OBSERVATION OF DOPPLER BROADENING IN β-DELAYED PROTON-γ DECAY

By

Sarah Schwartz

A THESIS

Submitted to

Michigan State University

in partial fulfillment of the requirements

for the degree of

Master of Science in Physics

2016
 ABSTRACT

OBSERVATION OF DOPPLER BROADENING IN β-DELAYED PROTON-γ

DECAY

By

Sarah Schwartz

The Doppler broadening of γ-ray peaks due to nuclear recoil from β-delayed nucleon

emission can be used to measure the energies of the nucleons. The purpose of this Thesis is

to test and apply this Doppler broadening method using γ-ray peaks from the 26 P(βpγ)25 Al

decay sequence. A fast beam of 26 P was implanted into a planar Ge detector, which was used

as a 26 P β-decay trigger. The SeGA array of high-purity Ge detectors was used to detect

γ rays from the 26 P(βpγ)25 Al decay sequence. Radiative Doppler broadening in β-delayed

proton-γ decay was observed for the first time. The Doppler broadening analysis method was

verified using the 1613 keV γ-ray line for which the proton energies were previously known.

The 1776 keV γ ray de-exciting the 2720 keV 25 Al level was observed in 26 P(βpγ)25 Al decay

for the first time and used to determine that the center-of-mass energy of the proton emission

feeding the 2720-keV level is 5.1 ± 1.0 (stat.) ± 0.6 (syst.) MeV, corresponding to a 26 Si

excitation energy of 13.3 ± 1.0 (stat.) ± 0.7 (syst.) MeV for the proton-emitting level. The

Doppler broadening method has been demonstrated to provide practical measurements of

the energies for β-delayed nucleon emissions populating excited states of nuclear recoils at

least as heavy as A = 25.

 TABLE OF CONTENTS

Chapter 1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.1 Introduction to the nucleus . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Types of nuclear decay . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
1.3 Beta-Delayed Nucleon Emission . . . . . . . . . . . . . . . . . . . . . . . . . 2
1.4 Previous work in Doppler broadening of β delayed nucleon-γ decay . . . . . 5
1.5 Example of Doppler broadening analysis . . . . . . . . . . . . . . . . . . . . 5

Chapter 2 Experiment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.1 Lab Layout . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
2.2 Detectors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

Chapter 3 Data Analysis and Discussion . . . . . . . . . . . . . . . . . . . . 13

3.1 Gamma-ray Intensities . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.2 βp Feeding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
3.3 Study of Doppler Broadened Peaks . . . . . . . . . . . . . . . . . . . . . . . 17
3.4 1613 keV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
3.5 1776 keV . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

Chapter 4 Conclusion and outlook . . . . . . . . . . . . . . . . . . . . . . . . 29

iii
Chapter 1

Introduction

1.1 Introduction to the nucleus

The atom is a building block of all matter. It is neutral and composed mostly of empty space.

Essentially all of its mass concentrated in the positively charged nucleus while the remainder

of its mass is in the negatively charged electrons which surround the nucleus. The nucleus

is composed of two different particles, the proton and the neutron. The proton is a fermion

consisting of two up quarks and a down quark, giving it one positive elementary charge.

The neutron is a fermion consisting of two down quarks and an up quark which results in a

neutral charge. Both of these nucleons are about the same mass. It is basic knowledge that,

due to the electromagnetic force, like charges repel while opposite charges attract. Even

though a nucleus consists of only positive and neutral particles they are bound due to the

fundamental force known as the strong nuclear force allowing them to be long-lived.

1.2 Types of nuclear decay

Nuclei tend to seek more energetically stable configurations by undergoing several types of

nuclear decays. A few of the different types of decays include α, β and γ decay. Alpha

decay refers to the spontaneous emission of an α particle. Alpha decay is mediated by a

combination of the attractive strong force and the repulsive electromagnetic force. There

1
are similar decays in which only one nucleon (a neutron or a proton) is emitted. These are

neutron and proton emission, named after the nucleon that was emitted, often from excited

states. The next fundamental decay is β decay, which is the emission of a β particle mediated

by the weak nuclear force. There are two types of β decays, β − and β + . In β − decay a

neutron decays into a proton, β − particle (electron) and electron anti-neutrino. In β + decay

a proton decays into a neutron, β + particle (positron) and electron neutrino. Finally, γ

decay is an electromagnetic transition of a nucleus in an excited state to a lower lying state

with the same number of protons and neutrons. A γ ray, or high energy photon, is emitted

during this transition.

These decays sometimes occur in a sequence, as in the case of beta-delayed proton-gamma

decay. First a parent nucleus beta decays. Then an excited state of the daughter nucleus

emits a proton. Finally the granddaughter nucleus de-excites via gamma decay. The decay
∗A → W ∗A−1 → W A−1 where A is the atomic number, Z is
occurs as follows XZA → YZ−1 Z−2 Z−2

the proton number, W , X, and Y are the nucleus names, and the asterisk denotes excited

states.

1.3 Beta-Delayed Nucleon Emission

When a nucleon is emitted from a nucleus the daughter nucleus recoils. For a given decay, if

the Center of Mass (CM) energy, ECM , is known then the velocity of the recoiling daughter

nucleus can be calculated by employing conservation of energy and momentum.

