LIFETIME MEASUREMENTS OF MAGNESIUM-32 IN THE N = 20

ISLAND OF INVERSION

By

Robert Martin Elder

A DISSERTATION

Submitted to

Michigan State University

in partial fulfillment of the requirements

for the degree of

Physics - Doctor of Philosophy

2020

LIFETIME MEASUREMENTS OF MAGNESIUM-32 IN THE N = 20 ISLAND OF

ABSTRACT

INVERSION

By

Robert Martin Elder

Understanding how nuclear shell structure and magic numbers emerge near stability but

change as nuclides move away from stability is one of the main goals of modern nuclear struc-

ture studies. The N = 20 island of inversion is one of the oldest and best-studied examples

of a magic number disappearing in exotic isotopes, but new observations are still emerging.

Nuclides in this region exhibit collective behavior driven by the neutron particle-hole in-

truder configurations across the reduced gap between the sd and pf shells. The assignment

of the 32Mg nuclide to the N = 20 island of inversion is supported from studies of the mass,

excited states, and transition rate to the 2+

standing of this classic island of inversion nuclide is still a challenge. The 0+

1 state of 32Mg. However, a complete under-
2 state of 32Mg

was observed at a lower energy than many models predicted, leading to a reinterpretation of

the collectivity in 32Mg. Reduced transition rates are valuable observables to characterize

collectivity. However, precise measurements of the reduced E2 transition rates in 32Mg are

still needed to test the validity of recent theoretical calculations.

This dissertation reports the results from an experimental study of the collectivity in

32Mg. The present work includes lifetime measurements for the lowest 0+, 2+, and 4+

excited states in 32Mg which are used to determine the reduced E2 transition rates. Life-

time measurements were performed using two different and complementary experimental

approaches. First, the TRIPLEX device was used along with the Recoil-distance Method

and Doppler-shift Attenuation Method to measure the lifetimes of the short-lived 2+

1 and

4+
1 states in 32Mg. The results reinforce the evidence from the energy systematics and ear-
lier B(E2) measurements which suggests that the ground-state band in 32Mg is collective.

Second, a novel in-flight technique referred to here as the Cascade Doppler-shift Method

was used to observe decays from the 0+

2 isomer in 32Mg. The 0+

2 state was confirmed and

the measured lifetime suggests a large collectivity in this state as well. The small partial

cross section populating the 0+

2 state in the 9Be(34Si,32Mg)X reaction provides experimental

evidence for the dominance of intruder configurations in the 0+

2 state.

In summary, the lifetime measurements presented in this dissertation indicate the collec-

tivity in 32Mg driven by intruder configurations that dominate the ground state as well as

low-lying excited states. Experimental methods based on in-flight Doppler-shift techniques

are demonstrated in this work and highlight how a broad range of lifetimes from 1 ps to tens

of ns can be measured in rare-isotope beam experiments.

Copyright by
ROBERT MARTIN ELDER
2020

ACKNOWLEDGMENTS

There are many people whom I must thank for their support during my thesis research.

Foremost, I am thankful for the guidance of my research advisor Dr. Hironori Iwasaki. His

advice has consistently encouraged me to reach greater rigor in my research and writing. I

have learned a great deal about physics and about how to succeed as a scientist from Dr.

Iwasaki. I am also deeply grateful for his sincere support when I would occasionally struggle

with my work.

I also wish to thank the members of my guidance committee, Dr. Alexandra Gade, Dr.

Heiko Hergert, Dr. Sean Liddick, and Dr. Stuart Tessmer for their encouragement and

advice. They have challenged me to produce better work and also ensured that I had all

that I needed throughout my research.

The experiment described in this work would not have been possible without the expertise

of a great many people. The past and present members of the lifetime group have always

provided me with valuable feedback. I am thankful for their hard work. I would also like to

thank the gamma group for their expertise and skill. I am grateful to the entire experimental

collaboration and the support staff at the National Superconducting Cyclotron Laboratory

who made the experiment a success with their professionalism and attention.

