
in biophysics: Prediction and monitoring
radiation doses received by astronauts in
space from GCR, SPE, SAA passes for
Low-Earth Orbit (LEO).
• Related subject: Radiation therapy for
cancer using similar beams of protons and
heavy ions at ~ few hundred MeV/amu.
Radiotherapy and Space Radiation
• Cancer therapy with heavy ions and space
radiation have much in common.
– Ions and energies of interest are similar.
– Intersection of nuclear physics and biology.
• A few places - GSI in Germany, NIRS in
Japan – have active research programs in
both areas.
Cancer Therapy with Charged Particles
• Charged particles give
•
better dose localization
than g or X-rays due to
Bragg curve.
Proton synchrotrons
with maximum
energies ~ 200-300
MeV are used in recent
designs.
Proton and Heavy Ion Therapy
• Proton therapy gaining acceptance: Loma Linda
University, Mass. General, MD Anderson, etc.
– More coming online in the next few years.
• Heavy ion therapy initially done at LBL in 1980’s
& 1990’s, dropped (no funding) , picked up in
Japan, Germany, and elsewhere in Europe.
– No facilities presently in US, nor are any planned.
• Efficacy of charged particle therapy vs.
conventional still has doubters but the physics
makes sense.
Heavy Ion Therapy Advantages
From Schardt et al.
• Even better dose localization – ions have higher
•
•
peak-to-entrance plateau ionization ratio.
LBL did only salvage cases  low success rate, but
GSI & HIMAC have shown clinical success.
US lagging – regulatory hurdles.
Heavy Ion Therapy Disadvantages
• Distal edge due to fragmentation slightly
reduces dose localization benefits.
• Machines are expensive.
– Costs coming down, may soon be competitive
with proton accelerators.
• Lack of FDA approval.
Heavy Ions in Space
• Same types of particles used to treat cancer
may also induce it, even at low doses.
• Long-duration spaceflight increases risk of
cancer, cataracts, maybe CNS damage.
– Earlier onset of cataracts seen in astronaut
cohort.
– No experience with long-duration flight outside
LEO.
Radiation Exposures in Space
• Complicated exposures – GCR gives a mix of
particles & energies.
– Huge unknowns in the biology.
– Some unknowns in the physics.
• Some of the exposure can be due to Solar
Particle Events (mostly protons).
– Modest shielding prevents acute effects, even in
intense events.
• In LEO, also get dose from passes through SAA
(trapped protons).
Exposure to GCR
• Dose mostly from protons and heavier ions, with
•
•
energy in the 10’s of MeV/nuc to ~ 10 GeV/nuc.
Health physics: Dose equivalent is the measure
of risk for cancer induction.
Conventional wisdom: 1 Sv  3% increase in
fatal cancer risk.
– Based largely on Japanese A-bomb survivors.
• Hard to estimate doses retrospectively.
– 3% is also the US legal maximum for job-related
health hazards.
Uncertainties in Risk Projections
10%
Maximum Acceptable Risk = 3%
▲
1%
ISS Mission
0.1%
.01%
Lunar
▲
▲
▲
Shuttle Mission
Increase in Individual’s Risk of Fatal Cancer
Mars
Mission
Uncertainties in Risk Projections
10%
Maximum Acceptable Risk = 3%
▲
1%
ISS Mission
0.1%
Lunar
▲
SPE??
“95% Confidence
Interval”
.01%
▲
▲
Shuttle Mission
Increase in Individual’s Risk of Fatal Cancer
Mars
Mission
Cosmic Ray Spectra
• GCR ~ 90% protons
• Most of the rest is He.
• Electrons and heavy
•
ions (Z > 2) are 1-2%
of the total flux.
But contributions to
dose go as Z2…
Abundance of Galactic Cosmic Rays by
Species in Unshielded Deep Space
1.0E+00
Flux
Badhwar- O’Neill Model
Relative Contribution
1.0E-01
1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1
3
5
7
9
11
13
15
17
Nuclear Charge
19
21
23
25
27
Abundance weighted by <LET>
1.0E+00
Flux
Dose
Relative Contribution
1.0E-01
1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1
3
5
7
9
11
13
15
17
Nuclear Charge
19
21
23
25
27
Abundance weighted by
average dose equivalent
1.0E+00
1.0E-01
Relative Contribution
Fe is
the
tallest
pole!