1 1
mn vn2 + md vd2 = ECM (1.1)
2 2

mn vn = −md vd (1.2)

2
Figure 1.1: An example of β delayed proton emission is 26 P(βpγ)25 Al. Above is a section of
the chart of nuclides [6] where the number of protons, Z, is on the Y-axis and the number
of neutrons, N , is on the X-axis. The blue arrow represents the β + decay from 26 P to 26 Si
which is then followed by the green arrow which represents the emission of a proton from
26 Si to 25 Al.

Figure 1.2: β-delayed proton-γ decay is a sequence of decays in which a parent nucleus β-
decays to become the system shown above. Next the proton is emitted causing the nucleus
above to recoil. Lastly a γ ray is emitted in an arbitrary direction with respect to the
surrounding γ-ray detectors, which results in a broadened peak in the γ-ray spectrum that
is centered on the unshifted energy. Figure courtesy of Erin O’Donnell

3
 where mn and md are the mass of the nucleon and daughter nucleus, respectively, and

likewise vn and vd are the velocities of the nucleon and daughter nucleus, respectively. Sim-

plifying, the velocity of the daughter nucleus can be written as

mn
vd = vn (1.3)
md

The daughter nucleus can be left in an excited state by the decay and can then de-excite

via γ decay. If the initial velocity is high enough and the life time of the excited state is

short enough then the γ ray will be emitted while the nucleus is still recoiling resulting in

the Doppler shift of the γ-ray-energy in the laboratory frame. When many of these decays

occur the resulting line in the γ ray spectrum will be Doppler broadened due to the isotropic

emission of the nucleons and of the γ rays from the daughter nucleus.

This also allows us to find the CM energy for nucleon emission decays that are unknown.

By modeling the broadening of the peak the velocity of the recoiling nucleus can be calculated

and ultimately the CM energy.

However in many cases the daughter nucleus is not in a vacuum and the medium it is in

has some stopping power that reduces the velocity. The velocity found via the broadening is

therefore lower than the initial velocity right after the decay. Naively assuming that the initial

and observed velocities of the daughter nucleus are the same results in the underestimation

of the CM energy.

This can be remedied by using Bethe’s stopping power equation which is incorporated

in software like Stopping Power and Range of Ions in Material (SRIM) [20]. The stopping

power of a material is dependent on the energy of the ion in the material. Since the ion is

constantly losing energy in the material the stopping power is also constantly changing for

4
the nucleus traversing it.

1.4 Previous work in Doppler broadening of β delayed

nucleon-γ decay

Doppler broadening of this kind has been observed from only one decay channel: β delayed

nucleon emission from the 11 Li(βnγ)10 Be decay channel [3, 10, 11, 16, 14]. The broadening

of these peaks was relatively clear since the recoiling nucleus, 10 Be, is a light nucleus allowing

the CM energy from neutron emission to give 10 Be a relatively large recoil velocity. There

were four broadened peaks observed in coincidence with neutrons from this decay at 2590,

2895, 3368 and 6263 keV. These peaks were clearly broadened in relation to other peaks in

the γ ray spectrum as can be seen in Figure 1.3. These peaks were observed to be broadened

in γ-γ coincidence spectra as well.

Further detail of the analysis of the 11 Li(βnγ)10 Be decay and for Doppler broadening

from β delayed nucleon emission in general were given in a paper by Fynbo in 2003 [10]. This

paper described the reason for the appearance of the line shapes in relation to the half-life

of the excited state and how information can be extracted from these broad peaks.

1.5 Example of Doppler broadening analysis

The ideal case for studying the Doppler broadening of a γ-ray peak from β-delayed nucleon-

γ decay is when a single state in the parent nucleus emits a nucleon to feed a state in the

daughter nucleus. Ideally there are also no higher lying states in the daughter nucleus feeding

the state of interest via γ decay. Furthermore, the decay occurs in vacuum and therefore it

5
Figure 1.3: Spectrum from 11 Li β decay which includes broadened peaks from 11 Li(βpγ)10 Be
from Sarazin etal [16].

is not necessary to incorporate the half-life of the decaying state. Then one can convolute

the detector response function with the underlying physical broadening by constructing a

function of identical detector response functions spaced evenly and centered at the non-

shifted energy in order to model the peak shape. This means the function used to describe

the broadened peaks is the detector’s response function spread out with a boxcar function.

To find how broad the peak is, the distance from the center of the highest (or lowest)

energy response function to the center of the central response function can be determined.

Then both of these energies would be used in a Doppler shift equation for energy where the

central energy of the center peak would be called E0 , or the unshifted energy, and the center

energy of the maximum peak is E 0 , or the shifted energy.

E 0c
E0 = (1.4)
c + vn

where c is the speed of light and vn is the velocity of the nucleus at the time of γ decay.

6
Now the velocity of the nucleus at the time of γ decay can be found. If the recoil and γ decay

of the nucleus had occurred in free space conservation laws could be applied at this point

to find the kinetic energy of the nucleon at the time of its emission and finally add it to the

kinetic energy of the recoiling nucleus to find the CM energy of the decay. Unfortunately

the situation is usually not so simple.

When the recoil occurs in a material, both the stopping power of the material and the

half life of the excited state of interest must be accounted for. Since stopping power is energy

dependent it is more accurate to iteratively determine the amount of energy lost in a fraction

of the half-life and then re-evaluate the stopping power again at this lower energy.

After computing how much energy was lost by the recoiling nucleus in the material, the

original kinetic energy of the nucleus can be found. Finally using conservation laws the

kinetic energy of the proton can be calculated and also the CM energy of the decay.

