An Overview of Collimator Design Progress

unkindnesskindUrban and Civil

Nov 15, 2013 (3 years and 6 months ago)

116 views

A
n Overview of
Collimator Design

Progress

Introduction

The focus of collimator design has shifted since the last design review. Following
MCNP investigations of neutron spectra at target and required shielding thickness
studies, the problem now appears sh
ielding dominated.
Reflection factors
1

obtained
through collimation concepts devised thus far do not warrant the use of a space intensive
reflector.
Observed g
ains in the high energy

spectrum

(capable of exciting
Oxygen

inelastic scatter:

E
>

6.5 MeV)

ar
e marginal a best, totaling 5
-
10% of the direct 14 MeV
contribution. Thermal reflection factors from 4 to 6 have been observed. However, the
utility of high energy neutrons to excite detectable characteristic gamma rays far
outweighs that of thermal neut
rons. Following the direction suggested by Dr. Dunn,
we
have begun to focus on minimizing the size of the system by combining the shield and
reflector at the expense of thermal reflection. Utilizing a very powerful neutron source
(10
13

n/s) we
hope to

be

able to ID the
standard
target
2

at 2m in around 20

seconds
by
analyzing the inelastic scatters from 14 MeV neutrons. In order to justify the selection of
shielding dominated concepts over expansive thermalizing reflectors, we have begun to
develop a
spec
tral effectiveness curve
.
The

purpose of this curve is to convert the
neutrons spectra observed at the target into a number of identifiable gammas emitted.

Using this curve we can evaluate reflector concepts which generate different spectral
distribution
s against one another
.



The first section of this

report substantiates the assertions made
in the introduction
through a review of the investigations we have performed.
A few broad concept
categories are extracted from the observations made during our M
CNP investigations as
well as from our concept generation sessions. The merits of these concepts are analyzed,
determining the

limits

on
target
flux enhancements which they create: how great of
reflection factors are attainable, and what is the spectral d
istribution.
Evaluation of the
increased thermal to high energy
flux ratio

attained using large reflectors
is made using a
basic

spectral effectiveness curve
.

Taking what we have learned by evaluating the
concepts with the spectral effectiveness, we conc
entrate on analyzing and enhancing two
of the most important concepts: the
Large RF Thermalizing Collimator
, and the
Absorbing Shielded Collimator
.


The second section of the report describes our analysis procedure.
This section was
created for our grou
p in order to focus our efforts on performing meaningful experiments
instead of performing many interesting but unnecessary investigations performed leading
up to design review 2.


In the final section,
a plan for upcoming investigations is outlined, hig
hlighting important
experiments and theoretical investigations to be performed
. We also pose questions that,
if answered, can help guide us to a solution tailored to our customer’s requirements.




1

Definitions for italicized terms are given in
Appendix B

2

See
ID Time

in
Appendix B


A Detailed Review of the Investigation
s

We Have Performed

T
he Concept Categories

Based upon our MCNP investigations we can identify two major

Collimation

Concepts.
The first is the
shielding dominated

design. In this design, little provision is made for
enhancing Reflection Factor (RF). Shield/Reflector materia
l contains integrated neutron
absorbers

like
boron, cadmium,

and lithium
. The second category is the
thermalizing
reflector
. This concept seeks to enhance the thermal neutron flux at target at the expense
of increasing size.
At this time, it seems imposs
ible to enhance the above 6.5 MeV flux
by more than a few percent.

Both concepts described above were developed as we
applied the observations that we made during our investigations of basic geometries
composed of various materials. These investigations
and observations are described in
the following section.


We began our collimation investigations by examining the at target flux generated by a
few basic geometries using MCNP. The studies were standardized so that the results
would be comparable. Our
standardization principle was fairly simple. A point isotropic
14MeV source was placed at the origin. An F5 (flux a
t a point) tally was placed at
x =
200 and x = 500cm
. Flux was binned into 100 equally spaced energy bins from 14 MeV
to 0 MeV. Following

each test a few important characteristics were measured and
tabulated for comparison of concepts. These characteristics include:


1.

The overall reflection factor

2.

Above 6.5 MeV flux reflection factor

3.

Thermal Reflection Factor

4.