Flux
Dose
Dose equivalent
1.0E-02
1.0E-03
1.0E-04
1.0E-05
1.0E-06
1
3
5
7
9
11
13
15
17
Nuclear Charge
19
21
23
25
27
What is Dose Equivalent?
• Unit of dose is Gy = 1 J/kg. Purely physics.
• Unit of dose equivalent is Sv, same units, but
•
with a weighting factor that varies by radiation
type to approximate biological effectiveness.
For exposure to a single radiation type, the
specific radiation weighting factor is used, wR.
– wR = 1 for g, e-, m; 5 for protons; 20 for a and
heavier nuclei.
• For a mixed field, one integrates the “LET”
spectrum (LET = dE/dx in water) vs. a “quality
factor” which depends only on LET, Q(L).
Mixed-field Quality Factor Q(L)
• Traditional view: effects
increase to ~ 100 keV/mm,
and then decline due to
“overkill.”
– Can’t kill a cell more than once.
(Dead cells aren’t really the
problem anyway.)
• Huge scatter in the RBE data.
– Lots of questions: what is (are)
the right endpoint(s)?
– Can we extrapolate cells or small
animals to humans?
Figure courtesy of F. Cucinotta, JSC
Nuclear Fragmentation
• N-N collisions  fragmentation of heavy
ions into lighter ions.
• “Projectile fragments” ~ preserve b and
direction of incoming ion.
• “Target fragments” are MeV’ish
– Charged particles are short-ranged.
– Neutrons with high wR are produced.
• For total dose, fragmentation helps:
– dE/dx ~ Si Zi2 and sum of squares < the
square of the sum.
– Effect on dose equivalent not so easy to
characterize.
Transport Calculations
• Analytic and Monte Carlo approaches.
• NASA codes date to 1960’s, when MC was
impractical due to CPU limitations.
• Analytic solves the Boltzmann equation
with many approximations and less
realism, e.g., problem is reduced to 1-d.
• Either way, need knowledge of many
nuclear fragmentation cross sections.
Cross Sections and Transport
J. Wilson et al., NASA TP, 1995
Z. Lin, Phys Rev C 75, 2007
Fragment
production
cross sections
Chargechanging
cross sections
Nuclear Cross Section Database
• Or, “What can a 3-person group do with a
tabletop experiment using 80’s technology?”
• Turns out there is (or was) a niche: Most
cross sections needed for transport models
not measured (& never will be).
– # of exclusive final states is ~ intractable.
– Inclusive reactions to be studied are of the form
(Zp, Ap) + (Zt, At)  (Zf, Af) + X
• Small subset of possible reactions was used
to “validate” models (30-50% errors o.k.).
Experimental Program
• Measure as many cross sections as possible.
• Silicon detectors upstream & downstream of target.
• Main issues are dynamic range of readout electronics –
•
need 3 orders of magnitude – and availability of beams.
Unlike earlier experiments, use small-acceptance
detectors to measure light fragments.
Large & Small Acceptance Spectra
• Upper histogram is typical
•
•
for this type of experiment –
resolve fragments down to
Zfrag ~ Zbeam/2.
Previously, cross sections
only reported for these
“peripheral” interactions.
Rough analogy to deep
inelastic scattering: if we
only test Zfrag  Zbeam/2,
we’re not going very “deep.”
Experimental Database
Ion
56Fe
Energy (MeV/nuc)
400
500
600
800
48Ti
40Ar
1000 3000 5000 10000
1000
290
400
35Cl
28Si
• Accumulated over 12
650
290
24Mg
400
•
650
600
1000
800
1200 3000 5000 10000
400
20Ne
290
400
600
16O
290
400
600
14N
290
400
12C
290
400
11B
290
400
10B
290
400
4He
230
•
1000
3000 5000 10000
years of ~ 1 week/yr
of beam time.
For each beam ion/
energy, full target set
was C, CH2, Al, Cu,
Sn, and Pb.
~ 200 chargechanging cross
sections, ~ 2000
fragment production
cross sections
measured.
Fragmentation Data vs. Models
• Simple model of overlapping
•
•
spheres describes chargechanging s’s to ~ 5-10%.
Older parametric models predict
monotonically decreasing
fragment production cross
sections as a function of DZ.
Missing physics:
– Odd/even effect
– Suppression of Z=9
– Increased cross sections for CNO.
• PHITS & FLUKA use different
flavors of QMD and do much
better than older models, but
still off.