7
Chapter 2

Experiment

2.1 Lab Layout

The data used in this thesis was obtained during a 26 P β-decay experiment (e10034) which

was carried out at the National Superconducting Cyclotron Laboratory (NSCL) at Michigan

State University. The Coupled Cyclotron Facility at NSCL is shown in Figure 2.1 [2].

A primary beam of stable 36 Ar was accelerated to about 150 MeV/u by the coupled

cyclotrons (K500 and K1200). Then to produce the isotope of interest, 26 P, the accelerated

beam was impinged upon a production target of 9 Be of thickness 1.55 g/cm2 . The processes

of nuclear fragmentation created not only 26 P, but many contaminants as well. To try to

purify the beam it next goes through the A1900, a series of tuned magnets which filters

out contaminants by magnetic rigidity [15]. This works since each isotope has a different

charge/mass ratio and therefore when moving perpendicular to a magnetic field each isotope

will start to curve in arcs of different radii. The magnets, and therefore the magnetic field,

can be tuned such that the isotope of interest stays in the beam line and others are stopped

by physical barriers. This filtering is not perfect since there are isotopes with very similar

magnetic rigidities. These isotopes end up being contaminants for the experiment. After

the beam passes through the magnets it is further purified by time of flight using a radio-

frequency fragment separator [1].

This results in a final beam that is 74% pure with major components being 24 Al and

8
Figure 2.1: The layout of the Coupled Cyclotron Facility at the National Superconducting
Laboratory. The beam starts as stable ions that are accelerated by the coupled cyclotrons,
the K500 and K1200. Once the stable beam is accelerated it is impinged upon a production
target, 9 Be in our case, which creates the isotope of interest (26 P) along with many other
isotopes. This new beam is delivered to the A1900 which helps to purifying the beam using
magnets. Next the beam goes to the RFFS which further purifies the beam. Finally the rare
isotope beam of interest is sent to the experimental setup.

22 Na. The particle identification plot in Figure 2.2 shows the energy loss of beam particles

versus time of flight. Each concentration of points indicates a different isotope; the points

circled in red are the 26 P particles of interest.

2.2 Detectors

At the end of the beam line, the 26 P beam was directed into the S2 vault to the experi-

mental setup. The setup consisted of two detectors: a central detector where the beam was

implanted, a Germanium Double Sided Strip Detector (GeDSSD), and a surrounding array

of 16 high purity germanium (HPGe) detectors, the Segmented Germanium Array (SeGA).

They can be seen in Figure 2.3 (A) and (B) respectively.

The GeDSSD [13] was the implantation system, which is a 1-cm thick planar germanium

detector. It is divided by electrodes into 16 segments that are 5 mm wide on the front and

9
Figure 2.2: The plot shows the energy loss of a particle versus the time of flight (TOF). Each
circled concentration of points is a different isotope. The red oval encompasses 26 P particles
while the green and yellow encompass the 24 Al and 22 Na particles, respectively.

Figure 2.3: (A)The Germanium Double Sided Strip Detector (GeDSSD) with the cryostat
open. The secondary beam was implanted into the GeDSSD which detected energy released
when a decay occurred. It was surrounded by SeGA in the experiment. (B) A computer
aided design (CAD) of the Segmented Germanium Array, SeGA. It is composed of 16 HPGe
crystals surrounding the GeDSSD, where the beam was implanted.

10
Figure 2.4: This conceptual shows the GeDSSD surrounded by the SeGA Detectors. Figure
courtesy of David Pérez-Loureiro

16 similar strips that are orthogonal on the back, providing 256 pixels. The final beam is

implanted into the GeDSSD about 300 µm into the germanium crystal. This is where all

the decays of interest would occur.

The γ-ray spectrum from SeGA was energy calibrated using the well known, strongly

populated background peaks at 1460.8 keV (40 K γ decay) and 2614.5 keV (208 Tl γ decay).

Efficiency calibrations were performed using three different sources, 154,155 Eu and 60 Co.

The sources were placed on the beam axis and data from the decaying sources was collected.

Then Monte Carlo simulations of the calibration runs using GEANT 4 [12] were performed

and compared to data collected.

Background in the SeGA γ-ray spectra was further reduced by using coincidence gating

on signals from β-decays in the GeDSSD. Only when high-gain events were recorded in

the GeDSSD would events be accepted by SeGA. The timing gate used was a 1.2 µs gate.

This reduces background since the setup is constantly detecting background radiation, for

11
example the 1461 keV γ ray from 40 K. The only time that any of the 40 K γ rays are included

in the final data is when they are detected within the 1.2 µs window surrounding a β or

β-delayed proton decay event in the GeDSSD.

12
Chapter 3

Data Analysis and Discussion

After the data was obtained and sorted information about the 26 P(βpγ)25 Al decay channel

to be extracted. The intensity of the population of each of the excited states of 25 Al could

be determined, their excitation energy and also information about the center of mass (CM)

proton energy could be determined for certain energy levels. Each of these will be discussed

in more detail in the following sections.

3.1 Gamma-ray Intensities

The peaks determined to be from the 26 P(βpγ)25 Al decay channel were observed at 452,

493, 844, 930, 944, 1338, 1613, 1776 and 1790 keV and are labeled in Fig. 3.1. The peaks

were first identified by searching for peaks at energies previously reported for this decay

in Ref. [18]. Peaks at 930 and 1776 keV were new to this decay channel and identified

as possible 25 Al γ-rays by comparing them to energy differences for known 25 Al level [8].