Energy binned and total neutron

flux at target (gamma flux considered in
some tests)

5.

Energy binned and total neutron and gamma flux at outer surface of collimator


For a few tests, neutron flux off axis was also measured to determine how well the
collimator performed. The sophisticatio
n of these measurements will increase in coming
studies once we acquire MCNP Version 5.


The results of basic geometry tests are summarized in
Figure
1

below.



From these results we see that there is o
nly marginal gain in high energy flux. As
expected, low energy flux gains are significant

for some geometries and materials
. But
how useful are these lower energy neutrons for bomb detection? The short ans
wer is that
we do not know with any confidence

a
t this time. The importance of spectral distribution
is addressed briefly below in the section

titled

Spectral Effectiveness Curve
.


In accordance with the
design space analysis procedures
3
, our objective following these
prelimin
ary investigations was to determine the limitations on the useful phenomena
encountered. The useful phenomena
have been

decomposed
in terms of

quantitative
design output parameters




3

See
Appendix A
-

Structure of Our Investigation
: Design Space Analysis Procedure

Figure
1


The basic geometries and their characteristics.

SCC


Carbon


CH2


D2O



SphSh


Pb


Carbon

Planar


Carbon


CH2


Pb


D2O


Scone


CH2


Figu
re
2

Comparison of spectra from thermalizing and shielding dominated collimators. The plot
shows target spectra generated by lead/water collimator and 1m diameter hole
-
lattice carbon reflector.

Borated Paraffin with Pb

CH2 Hole Lattice Cylindrical
Collimator Cd ¼ in with
½ inch Pb
Biological Shield

With SCC 2m point target


D20 Hole Lattice Cylindrical with
Collimator Cd ¼ in with ½ inch Pb
Biological Shield

With SCC 2m point target


Materials Selection

The two concept categori
es require different types of materials. We have performed
investigations to determine which materials are good for each

(See
Error! Reference
source not found.
)
.

Hydrogenous Densities

Lithium Compounds


In order to enhance the
high energy reflection gains, it is necessary to find a material with
High Density, High Atomic Mass, Low
(inl +

n,2n)/elastic

cross section ratio.
T
he
material

could alternatively
have
inelastic

cutoff

energy greater than that of Oxygen 6.5
M
eV.



Integrated Shield and Reflector

Investigation into the integration of shield and reflector has been performed. By
combining reflector and shield, we seek to minimize overall system size
. It was
originally planned to
surround the s
ource with a reflector designed only to enhance flux
at target without regard to biological shielding. A biological shield would then be
constructed outside of the reflector. No absorbing material would be included in the
reflector.

Integrated shield/ref
lector concepts do not exclude neutron absorbing material from the
region near the source.

Figure
3

The search for ideal reflector material and comparison to real material characteristics.

Shielding Materials Comparison
Spherical Shield R=50cm
0.E+00
1.E-03
2.E-03
3.E-03
4.E-03
5.E-03
6.E-03
7.E-03
8.E-03
0
2
4
6
8
10
12
14
16
Energy (MeV)
Neutrons per source N
Pb and CH2 1:1 Volume
CH2
Pb and CH2 1:3 Volume



Table
1

i
n
Appendix A

lists design space variables

used in
our investigations


Repeated Structures

Repeated

structure concepts are being evaluated to determine if it is possible to obtain
very large (30 or more) reflection factors as a result of increased thermal neutron cross
sections and resulting short neutron mean free paths. The large thermal flux
enhance
ments might make a thermalizing concept more effective than the shielding
dominated concepts with only high energy spectra which we are currently focusing our
efforts on.

The repeated structures concept was evaluated as shown below



Figure
4

The lattice of holes reflec
tor: a
n example of
a
Repeated Structures

concept.

Spectral Effectiveness Curve
s

We are developing a spectral effectiveness curve to evaluate collimators which produce
differently weighted spectra


(
See
Figu
re
2
)
. The Spectral Effectiveness curve is quite
simplistic at thi
s time and may not adequately represent the relationship between target
flux and detectable gamma emission. Currently, the curve is created as follows.