My family and friends have been a constant source of affirmation and love for a long time

before this research began. I could not have reached where I am now without my parents

Joan and Andy. Thank you both for everything. Before I could take my first steps, you were

there to raise and guide me. I also thank my friends from Michigan State University, from

Ohio Wesleyan University, and from everywhere else who have kept me in good cheer even

when I was faced with self-doubt. Finally, I am thankful for my son Kolten who has been

v

incredibly patient as I write this dissertation. The joy and pride he gives me has made me

a better person.

vi

TABLE OF CONTENTS

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

ix

LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

x

Chapter 1

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.1 The Atomic Nucleus
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.2 The Shell Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.3 Nuclear Collectivity . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.4 Electromagnetic Transitions . . . . . . . . . . . . . . . . . . . . . . . . . . .
Islands of Inversion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
1.5
1.6 Collectivity in 32Mg
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.1.1 Gamma-ray Interactions in Matter
2.1.2 Relativistic Doppler Shift

Chapter 2 Experimental Techniques . . . . . . . . . . . . . . . . . . . . . . .
2.1 Gamma-ray Spectroscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .
2.2 Lifetime Measurement Techniques . . . . . . . . . . . . . . . . . . . . . . . .
2.2.1 Doppler-shift Attenuation Method . . . . . . . . . . . . . . . . . . . .
2.2.2 Recoil-distance Method . . . . . . . . . . . . . . . . . . . . . . . . . .
2.2.3 Cascade Doppler-shift Method . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2.3 Direct Reactions

Chapter 3

Setup and Calibration of Experimental Devices . . . . . . . . .
3.1 Superconducting Source for Ions (SuSI) . . . . . . . . . . . . . . . . . . . . .
3.2 The Coupled Cyclotrons, K500 and K1200 . . . . . . . . . . . . . . . . . . .
3.3 A1900 Fragment Separator . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4 S800 Spectrograph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.1 Cathode Readout Drift Chambers . . . . . . . . . . . . . . . . . . . .
Ionization Chamber . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.2
3.4.3 E1 Timing Scintillator
. . . . . . . . . . . . . . . . . . . . . . . . . .
IsoTagger Hodoscope . . . . . . . . . . . . . . . . . . . . . . . . . . .
3.4.4
3.4.5 Run-by-run Corrections
. . . . . . . . . . . . . . . . . . . . . . . . .
. . . . .
3.5.1 GRETINA Description . . . . . . . . . . . . . . . . . . . . . . . . . .
3.5.2 GRETINA Calibration . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6 TRIple PLunger for EXotic Beams (TRIPLEX) . . . . . . . . . . . . . . . .
3.6.1 TRIPLEX Description . . . . . . . . . . . . . . . . . . . . . . . . . .
3.6.2 TRIPLEX Calibration . . . . . . . . . . . . . . . . . . . . . . . . . .
3.7 Experimental Setup to Study Short-lived States . . . . . . . . . . . . . . . .
3.8 Experimental Setup to Study Isomeric States . . . . . . . . . . . . . . . . . .
3.9 Data Analysis Tools
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

3.5 Gamma-Ray Energy Tracking In-beam Nuclear Array (GRETINA)

1
1
4
9
12
16
18

22
23
23
25
28
29
31
34
36

38
39
40
40
42
44
48
51
54
55
58
58
61
64
65
70
74
78
81

vii

SpecTcl Online Data Analysis . . . . . . . . . . . . . . . . . . . . . .
3.9.1
3.9.2 GrROOT and ROOT Offline Data Analysis
. . . . . . . . . . . . . .
3.9.3 G4Lifetime Simulation . . . . . . . . . . . . . . . . . . . . . . . . . .

81
81
82

Chapter 4 Data Analysis and Results

4.1

4.2
4.3
4.4
4.5

. . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . . . .

88
32Mg Short-lived States
89
32Mg Target-only Setting
90
4.1.1
32Mg TRIPLEX Large-separation Setting
95
4.1.2
32Mg TRIPLEX Short-separation Settings . . . . . . . . . . . . . . . 101
4.1.3
30Mg 2+
1 State
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108
20O 2+
1 State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111
31Mg (7/2−) Isomer
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 116
32Mg 0+
2 Isomer . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 119
4.5.1 Energy Measurement and Distribution of Decay Points . . . . . . . . 120
4.5.2 Constraints on the Cross Section and the 0+
2 Lifetime . . . . . . . . . 121
Independent Cross Section Measurement . . . . . . . . . . . . . . . . 123
4.5.3

Chapter 5 Discussion of Results . . . . . . . . . . . . . . . . . . . . . . . . . . 131
5.1 B(E2) Values in 32Mg and Even-even Mg Isotopes
. . . . . . . . . . . . . . 131
5.2 The 32Mg 4+
1 State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135
5.3 The 32Mg 0+
2 State . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 137
5.4 Cross Section to 32Mg 0+ States . . . . . . . . . . . . . . . . . . . . . . . . . 138

Chapter 6 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . 140

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 143

viii

LIST OF TABLES

Table 3.1: The separations between the foils in the TRIPLEX device. T refers to the
target, 1D refers to the first degrader, and 2D refers to the second degrader.
The S800 setting refers to whether the S800 spectrograph was tuned to the
unreacted 34Si secondary beam or the 32Mg reaction product.
. . . . . .