They correspond to the transitions from excited states at 2720 to 1790 keV and 2720 to 945

keV, respectively, as see in Table 3.2. Coincidence gating was also performed to support the

previously known level scheme and the peaks observed in coincidence are reported in Table

3.1.

To fit the peaks and determine the number of events pertaining to each energy the

13
Figure 3.1: 26 P β-delayed γ-ray spectrum. All γ-ray peaks attributed to the 26 P(βpγ)25 Al
decay are labeled by their energy in keV (black). Other peaks are labeled by the γ-ray
emitting nuclide, with escape peaks denoted by an asterisk (gray). Selected regions are
shown in more detail in Fig. 3.7 and Fig. 3.8.

Table 3.1: 26 P(βpγ)25 Al γ rays observed in the present work. The measured γ-ray energies
are reported in the 1st column with their statistical uncertainties only; the global systematic
uncertainty is 0.5 keV. An asterisk denotes γ rays observed for the first time in 26 P β-decay.
The γ-ray intensity per 26 P decay is reported in the 2nd column, where the intensity of the
1613-keV line from [18] was used for normalization. The 3rd column lists γ rays observed in
coincidence.

Energy (keV) Intensity (%) γ-ray coincidences

451.9(3) 2.6(3) 493, 844, 930, 1338, 1776
493.1(4) 2.4(3) 452, 844, 1776
843.5(3) 0.8(2) 452, 493, 944
930.4(5)∗ 0.09(5) 452, 944
944.4(2) 1.2(1) 844, 930, 1776
1338.0(2) 0.8(1) 452
1613.1(3) 2.2(2)
1775.5(3) ∗ 1.2(1) 452, 493, 944
1790.2(3) 0.8(3)

14
response function of SeGA had to be determined. The response function used to fit the

peaks in the γ-ray spectrum was an Exponentially modified Gaussian (EMG).

A erf( x−µ τ 2
τ + σ ) ) exp( τ + x − µ )
f (x) = (1 − √ (3.1)
2σ 2 2σ 2 σ

where A is the amplitude or integral of the peak, x represents energy, µ represents the

center energy, σ is the width and τ describes the decay of the exponential component.

The reason why the EMG works well as a response function for this spectrum is that it

incorporates the Gaussian resolution of the Ge detectors with a low energy tail.

This response function worked well for most of the 25 Al peaks. However, the fitting of

the Doppler broadened peaks at 1613 and 1776 keV was more complex. The modeling of

these peaks will be discussed in detail in the following sections.

Using well-known 24 Mg γ-ray energies from β decay of the 24 Al beam contaminant [9] a

2nd -degree polynomial energy-calibration function was created. The calibration used well-

known room background peaks in the γ-ray singles spectrum to verify its accuracy of 0.5

keV. The calibrated γ-ray energies are reported in Table 3.1. The 844 keV peak contains

a small contribution from an unresolved 26 P(βγ)26 Si line at nearly the same energy. In

this case, we report the energy of the combined peak. The energies are all consistent with

previously reported values when known [8].

Simulations using the GEANT4 Monte-Carlo package [12] were then compared to data

taken offline using an absolutely calibrated 154,155 Eu source and the relative intensities of

the 24 Mg lines from online data to establish the efficiency as a function of energy of the

SeGA detectors. The efficiency curve allowed for the relative intensities of the 25 Al γ-rays

to be determined. Due to difficulties in determining the total number of 26 P β-decays that

15
occurred in the experiment the absolute intensities were found by normalizing to the 1613

keV γ ray, which is known to have an absolute intensity of 2.2 ± 0.2 % [18] based on the

proton feeding of the 1613 keV excited state (Table 3.1). Due to the close proximity of the

844 keV peak and the 842 keV 26 Si peak, its intensity and uncertainty were determined by

combining the acquired 26 Si data set with sd-shell model calculations [4] to predict, and

subtract, the small contribution of 0.33 ± 0.17 % from the 26 P(βγ)26 Si line.

3.2 βp Feeding

The feeding of each 25 Al level via βp decay of 26 P was calculated by subtracting the intensity

of γ-decay branches feeding it from the intensity of γ-decay branches de-exciting it. For

example, as seen in Figure 3.2, the 944 keV excited state is not only fed by proton emission,

but also the de-excitation of higher-energy states via the emission of 844 and 1776 keV γ

rays. The 944 keV state de-excites via the emission of 493 and 944 keV γ rays. The feeding

of the 944 keV state from proton emission is equal to the sum of the intensities of the emitted

γ rays minus the sum of the intensities of the γ rays which feed the 944 keV state, all of

which can be found in Table 3.1.

Since this experiment was not sensitive to the feeding of the ground state of 25 Al the

value of this feeding reported in Table 3.2 and Fig. 3.2 are adopted from [18]. The βp feeding

is summarized in Table 3.2 and illustrated in Fig. 3.2. There is good agreement with most of

the proton-feeding values from Thomas et al. [18]. The only exception is the feeding of the

first excited state at 452 keV. Some of the difference can be attributed to their insensitivity

to the 2720 keV state, which provides a significant γ-ray feeding of the 452 keV state. The

small proton feeding can be explained by the need for an ` ≥ 2 proton to populate this

16
Table 3.2: The 26 P(βp) feeding of 25 Al states found in the present work and previous work
[18]. The 452 keV β-feeding is given as an upper limit. Upper limits are calculated at a 95%
confidence-level. An asterisk denotes evidence for excited states observed for the first time
via this decay channel. The intensities are normalized to the feeding of the 1613-keV level
from [18].
25 Al excitation energy (keV) Proton feeding
Present work Ref [18]
(%) (%)
Ground state 27.3 (4)
452 <0.34 2.1 (1)
944 1.6(3) 2.1 (5)
1613 2.2(2) 2.2 (2)
1790 2.3(2) 2.3 (2)
2720* 1.1(1)

J π = 1/2+ state from the J π =(2,3,4)+ 26 Si states fed by allowed 26 P β decay transitions.