Shielding Effectiveness: Another Way
to Look at the Same Data
• 1 GeV/amu 56Fe beam with many
materials, including spacecraft
shielding candidates.
• Nobody has a metric of shielding
performance, so we made one up
based on physics (not biology).
– dDn = change in avg. dose per
particle behind the target.
• Take out depth effects, get
unsurprising result: hydrogen best.
– Fragmentation drives dose reduction.
– Mass ~ A, s ~ A2/3, so as A increases
the # of interactions per unit target
mass goes down.
• Predicted by Wilson et al. for GCR,
nice to validate in lab.
Beam Dependence of Shielding Results
• Wilson calculated for GCR.
• Are the experimental results
beam-dependent?
– Can a single beam/energy
serve as proxy for the GCR?
• Looked at dose reduction
•
with the same CH2 target in
several different beams.
It seems any energetic heavy
ion beam of Z  14 and E 
600 MeV/amu or greater
yields ~ similar results.
TEPC Characterization
• Experiments with T. Borak (CSU)
•
to characterize response of
tissue-equivalent proportional
counters to heavy ions.
Simple device with complicated
response function.
– Widely used in health physics.
– Used by NASA on ISS, Shuttle.
– Response not systematically
studied.
• TEPCs measure “lineal” energy
transfer, typically equated to LET
(but not rigorously correct).
• Avg. energy deposited vs.
impact parameter looks
strange…
Tissue Equivalent Proportional Counters
• For these experiments, used spherical
TEPC with field-shaping helical wire
surrounding anode.
• Tissue equivalent wall - wall effects crucial.
• Fill with TE gas @ ~ 100 millitorr to
simulate a 1 mm diameter tissue volume.
• Helical wire causes artifacts, too
– Flight TEPCs are cylindrical, no wire.
• Result shown on previous slide (e vs. b)
greeted with scorn by “experts” at first.
Modeling Explains Measurement
• Proposed mechanism for
•
•
GEANT4 simulations by Taddei, Zhang, and Borak
“boomers” – ionization
cloud formed in wall near
cavity boundary leaks in.
First good model
calculation by Nikjoo
agreed with the data.
Later modeling by Taddei
et al. using GEANT4
supports conclusion.
Flight Measurements
• Impossible to measure all particles of interest –
•
many trades to make a flight instrument.
Typical allocation for a NASA ESMD instrument:
few kg of mass, a few W of power, a few $M.
– Hitch a ride on an interplanetary mission (MARIE on
Mars Odyssey, RAD for MSL, CRaTER on LRO), or to
ISS.
• NASA JSC has operational responsibility for crew
dose monitoring – they rely on TEPC, passive
dosimeters, and some data from international
partners.
– JSC particle spectrometers flown to date on ISS,
Shuttle had lots of problems.
MARIE
• Martian Radiation Environment Experiment flew
on 2001 Mars Odyssey spacecraft.
– Original plan: Odyssey was orbiter + lander. Idea was
to have an instrument on both.
– After MPL crashed, NASA got cold feet & Odyssey
became an orbiter only. (Lander became Phoenix.)
• Straightforward silicon telescope built by NASAJSC, similar to ISS instruments.
– Worked for 20 months, failed in Halloween ’03 SPE.
– Many hardware & firmware problems.
CRaTER (Cosmic Ray Telescope for
the Effects of Radiation)
• Built by Spence et al.,
•
•
Boston University.
Si telescope with
alternating pairs of
detectors and pieces of
tissue-equivalent plastic.
Will fly on Lunar
Reconnaissance Orbiter
(LRO), launch scheduled
for May.
Mars Science Laboratory (MSL)
• MSL is the largest Mars
rover to date
– 850 kg, 10 instruments
• Launch date – Fall 2011
•
•
(was to be Fall ’09).
Arrives at Mars between
July & September 2012.
Prime mission duration
1 Mars year (687 days)
RAD for MSL
A SolidStateDetector (SSD) A
D CesiumIodide(CsI)
B SSDB
E NeutronChannel (Bicron430Mscintillatingplastic)
C SSDC
F Anti-coincidenceShield
• Kiel University built sensor head, SwRI built electronics.
• 1.5 kg, 4 W
• ~ 104 dynamic range in DE for charged particles: Si
telescope (A+B+C) & CsI(Tl) (D).
n
• CsI stops protons up to ~ 100 MeV (good for SEPs).
• Plastic scintillator (E) for neutrons with energy > 10 MeV.