Table 4.1: Prompt decays in 32Mg observed in this experiment. The populations
. . . . . . . . . . . .

observed in this work are consistent with Ref. [60].

76

95

Table 4.2: Two-proton removal cross sections in the 9Be(34Si,32Mg)X reaction are
shown. The column labeled σexp is from the present experiment and the
column labeled σprev shows the results of the work from Ref. [60]. The
upper limit for the exclusive cross section of the 0+
2 state is based on the
unobserved gamma-ray intensity as described in the main text.

. . . . . . 130

Table 5.1: Energy ratio and B(E2) ratio observed for 32Mg and from several predic-

tions. The EKK predictions are from Ref. [22].

. . . . . . . . . . . . . . . 136

Table 5.2: B(E2) values in 32Mg and neighboring even-even isotopes. . . . . . . . . . 137

ix

LIST OF FIGURES

Figure 1.1: A Segr´e chart depicting the known nuclides. The nuclides are arranged
by increasing N along the x-axis and increasing Z along the y-axis. The
decay mode of each nuclide is represented by the color. The blue and
black narrow rectangles represent the proton and neutron magic numbers.
Magic number 2 is not labeled in this figure. The figure is adapted from
Ref. [1] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1.2: The energy of the first excited states in even-even nuclides. The large
excitation energies at proton or neutron number of 8, 20, 28, 40, 50, 82,
and 126 are evidence for closed shells at these magic numbers. The figure
is adapted from Ref. [2].
. . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1.3: The single-particle orbits for neutrons in the 208Pb nucleus, calculated for
the harmonic oscillator potential, Woods-Saxon potential, and Woods-
Saxon potential with spin-orbit term. The figure is adapted from Ref. [2].
The number in brackets is the occupancy of that orbit and is followed
by the cumulative occupancy in that orbit and all lower-energy orbits.
The orbits are labeled by the harmonic oscillator quantum number N , the
orbital angular momentum quantum number l = s, p, d, f, g, h, i mean-
ing l = 0, 1, 2, 3, 4, 5, 6, and twice the total angular momentum quantum
number 2j.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1.4: The forms of the central potentials discussed in the text. The harmonic
oscillator potential is shown as a blue dashed line and the Woods-Saxon
potential is shown as a solid red line. The Woods-Saxon potential is shifted
vertically by one energy unit to align with the minimum of the harmonic
oscillator potential.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 1.5: B(E2; 0+

1 → 2+

1 ) values for argon isotopes, adapted from Ref. [9] . The
B(E2) value is minimized for isotopes with a conventional magic number
of neutrons, N = 20 or N = 28.
. . . . . . . . . . . . . . . . . . . . . . .

Figure 1.6: The energy of the first 2+ states in even-even nuclides. For many nuclides
with the magic number N = 20 or N = 28 the first 2+ energy is larger
than the neighboring even-even isotopes. This is not the case for 32Mg at
N = 20 and for 40Mg and 42Si at N = 28.
. . . . . . . . . . . . . . . . .

Figure 2.1: The mass attenuation coefficients for gamma-ray interactions in sodium
iodide. Photoelectric absorption, Compton scattering and pair production
interactions account for nearly all of the attenuation. The figure is from
Ref. [48]. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

x

2

4

7

8

15

17

24

Figure 2.2: Several methods used to measure the lifetimes of excited states. The figure
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

is from Ref. [52].

Figure 2.3: A schematic of a generic plunger device used in Recoil-distance Method
(RDM) measurements. The image is from Ref. [53]. . . . . . . . . . . . .

Figure 2.4: A generic level scheme in which the highest state C has a long lifetime and
is followed by a cascade of gamma rays γ1 and γ2 involving a short-lived
. . . . . . . . . . . . . . . .
intermediate state B and a ground state A.