3.3 Study of Doppler Broadened Peaks

Two peaks in the γ-ray spectrum clearly exhibit Doppler broadening from the recoil due

to proton emission. Those peaks are found at 1613 and 1776 keV in the spectrum of Fig.

3.1. The peaks at 1613 keV (Fig. 3.7) and 1776 keV (Fig. 3.8) are from the de-excitation

of the 1613 and 2720 keV excited states of 25 Al respectively. Since all of the 25 Al peaks

in the spectrum are populated due to βp decay one might expect that Doppler broadening

should be observed in all of the 25 Al peaks analyzed in this work. However, due to the

proportionality of the Doppler shift on γ-ray energy, we were not sensitive to the Doppler

broadening of the lines at 1338 keV and below. Unfortunately, due to the close proximity of

the intense 1797 keV 26 Si peak to the 1790 keV 25 Al peak, it was not possible to study the

17
Figure 3.2: 25 Al level scheme from 26 P(βpγ)25 Al decay deduced from the present work.
The γ-ray transitions observed are denoted by arrows with thicknesses proportional to their
intensities and labeled by γ-ray energy in keV. The βp feeding of the different excited states
is depicted by the arrows on the right, which are labeled by the intensities. The single
asterisk denotes a value adopted from [18]. The double asterisks denotes the upper limit of
the β-feeding of the 452 keV state at the 95% confidence level.

broadening of the 1790-keV peak precisely. The physical process creating broadened peaks in

our γ-ray spectrum was discussed in detail in the Introduction. Fitting these peaks requires

several steps. If the broadened peak is from a state fed by a previously known proton energy,

then the kinetic energy of the recoiling nucleus that emits the γ-ray can be calculated. For

example, consider only the 2288 keV CM proton energy which feeds the 1613 keV excited

state in 25 Al. One can find the initial velocity of the excited 25 Al nucleus by conservation

of momentum and energy as seen below.

18
 1 1
E = mp vp2 + mAl vAl
2 = 2288keV (3.2)
2 2

p = mp vp + mAl vAl (3.3)

1 m v 1
E = mp ( Al Al )2 + mAl vAl
2 = 2288keV (3.4)
2 mp 2

m
vAl = 8.2 · 105 (3.5)
s

Where m is mass, v is velocity and subscripts p and Al stand for proton and 25 Al nuclei

respectively. Assuming that the γ decay occurred in free space this would be the velocity

of the nucleus at the time of γ decay and the velocity can be used directly to calculate the

Doppler broadening.

0 c
E(max) =( )Eγ = 1616keV (3.6)
c − val

0
where E(max) is the maximum shifted energy that can be detected due to Doppler shift
0
and Eγ is the original unshifted energy of the γ ray. E(max) can either be the maximum

or minimum possible Doppler shifted energy depending on the sign of val . We can find the

difference in these two energies, 3 keV, which will be used as the parameter which determines

how broad our peak of interest is [10]. For these broadened peaks, a model using a boxcar

function convoluted discretely with the EMG response function is employed. A simplified

version can be seen in Equation 3.7.

19
Figure 3.3: This plot shows how the boxcar function convoluted EMGs produce a broadened
peak. In this example µ= 1613 keV, σ = 1.54, τ = 1.2, λ = 3 keV and the amplitude of the
smaller peaks are set to one.

A erf( x−µ+λ
τ + στ ) τ2 x−µ+λ
f (x) = (1 − √ ) exp( 2 + )
2σ 2 2σ σ
A erf( x−µ+0.5λ
τ + στ ) τ2 x − µ + 0.5λ
(1 − √ ) exp( 2 + )
2σ 2 2σ σ
A erf( x−µ
τ + στ ) τ2 x−µ (3.7)
(1 − √ ) exp( 2 + )
2σ 2 2σ σ
A erf( x−µ−0.5λ
τ + στ ) τ2 x − µ − 0.5λ
(1 − √ ) exp( 2 + )
2σ 2 2σ σ
A erf( x−µ−λ
τ + στ ) τ2 x−µ−λ
(1 − √ ) exp( 2 + )
2σ 2 2σ σ

where A is the amplitude, µ is the central energy, τ is the decay constant, σ is the width
0
of the Gaussian and λ is the “stretch parameter”, which is the difference in E(max) and Eγ

calculated above. Using the parameters relevant for the 1613 keV peak, we can plot the

simplified version of the peak and all of its contributions.

However, in the actual case of the 1613 keV γ-ray peak, it is known that two different

CM energy protons feed the 1613 keV level in 25 Al [18] as seen in Fig.3.4. Therefore,

20
the function used in modeling this peak requires a linear combination of two boxcar EMG

functions. These boxcar functions are centered at the same energy and have the same τ and

σ parameters. The difference in the two boxcar functions is that one of the functions has an

amplitude that is twice the other [18] and that their λ stretch parameter is different.