– Large neutron flux from RTG at lower energy.
• Plastic scintillator anticoincidence shield (F).
MSL RAD Scintillators
• Scintillators read out
CsI (D)
by photodiodes.
D Readout
Diode
CSA
– Keeps mass down.
– Only one bias voltage
(-70V) needed, no risk
of coronal discharge.
– Anticoincidence also
read out by photodiodes.
Retainer
Neutron Channel (E)
E Readout Diode
RAD Electronics
• Preamplifiers & 1st-stage shaping amps built into
•
•
•
sensor head.
Analog outputs  electronics box, into VIRENA
(36-channel custom ASIC).
VIRENA multiplexes outputs for the 14-bit ADC.
Digitized data processed in Level 3 by a virtual
8051 instantiated in the RDE FPGA.
– Most events histogrammed, a few high-priority (high
LET) events will be telemetered down.
– Only sending down ~ 400 kB/day.
VIRENA ASIC
• Built by NOVA R&D.
• Fast & slow discriminators, additional shaping stage,
sample & hold, output multiplexer.
• Highly configurable.
• Controlled by the “RAE” FPGA which runs Level 1
software for trigger & readout control.
• Same FPGA runs Level 2 software (uses digitzed data)
for low-level data processing – applies calibration
constants, chooses best gain readout, etc.
Beamline Setup for Calibration
target
Beam incident
from left.
Some beam ions survive the target,
others fragment into lighter ions. (Just
like fragmentation experiments.)
Need to calibrate quenching, esp.
in E, for heavy charged particles.
Species Resolution in Si Detectors
Scale PH to ion charge
RAD for ISS
• Charged particle capability ~ same as MSL RAD.
– Slightly enhanced – CsI replaced by larger BGO for
more stopping & to eliminate possible issue with
moisture. (ISS can get humid.)
• Greatly enhanced neutron capability.
• Neutrons are created by nuclear interactions.
– Important behind large depths of shielding.
– Estimated to contribute 30-50% of dose equivalent in
ISS.
– ISS shielding highly variable, averages 20 g cm-2.
Neutron Coverage
• MSL RAD not designed
•
for lower energies
because of RTG.
ISS RAD will be used
for crew dosimetry,
must measure 0.5-10
MeV neutrons.
Neutron Coverage
• MSL RAD not designed
•
for lower energies
because of RTG.
ISS RAD will be used
for crew dosimetry,
must measure 0.5-10
MeV neutrons.
MSL
RAD
Neutron Coverage
• MSL RAD not designed
•
for lower energies
because of RTG.
ISS RAD will be used
for crew dosimetry,
must measure 0.5-10
MeV neutrons.
ISS
RAD
MSL
RAD
Adding Neutron Spectrometry to RAD
• Plan: Use boron-loaded plastic scintillator (Bicron
•
•
•
BC-454 or Eljen EJ-254) and capture-gating.
Unique signature of capture events: first pulse
from interaction(s) with protons, neutron is
thermalized, captured by 10B.
Amplitude of 1st pulse  neutron energy  no
unfolding.
Second pulse is due to 10B (n,a)7Li* reaction.
– About 2.3 MeV released, but light output is heavily
quenched  60 keVee.
– Dt distribution is exponential, <t> ~ 2 msec for usual
B concentration.
B-Loaded Plastics in Flight Experiments
• Double-pulse method used by instruments
built for planetary science:
– Mars Odyssey Neutron Spectrometer (LANL).
– Mercury Messenger Gamma-Ray and Neutron
Spectrometer (JHU APL).
• Both use PMT’s for readout.
• GRNS uses scanning ADC with FPGA doing
all the coincidence logic.
Capture Gating
• Can’t see 2nd pulse above noise with pin
diodes – noise is 1 MeVee FWHM at best.
• Investigating APD’s + optimized preamp
design.
– Present preamp excellent for low-capacitance
detectors; APD’s are high-capacitance.
– Have to add ~ 400V HVPS.
– Fallbacks are SiPM’s and (worst case) PMT
since constraints are greatly relaxed for ISS.
Conclusions
• Space radiation health presents many
challenges – small but important role for
physics.
Conclusions
• Space radiation health presents many
challenges – small but important role for
physics.
• Severely resource constrained, but that’s
part of what makes it a challenge.
Conclusions
• Space radiation health presents many
challenges – small but important role for
physics.
• Severely resource constrained, but that’s
part of the challenge.
• Thanks for your attention!