Figure 2.5: A generic experimental setup in which two gamma rays γ1 and γ2 are
emitted at angles θ1 and θ2 from an in-flight ion and are detected. The
gamma-ray detector array GRETINA is depicted here [50].
. . . . . . .

Figure 2.6: A depiction of a two-nucleon knockout reaction. The relevant angular
. . . . . .

momentum couplings are shown. This figure is from Ref. [57].

29

32

34

35

37

Figure 3.1: A schematic of the S800 spectrograph. The figure is from Ref. [71].

. . .

42

Figure 3.2: The S800 focal plane. The cathode readout drift chambers (CRDC’s),
plastic scintillator (E1 scintillator), ionization chamber, and IsoTagger
hodoscope are depicted. The image is from Ref. [71].
. . . . . . . . . . .

Figure 3.3: A schematic of a CRDC detector used in the S800 focal plane. The beam
axis z, dispersive axis x, and non-dispersive axis y are shown. The image
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
is from Ref. [71].

Figure 3.4: A schematic of the CRDC detectors used in the S800 focal plane and the
position and angular information of the ion. The beam axis z, dispersive
axis x, and non-dispersive axis y are shown. The first CRDC provides a
measurement of the x and y position at the focal plane, xf p and yf p. The
angle information af p and bf p comes from the difference in the x and y
. . . . . . . . . . . . . . . . . . . .
positions between the two CRDC’s.

Figure 3.5: The 224 channels of each CRDC must have their signals matched so that
there is no bias for a particular channel due to the variations in the gain
of the electronics. The signal from the channels before matching is shown
in the top spectrum for the first CRDC, and the signal from the same
channels after matching is shown in the bottom spectrum. The events are
for the incoming 34Si secondary beam and 32Mg reaction product observed
at the focal plane.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

43

44

45

47

xi

Figure 3.6: The top section shows the uncalibrated x and y positions of the first
CRDC with the mask placed in front. The holes in the mask allow ions to
pass through to the CRDC. The lower section shows the linear calibration
between the raw position values ch and the calibrated positions y. The
line is defined by y = −0.156248ch + 107.28 . . . . . . . . . . . . . . . .

Figure 3.7: The top section shows the uncalibrated energy loss measured by the ion-
ization chamber as a function of xf p. An exponential line is fit to the data
and shown in red. The exponential fit very nearly resembles a straight
line in this case. The bottom section shows the ionization chamber energy
loss calibrated for xf p.
. . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 3.8: The uncalibrated time-of-flight (TOF) is shown with respect to af p and
xf p. The calibrated TOF is also shown, presenting a more uniform TOF
distribution which allows for a more accurate particle identification with
the S800 spectrograph. . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 3.9: A diagram of the IsoTagger hodoscope. The 32 sodium-doped cesium
iodide crystals are arranged into eight rows of four detectors each. The
beam is stopped in an aluminum plate and gamma rays emitted from the
decay of isomers are detected by the IsoTagger hodoscope.
. . . . . . .

Figure 3.10: An IsoTagger spectrum gated on 32Al ions at the focal plane of the S800
spectrograph. The gamma-ray transitions at 221.9 and 734.6 keV which
follow the decay of the isomer at 956.6 keV are observed [76].
. . . . . .

Figure 3.11: yf p distributions, gated on 32Mg reaction products. The distribution be-
fore run-by-run corrections is shown in black, while the distribution after
run-by-run corrections is shown in red. The correction causes the distri-
. . . . . . . .
bution to be centered in the CRDC and slightly narrower.

Figure 3.12: Diagram of a GRETINA module. The full module is shown in the bottom
section. The four high-purity germanium crystals in a single module are
depicted in the top-left. The segmentation of a single crystal is depicted
in the top-right. The figure is adapted from Ref. [67].
. . . . . . . . . .

Figure 3.13: The efficiency of GRETINA measured using a 152Eu radioactive source.
For this measurement, the source was placed at the center of GRETINA.
The efficiency of GRETINA in the GEANT4 simulation used in this work
is also shown with red cross (+) symbols. The simulation efficiency is
greater than the experimental efficiency by a factor of 1.13(2). The simu-
lation efficiency scaled down by 1.13 is shown with red saltire (×) symbols
and matches the data well. . . . . . . . . . . . . . . . . . . . . . . . . . .