Figure 3.4: The previously known decay scheme for the 26 P β-delayed proton emission to
the 1613 keV excited state of 25 Al [18]. Two proton-unbound 26 Si states feed the 1613 keV
25 Al excited state, causing two different 25 Al recoil velocities following proton emission.

Another way that our simplified case is not appropriate for our actual data is that the

decay does not occur in free space. It instead occurs in the GeDSSD, which is germanium

and, therefore, the recoiling nucleus will begin to slow down as soon as it is emitted. It is

necessary to incorporate the half life of the excited state of 25 Al that the γ-ray is emitted from

and it is used along with the stopping power of the material in which the nucleus is moving

to determine the velocity of the 25 Al at the time of γ decay. As shown previously, once the

velocity at the time of γ-ray emission is determined along with the expected energy of the

γ ray the maximum possible γ-ray energy that can be detected due to Doppler broadening

can be calculated.

The final piece of information missing from our simplified case is the fact that not all of

the decays occurs exactly at the half-life. Instead that is the time by which 50% of decays are

21
expected to have occurred. To more accurately depict this in our model we expanded what

was one boxcar function for each proton into 5 boxcar functions each of which represent

a different time interval between proton emission and γ-ray emission. We discretized the

decay,

t
1 t1
N (t) = N0 2 (3.8)
2

where N (t) is the current number of particle that have not yet decayed, N0 are the

number of original particles that can decay, t 1 is the half-life and t is time.
2
Using the known value of τ , the times at which 10, 30, 50, 70 and 90% of the decays

are expected to have occurred were calculated. Each of these discrete times was used to

represent 20% of the decays, providing five relationships between the initial velocity and the

velocity at the time of γ-ray emission through the stopping power. Figure 3.5 shows the

exponential decay for a particle with a theoretical half-life of 10 fs and initial count of 100

particles. Black lines divide the decay where 20%, 40%, 60% and 80% of the particles have

decayed. For each C. M. proton energy a boxcar step function was created corresponding

to each of the five times such that they had equal integrals and stretched in unison. The

stopping power was treated by utilizing SRIM tables generated for 25 Al ions in germanium

[20]. For the 1776-keV case the stopping power was treated iteratively in 25 fs steps to

account for its energy dependence due to its relatively long half-life of 201(14) fs [9]. Figure

3.6 shows how the shape of the peak changes when evaluated at different times along the

decay for the 1776 keV γ-ray case. It illustrates why it is important to evaluate the Doppler

shift at more than just the half-life of the decay. There are more accurate ways to treat

the half-life and stopping power; however, this discrete treatment was sufficiently accurate

22
considering the statistical uncertainties in the spectrum and the absolute uncertainty in the

stopping power, which was approximated to be 10% based on the scatter in the experimental

data plots from SRIM for aluminum ions in germanium.

Figure 3.5: The decay of a theoretical particle with initially 100 daughter particles and a
half-life of 10 fs is shown above. For the analysis the decay curve was divided into five
sections like shown above. The time used to represent the Doppler broadening of the first
20% of decaying particles would be the time at which it is expected for 10% of the particle
to have decayed. The other fifths are evaluated in the same manner.

3.4 1613 keV

The C. M. energies and relative intensities for the protons that feed the 1613 keV excited

state are well known [18]. With this previously reported information we could test the

Doppler broadening method by adopting these energies along with the known lifetime to

constrain the fit (Fig. 3.7).

The Compton background in the 1613-keV region was continuous and flat and was there-

23
Figure 3.6: The components of the final shape of the 1776 keV γ-ray peak are shown above.
Each peak has the same number of counts, but the spreads due to Doppler broadening are
different due to evaluating them at the time here 10% (green solid line), 30% (orange dashed
line), 50% (grey dotted line), 70% (yellow dot-dashed line) and 90% (blue double dot-dashed
line) of the particles have decayed. The response function used for this figure is Gaussian
for simplicity.

fore modeled with a straight line. The neighboring contaminant 26 Al peak at 1622.26(3)

keV (from 26 Si β decay) [7] was not broadened and could be modeled with a simple EMG

with the amplitude and the centroid as free parameters. Another background peak is in

the region at 1611.807(11) keV [7] and is the result of the well-known β decay of 25 Al to

excited states of 25 Mg. Since the absolute intensity of this γ ray is known it was possible

to fix the intensity of the corresponding peak to be 0.112 times that of the 1613 keV peak.

This confirms that the overlap of the weak 25 Mg γ-ray line to the 25 Al γ-ray line does not

contribute significantly to the broadening of the 1613 keV peak.

Initially the 1613 keV peak was treated as if there was no Doppler broadening by using

the same response function used to model the other peaks in the spectrum. This yielded

24
 800

700

600
Counts
500

400

300

200

1605 1610 1615 1620 1625

Energy (keV)
Figure 3.7: 26 P β-delayed γ-ray spectrum in the region of the 25 Al peak at 1613 keV. The
peak at 1613 keV is broader than the neighboring 26 Al peak at 1622 keV. The solid blue
line is the overall fit including Doppler broadening and the red dotted line represents the
Compton scattering background. Below the data and fit, the individual peak components
are shown. The 1611 keV 25 Mg, 1613 keV 25 Al and 1622 keV 26 Al γ-ray lines are represented
by the green dot-double-dashed, gold dashed and light blue dot-dashed lines, respectively.

a high χ2 per degree of freedom of 138/27 corresponding to a p-value of 0.0001. It was,

therefore, clear that the assumption that the 1613 keV peak should be modeled with the

same EMG response function without broadening is incorrect.