Bomb materials considered are N, C, O, and H. These materials inelastically scatter
neutrons

(exce
pt H)
: the emission of characteristic gamma rays result
s
. The target
materials also create capture gamma rays (N, C, H, and O)
. However, capture cross
sections are very small for O and C

and quite small for N as well
. As a result, a smaller
number of ga
mma rays will be emitted for every thermal neutron that reaches the target,
compared to the number of gammas created by a 14 MeV neutron. So the time required
to ID a target with a given flux of thermal neutrons is longer than that required
for the

same f
lux of high energy neutrons. It is stated here qualitatively, but to truly evaluate the
effectiveness of a collimator design, we will need a quantitative evaluation.

Dr. Dunn,
Dr. McGregor, and Dr. Shultis may be able to give us some guidance on developi
ng this
curve
.

At this time two complimentary strate
gies have been tried to produce a spectral
effectiveness curve. The first is a mathematical treatment in which simulated target
spectra are convolved with the cross sections for the gamma yielding (inel
astic and
capture) interactions. The second treatment uses MCNP to expose simulated targets to
neutrons and mea
sures the emitted gamma spectra.


At this time, the second method is believed to be more accurate, since the first method
fails to account for m
ultiple interactions in the target. Method 1 is illustrated in
Figure
5

and the second in
Figure
6
.


The

basic spectral effectiveness curve has been composed by exposing
targe
ts of material
to 14
MeV
, 6.7MeV and thermal
neutrons and measuring the spectra and intensity of
gammas emitted. The spectra are combined (summed) and not weighted at this time.
Weighting factors may be determined in the future if this is
determined to be

important
.


Figure
5


Cross sections contributing to the spectral effectiveness curve. Only H,N,O,and C cross
sections are considered here.



Applying the preliminary spectral effectiveness curve to the spectra from the two best
concepts, we should see that only a marginal improvement
will be

realized

from the gains
in the low energy spectra, with the

exception of hydrogen. Hydrogen capture cross
section is quite large, and it has no inelastic scatter, so it cannot be detected except
through its capture gammas.

The preliminary

Figure
6


MCNP S
pectral Effectiveness
Curves. Data from MCNP 10 cm cube of material with
1g/cm3 target density, monodirectionally irradiated with 14 MeV and thermal (Maxwellian 0.025eV)
neutrons.


0.000
0.050
0.100
0.150
0.200
0.250
0.29
0.57
0.86
1.14
1.43
1.71
2.00
2.29
2.57
2.86
3.14
3.43
3.71
4.00
4.29
4.57
4.86
5.14
5.43
5.71
6.00
"C12 (14MeV)"
"H1 (thermal)"
0.000
0.010
0.020
0.030
0.040
0.050
0.060
0.67
2.00
3.33
4.67
6.00
7.33
8.67
10.00
11.33
12.67
14.00
"N14 (14 MeV)"
N14 (thermal)
0.000
0.010
0.020
0.030
0.040
0.050
0.060
0.48
0.95
1.43
1.90
2.38
2.86
3.33
3.81
4.29
4.76
5.24
5.71
6.19
6.67
7.14
7.62
8.10
8.57
9.05
9.52
10.00
O16 (14MeV)
O16 (thermal)
Gamma Noise at the Target

One potential problem in collimator design that
has not been addressed yet is the gamma
noise spectrum emitted from the collimator. Since the collimator will certainly contain
hydrogen and either oxygen or carbon, many inelastic and capture gammas will be
generat
ed within the collimator. These gammas
are the same gammas we intend to excite
in the target
in order
to make an ID. Some provision must be made to shield these
gammas from the detection
system
. Direct contributions can

easily be attenuated by lead;
h
owever, gammas emitted down the axis of th
e collimator will not encounter attenuating
material and will hit the target directly, backscattering to the detector. This undesired
signal may be quite large
and might even
mask the target generated spectrum. At this
time we proceed on the assumption t
hat scattered collimator born gammas can be
differentiated from target emitted gammas since they must necessarily have backscattered
off of electrons in the target in order to reach the detector, losing energy and shifting off
of the characteristic peaks.

If it becomes necessary to shield these gammas, this could
necessitate dramatic
reevaluation

of collimator design, so it is very important to address
this issue early in the next phases of design.

The Next Phase and Future Investigations

Direction

Revisi
ng the vision
---

a general course correction
. A good concept sketch might be
handy here.