48

50

53

54

56

57

59

62

xii

Figure 3.14: The efficiency of GRETINA for the 779-keV gamma ray from a 152Eu
radioactive source placed at various distances upstream of the center of
GRETINA. The measured efficiency is shown with blue circles and error
bars and the efficiency of GRETINA in the GEANT4 simulation is shown
as red saltire (×) symbols.
. . . . . . . . . . . . . . . . . . . . . . . . .

Figure 3.15: The TRIPLEX device used in this experiment. The schematic shows (A)
the outer casing, (B) a piezoelectric motor, (C) the outermost tube, (D)
the middle stationary tube, (E) the innermost tube, (F) the target posi-
tioned by the innermost tube E, (G) the first degrader held stationary by
the middle tube, and (H) the second degrader positioned by the outermost
. . . . . . . . . . . . . . . . . . . . .
tube. The figure is from Ref. [81].

Figure 3.16: A schematic of a motor used to drive a tube and foil in the TRIPLEX
. . . . . . . . . . . . . . . . . . . .

device. The image is from Ref. [86].

Figure 3.17: A cartoon depicting the operation of the linear actuators used within the
. . . . . . .

motor of the TRIPLEX device. The image is from Ref. [86].

Figure 3.18: A diagram showing the connections between the TRIPLEX systems and
the PC. The PC uses the multi-channel analyzer (MCA) GUI, the Motor
GUI, and the Micrometer GUI to access the three systems represented
here. The Pulser sends a trigger to the MCA and a pulse to Degrader 1
which causes induced pulse responses on Target and Degrader 2 due to
the capacitance with these foils. NIM modules are used to manipulate
the trigger signal and the induced pulse responses. A motor controller
operates and reads the displacement value from each motor. The TESA
controller reads the measurements from the TESA micrometers.
. . . .

Figure 3.19: The inverse of the induced voltage on the target foil at various separations
from the first degrader. The separation is measured relative to the location
where the induced voltage saturated, indicating electric contact between
the foils. The change in separation is measured by the motor displacement.
The red dots are data points which are fit with a linear function shown
by a black dotted line. The fit matches the data well for separation <
150 µm but deviates somewhat at larger separation. At large separation
the assumption that the foils behave like a parallel-plate capacitor breaks
down and the induced voltage value changes by small degrees that are
difficult to resolve.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 3.20: The setup for the part of the experiment with the upstream target, de-
2 isomer of 32Mg. The 9Be(34Si,32Mg)X reaction
2 isomer which emits a cascade of gamma rays γ1 and γ2
. . . . . . . . . . . . .

signed to measure the 0+
populates the 0+
at angles θ1 and θ2 relative to the ion trajectory.

64

66

68

69

71

73

79

xiii

Figure 3.21: The ata, bta, dta and yta distributions of the 34Si secondary beam. Data
.

is shown as a black line and a scaled simulation is shown as a red line.

Figure 3.22: The ata, bta, dta and yta distributions of the 32Mg reaction product with
all three foils installed in the TRIPLEX device. Data is shown as a black
line and a scaled simulation is shown as a red line.
. . . . . . . . . . . .

Figure 4.1: The particle identification (PID) for the 34Si secondary beam. The x-axis
is the TOF in arbitrary units between the E1 and OBJ scintillators of
the S800 spectrograph. The y-axis is the TOF in arbitrary units between
the E1 scintillator of the S800 and the XFP scintillator of the A1900
fragment separator. Both time values were digitized with the Mesytec
multi-hit TDC. The software gate used to select 34Si ions is shown as a
solid black line.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Figure 4.2: The PID for reaction products resulting from the 34Si secondary beam
and the 9Be target in the target-only setting. The x-axis is the TOF in
arbitrary units between the E1 and OBJ scintillators of the S800 spectro-
graph. The y-axis is the energy loss of ions traveling through the ionization
chamber in arbitrary units. The software gates used to select the 32Mg
and 30Mg reaction products are shown with solid black lines. . . . . . . .

Figure 4.3: The ata, bta, dta and yta distributions of 32Mg. Data is shown as a solid
. . . . .

black line and a scaled simulation is shown as a dashed red line.

Figure 4.4: Doppler-corrected energy spectra showing four transitions in 32Mg corre-
sponding to prompt decays observed in this experiment. Data is shown
with points and error bars while the simulation is shown with a solid red
line.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

83

84

90

91

92

93

Figure 4.5: Partial level scheme of 32Mg.