Next Doppler broadening due to the proton emission was incorporated into the model,

as described previously. There are two different excited states above the proton threshold

in 26 Si that emit protons populating the 1613 keV excited state of 25 Al, which undergoes a

γ-ray transition to the ground state [18]. The C. M. energies of the two protons are 2288(3)

keV and 5893(4) keV and they have a relative intensity of I2288 /I5893 = 2.0 (Fig. 3.4) [18].

There was substantial improvement of the fit after including Doppler broadening with these

known values (Fig. 3.7) and it is reflected in the improvement in the χ2 per degree of freedom

to 31.5/27 corresponding to a p-value of 0.25. This confirmed the Doppler broadening of

this line as well as the accuracy of the Doppler broadening analysis technique [10]. We were

25
 25
500 25
Al
Al
214 26
Po, Si 26
26
Si
400 Si
Counts

300 (25Al)

200

100

1750 1760 1770 1780 1790 1800 1810

Energy (keV)
Figure 3.8: (Color online) 26 P β-delayed γ-ray spectrum (red crosses) in the region of the
1776- and 1790-keV peaks with fit (blue solid line). The dashed green line shows the con-
tribution of the Doppler broadened 1776-keV peak to the overall fit. The dot-dashed purple
line shows the contributions of all other peaks. The small excess near 1810 keV could not
be confirmed to be from 25 Al.

further encouraged to utilize this method to try to extract the CM energy from the 1776

keV peak which has no known proton CM energy.

3.5 1776 keV

The 1776 keV γ-ray transition from the 2720 keV excited state of 25 Al was observed for the

first time in 26 P β decay (Fig. 3.8) [17]. Since it has never been observed via this decay

mechanism the C. M. proton energy feeding the 2720 keV excited state was unknown. This

allowed us to extract this new information using the application of the Doppler broadening

method.

As in the 1613 keV region, the continuous Compton scattering component of the back-

26
Figure 3.9: The blue dots show the χ2 value for each of the initial 25 Al kinetic energy tested.
The blue line shows the quadratic fit line that was used to determine the energy with the
minimum χ2 . The minimum was found to be at 195 +41 −50 (stat.) ± 18 (syst.) keV

ground was modeled to be linear. To provide an accurate representation of the background

underneath the 1776-keV peak it was necessary to include the other peaks in the region in

the fit function. The 1790 keV 25 Al peak on the shoulder of the strong 1797-keV 26 Si line

was modeled to be Doppler broadened; its shape was constrained using the three known

proton energies [18] feeding it, their ratios and the known half life of the 1790 keV excited

state [8].

To find the CM energy for the 1776 keV peak while incorporating stopping power and

the half-life we used hypothesized initial 25 Al recoil energies and varied them in the fit from

100 to 325 keV in 15 keV steps. The best fit and χ2 value were found for each of these

energies.

Plotting the CM energies versus the χ2 value of the fit resulted in an optimal 25 Al

initial kinetic energy of 195 +41

−50 (stat.) ± 18 (syst.) keV based on the minimum value of

χ2 as seen in Fig. 3.9. Using conservation of energy and momentum the corresponding

C. M. proton energy was found to be 5.1 ± 1.0 (stat.) ± 0.6 (syst.) MeV corresponding

27
Table 3.3: Sources of uncertainty in the 5.1 ± 1.0 (stat.) ± 0.6 (syst.) MeV 26 P β-delayed
proton C. M. energy feeding the 2.72 MeV 25 Al state.

Source of Uncertainty
uncertainty (MeV)
statistics 1.0
response function 0.3
stopping power 0.5
background 0.1

to a proton-emitting 26 Si level at an excitation energy of 13.3 ± 1.0 (stat.) ± 0.7 (syst.)

MeV. Systematic uncertainties for the proton energy were derived from the uncertainties

in the shape parameters of the response function, uncertainties in the stopping power, and

uncertainties in the background. A summary of the uncertainties can be found in Table 3.3,

which shows that the stopping power contributes the dominant systematic uncertainty. The

only proton unbound state of 26 Si that is consistent with our measured proton energy is the

isobaric analog state (IAS) of 26 P at 13.015(4) MeV (the only known excited state above

10.8 MeV), which is also known to be strongly populated in 26 P β decay [18].

28
Chapter 4

Conclusion and outlook

While the method demonstrated in the present work is not quite of similar accuracy to direct

techniques which measure the proton CM energies from β-delayed proton-γ decay, it is much

more difficult to measure β-delayed neutron CM energies. The present method may prove to

be very useful to the measurement of neutron energies since it does not require the emitted

nucleon to be charged. There have been proposals to further refine the way γ-ray data is

collected which will be a next potential step for this method.

To obtain more precise results in the future it is possible that Crystal diffraction spec-

trometer (CDS) measurements will be the key [5]. It has already been proposed to use

CDS at ISOL@MYRRHA (Isotope separator on-line) in Belgium to study β-delayed neu-

tron emission [19]. The advantages to using CDS is that the resolution of γ ray energies is

much better than that of HPGe crystal detectors. The down side it that the detector for

CDS necessarily has a low acceptance, giving it a much smaller solid angle, and therefore

a much lower efficiency. One way to combat this is to use very intense beams (like those

produced at ISOL@MYRRHA or, potentially, FRIB) and to have long experimental run

times to allow for the sufficient collection of statistics.