Specific Known Deficiencies in Our Analysis


Revised Project Plan

We a
re

rewriting the project plan and reallocating team resources now that the neutron
source inv
estigation is nearing completion. Based on suggestions from the panel at DR 2
we will perform detailed investigations of ___ and ___ and ___. Some

justifications need
to be made.

Experiments Plan

This is an extension of the project plan intended to guide

and coordinate research efforts
of team members.
The Planned Experiments List is presented in
Appendix D
. A
representative experiment

from this list is presented in

Figure
7

below.


Experiments Plan

Ti
tle:


Lattice Reflector

Purpose:

Determine flux enhancements caused by lattice of holes. Demonstrate principle of differentiation.

Deliverable:


1.

Graph relating void fraction to flux ratio

2.

Effect of 1 hole that can be applied to collimators with non parall
el far separated holes

Directions:


1.

Run MCNP on test structures binning current (100 E bins) on both sides of the test structure



By explicitly stating the title, purpose, and directions behind major

experiments to be
performed, we will increase time management efficiency and distribute tasks more
effectively

amongst team members.

Figure
7

An example experiment from Planned Experiments List in
Appendix D

Questions for Our Director and Panel


Appendix A

Structure of Our Investigation
: Design Space Analysis
Procedure

In this
section we analyze the strategies behind our design procedures. To optimize the
time we invest, it is prudent to develop a vision of the path we will take to the
final
solution. This vision will be created by developing a
Revised Project Plan

and an
outl
ine
covering the
Investigations

to be
Performed
.

To develop these reports, it will be useful to
analyze the nature of our design problem from problem statement, through concept
generation, to final design. This progression is presented
below
in the conte
xt that we are
searching through a vast multi
-
dimensional array of possible designs for an optimal
solution.
We refer to this array a
s the
design spac
e
.




The nature of our design project is one of investigation


an open rather than a closed
ended que
stion.
Instead of being asked to
meet a specific s
et of desired output
parameters
,

we

have been asked to “…design the best

portable, on/
off, neutron
collimator
and source system which

maximize
s

neutron flux at a remote target ...”

In order to find
the be
st solution, we must consider a very broad region of the Design Space
. S
ince we
have a limited time to perform the analysis, we must identify the regions of the design
space in which the optimal solutions exist without performing exhaustive studies in all

regions. The procedure which solves this problem is classically referred to as Conc
ept
Generation and Evaluation.




… this needs more and I have more but not the time…
Design Space Variables
Wall Thickness
Material Type
Moderator Stoichiometry
Design Output Parameters
RF
Surface Dose
Gamma Flux at Target
Table
1

Design Space Variables and Output Parameters












Figure
8

Parameter Studies. Many parameter studies have been performed. These studies reveal
the sensitivity of output parameters to the design space variables they cover.

C
urves from Shielding and Planar
Reflector Studies

Appendix B

Glossa
ry and Definitions

Spectral Effectiveness Curve

Standard Target

Reflection Factor

Shielding Dominated Design

Thermalizing Reflector

High Energy Reflection Factor

Thermal Reflection Factor

Design Space

Design Input Variable

Design Output Parameter

ID Time

Appendix C

Theoretical Analysis of the Repeated Structures Concept

The theory of differentiation.

Statistical Analysis of Radiation Transport to show what minimum size is required to
obtain RF of x.

Is it really possible to do this?

Plotting of Particl
e Tracks

In order to get a feel for the transport of neutrons through the reflector, it is desired to
overlay a plot of the neutrons tracks form source to target. This should clarify where the
neutrons are coming from. Are they scattering deep into the c
ollimator, thermalizing, and
then exiting the repeated structures from deep within the reflector, or are they simply
diffusing toward the collimator’s face?… (A poor explanation here). I will add the
pictures and the diagrams to make this clear)

Thermaliz
ing Collimators

In order for thermalizing concepts to work, we must realize two things. First, we must
use a neutron efficient reflector material. Since there are necessarily more than 30
isotropic scatters required to obtain reflection factor of 30, the

ratio of elastic scatter to
capture must be much greater than
30:1.

The
second is

an advanced geometry like the
repeated structures described in that section below.