The great resolution of CDS would allow for the CM energies from higher mass cases to

be determined. It will also yield more precise energies for lower mass cases. While this may

not be a competitive method to obtain β-delayed proton CM energies, it is significant to the

detection of the CM energies of β-delayed neutrons, due to their neutral nature.

29
 To conclude, this was the first observation of the radiative Doppler broadening from β-

delayed proton-γ decay, and it was also the highest mass (A=25) for which Doppler broad-

ening has been observed in any β-delayed nucleon-γ decay. Since there was not one but

two peaks with obvious Doppler broadening this not only allowed us to test the technique

presented in [10] with a case with a known CM proton energy, but also apply the technique

to determine a previously unknown CM proton energy. Absolute intensities of 8 different

energy γ rays from 26 P(βpγ)25 Al were determined as was the β feeding of 5 25 Al excited

states from the same decay. The present results have been published in Ref. [17].

30
Bibliography

[1] D. Bazin et al. “Radio Frequency Fragment Separator at NSCL”. In: Nucl. Instrum.

Methods A 606 (July 2009), pp. 314–319. doi: 10.1016/j.nima.2009.05.100.

[2] M. B. Bennett et al. “Classical-Nova Contribution to the Milky Way’s Al26 Abundance:

Exit Channel of the Key Al25(p,γ)Si26 Resonance”. In: Phys. Rev. Lett. 111.23 (2013),

p. 232503. doi: 10.1103/PhysRevLett.111.232503.

[3] M. J. Borge et al. “Elucidating halo structure by β decay: βγ from the 11 Li decay”.

In: 55 (Jan. 1997), p. 8. doi: 10.1103/PhysRevC.55.R8.

[4] B. A. Brown. private communication. 2013.

[5] SCK CEN. MYRRHA: Multi-purpose hybrid research reactor for high-tech applications.

2016. url: http://myrrha.sckcen.be/en (visited on 03/25/2016).

[6] S. Y. F. Chu, L. P. Ekström, and R. B. Firestone.

[7] P. M. Endt. “Supplement to Energy Levels of A=21-44 Nuclei (vii)”. In: Nucl. Phys.

A633 (Apr. 1998). www.nndc.bnl.gov/ensdf, pp. 1–220.

[8] R. B. Firestone. In: Nucl. Data Sheets 110 (July 2009), p. 1691. doi: 10.1016/S0168-

583X(03)00570-6.

31
 [9] R. B. Firestone. “Nuclear Data Sheets for A = 24”. In: Nucl. Data Sheets 108 (Nov.

2007), p. 2319. doi: 10.1016/j.nds.2007.10.001.

[10] H. O. U. Fynbo. “Doppler broadened γ-lines from exotic nuclei”. In: Nucl. Instrum.

Methods Phys. Res. B 207 (2003), p. 275. doi: 10.1016/S0168-583X(03)00570-6.

[11] H. O. U. Fynbo et al. “New information on the β-decay of 11 Li from Doppler broadened

γ lines”. In: Nucl. Phys. A 736 (May 2004), pp. 39–54. doi: 10.1016/j.nuclphysa.

2004.02.020.

[12] Geant4 Collaboration and S. Agostinelli et al. (Geant4 Collaboration). “Geant4-a sim-

ulation toolkit”. In: Nucl. Instrum. and Methods A 506 (July 2003), pp. 250–303.

[13] N. Larson et al. “High efficiency beta-decay spectroscopy using a planar germanium

double-sided strip detector”. In: Nucl. Instrum. Methods Phys. Res. A 727 (2013),

p. 59. doi: 10.1016/j.nima.2013.06.027.

[14] C. M. Mattoon et al. “Line-shape analysis of Doppler-broadened γ lines following the

β decay of Li11”. In: 80.3, 034318 (Sept. 2009), p. 034318. doi: 10.1103/PhysRevC.

80.034318.

[15] D. J. Morrissey et al. “Commissioning the A1900 projectile fragment separator”. In:

Nucl. Instrum. Methods B 204 (May 2003), pp. 90–96. doi: 10.1016/S0168-583X(02)

01895-5.

[16] F. Sarazin et al. “Halo neutrons and the β decay of 11 Li”. In: 70.3, 031302(R) (Sept.

2004), 031302(R). doi: 10.1103/PhysRevC.70.031302.

[17] S. B. Schwartz et al. “Observation of Doppler broadening in β-delayed proton-γ decay”.

In: Phys. Rev. C 92 (3 2015), p. 031302. doi: 10.1103/PhysRevC.92.031302. url:

http://link.aps.org/doi/10.1103/PhysRevC.92.031302.

32
[18] J.-C. Thomas et al. “Beta-decay properties of 25 Siand26 P ”. In: Eur. Phys. J. A 21

(Sept. 2004), pp. 419–435. doi: 10.1140/epja/i2003-10218-8.

[19] P. C. Womble et al. The Doppler-broadening of Gamma-ray Spectra in Neutron-Based

Explosives Detection Systems. 2009. url: http : / / www - pub . iaea . org / MTCD /

publications/PDF/P1433_CD/datasets/abstracts/sm_en-16.html.

[20] J. F. Ziegler and J. P. Biersack. “The Stopping and Range of Ions in Matter”. In: Trea-

tise on Heavy-Ion Science, by Bromley, D. Allan, ISBN 978-1-4615-8105-5. Springer-

Verlag US, 1985, p. 93. Ed. by D. A. Bromley. 1985, p. 93. doi: 10.1007/978- 1-

4615-8103-1_3.

33