Appendix D

Experiments Plan

Title:


Lattice Reflector

Purpose:

Determine flux enhancem
ents caused by lattice of holes. Demonstrate
principle of differentiation.

Deliverable:


3.

Graph relating void fraction to flux ratio

4.

Effect of 1 hole that can be applied to collimators with non parallel far
separated holes

Directions:


2.

Run MCNP on test str
uctures binning current (100 E bins) on both
sides of the test structure



3.

Vary Materials: C, D2O, CH2, Pb

4.

Vary source energy



Title:


Integrated Reflector: Maximized Shielding

Purpose:

Determine the best shielding materials c
omposition

Idea:


Chunks of Pb can be mixed in with Paraffin or Poly to stop more neutrons.



We want to find optimum volumetric ratio.

Deliverable:


1.

Spherical 50 cm test comparisons with Lead in Paraffin: MCNP file
called p1pe3 in folder …mcnp test resu
lts
\
shielding
\
shielding
dominated
\

2.

Shielding Materials Comparison
Spherical Shield R=50cm
0.E+00
1.E-03
2.E-03
3.E-03
4.E-03
5.E-03
6.E-03
7.E-03
8.E-03
0
2
4
6
8
10
12
14
16
Energy (MeV)
Neutrons per source N
Pb and CH2 1:1 Volume
CH2
Pb and CH2 1:3 Volume


3.

From this plot it is apparent that Heterogeneous Mixtures of Pb and
CH2 can shield better than CH2 alone: The best volumetric ratio of
the two should be determined, by varying material card and density in
MCNP.

4.

For
the final result, the curves above must be integrated, and
convolved with an appropriate response function.

Directions:

Run MCNP varying densities. The procedure to determine the densities is
illustrated below.


For ½ CH2 and ½ Pb (Volumetric Ration 1:1
), mass densities are:


Lead


=

CH2


=

C


=

H


=

Total Density



5.65 + 0.47 = 6.12
For Volumetric Pb:CH2 = 1:3 we have

Pb


=

CH2


=

C


=


H


=

Total Density



2.825 + 0.705 =3.53

These are the densities o
f the elements in g/cm3



Here is the basic MCNP File for this experiment


And a sample output file with the Excel Sheet that I used to compare spectra



C:\Documents and
Settings\Owner.YOUR-D1E461F6E9\Desktop\HW 06 07\ME 574\MCNP Test Results\MCNPtests11-20\Shielding\Shielding Dominated\Shielding Materials Comparison.xls

Title:


Maximum Reflection Factor
Observed and Spectral Distribution

Purpose:

To determine if flux enhancements resulting from the use of a non
-
absorbing collimator justifies the use of the space required.

Deliverable:


1.

Target Spectrum for the highest RF observed in a collimator with
diame
ter less than or equal to 1 meter

2.

Target Spectrum for the highest above 6.5 MeV Flux

Directions:


1.

Run MCNP on test structures binning current (100 E bins) on both
sides of the test structure





Title:


High Energy Reflection Gains

Purpose:

Find the maxim
um above 6.5 MeV RF

Deliverable:


1.

Comparison of spectra at target for a few geometries and for a few
common (best we know of) reflector materials (Graphite, Pb, Fe, CH2,
D2O)

Directions:


1.

We don’t need the absolutely highest above 6.5 MeV gain. What we
n
eed is some demonstration of gain


or demonstration that it is not
feasible with simple geometry and materials. In the intro paragraph of
the report I have stated that these gains are marginal (5
-
10%). This
assertion is based only upon my limited invest
igations. It needs to be
substantiated.






Title:


SCC
and
Lattice CC Comparisons

Purpose:

To compare simple cylindrical collimator with hole lattice cylindrical
collimator

Deliverable:


1.

Comparison of spectra at target for a few geometries and for a f
ew
common (best we know of) reflector materials (Graphite, Pb, Fe, CH2,
D2O)
. Quantify the gains resulting from each material. How are they
different?

Directions:


1.

We don’t need the absolutely highest above 6.5 MeV gain. What we
need is some demonstrati
on of gain


or demonstration that it is not
feasible with simple geometry and materials.
The basic MCNP input
file needed for this investigation is included below:


2.