. . . . . . . . . . . . . . . . . . . . . . . .

94

Figure 4.6: The PID for reaction products resulting from the 34Si secondary beam and
the 9Be target in the three-foil settings. The x-axis is the TOF in arbitrary
units between the E1 and OBJ scintillators of the S800 spectrograph. The
y-axis is the energy loss of ions traveling through the ionization chamber
in arbitrary units. The software gates used to select the 32Mg, 30Mg, and
20O reaction products are shown with solid black lines.
. . . . . . . . . .

96

xiv

Figure 4.7: Doppler-corrected energy spectra showing the two components of the 885-
keV peak in the large separation setting. Only gamma rays observed in
the laboratory frame with an emission angle of θ < 70◦ are shown. The
solid red line depicts the result of a G4Lifetime simulation that assumes a
ratio of r(2+
1 ) = 3.2 for the number of reactions on the target divided by
the number of reactions on the first degrader that populate the 2+
1 state
of 32Mg. The dashed gray line is the background contribution from the
neutron-induced 1014-keV gamma ray from 27Al and it is scaled by ×5 in
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
this image.

Figure 4.8: χ2 distribution for the ratio of the number of reaction on the target to
1 state of 32Mg. The χ2 statistic
1 ) = 3.2(6). Another local minimum
. . . . . . . . . . . . . . . . . . . . . . . .

those on the first degrader for the 2+
obtains an absolute minimum at r(2+
can is present at r(2+

1 ) = 2.6.

97

99

Figure 4.9: Doppler-corrected energy spectra showing the lineshape of the 1437-keV
peak in the large separation setting. Only gamma rays observed in the
laboratory frame with an emission angle of θ < 70◦ are shown. The solid
red line depicts the result of a G4Lifetime simulation that assumes a ratio
of r(4+
1 ) = 0.9 for the number of reactions on the target divided by the
number of reactions on the first degrader that populate the 4+
1 state of
32Mg. The dashed gray line assumes a ratio of r(4+
. . . . . . . 101

1 ) = 1.6.

Figure 4.10: χ2 distribution for the ratio of the number of reaction on the target to
1 state of 32Mg. A clear minimum is

those on the first degrader for the 4+
visible at r(4+

. . . . . . . . . . . . . . . . . . . . . . . . . . 102

1 ) = 0.9(2).

Figure 4.11: Doppler-corrected energy spectra showing the two components of the 885-
keV peak. A gate is placed on the gamma-ray emission angle θ < 50◦.
The red lines depict the results of G4Lifetime simulations that assume a
lifetime of 19 ps for the 2+

. . . . . . . . . . . . . . . . . 103

1 state of 32Mg.

Figure 4.12: Doppler-corrected energy spectra showing the reduced and slow compo-
nents of the 885-keV transition in the 0.7-mm setting. The red solid line
depicts the results of G4LifeTime simulation that assumes a lifetime of
28 ps for the 2+

. . . . . . . . . . . . . . . . . . . . . . . 105

1 state of 32Mg.

Figure 4.13: The best-fit lifetime for each of the four short-separation settings is shown
1 ) =

with the points and error bars. The weighted-average lifetime of τ (2+
18.7 ± 1.2(stat.) ps is shown as a solid red line.

. . . . . . . . . . . . . . 106

xv

Figure 4.14: Doppler-corrected energy spectra showing the lineshape of the 1437-keV
transition. The solid red line depicts the results of GEANT4 simulations
that assume a lifetime of 0.9 ps for the 2+
1 state of 32Mg, while the gray
. . . . . . . . . . . . . . . . . . 107
dashed line assumes a lifetime of 1.6 ps.

Figure 4.15: Partial level scheme of 30Mg.

. . . . . . . . . . . . . . . . . . . . . . . . 108

Figure 4.16: Doppler-corrected energy spectra of gamma rays from 30Mg in the three-
foil, large-separation setting. The red solid line depicts the results of
G4Lifetime simulations with the best-fit intensity of the gamma rays. No
gate was placed on the emission angle θ in order to increase the statistics
of the observed peaks.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 109

Figure 4.17: Doppler-corrected energy spectra of gamma rays from 30Mg in the three-
foil, short-separation setting. A gate was placed on the gamma-ray emis-
sion angle θ < 50◦. The solid red line depicts the simulation results
assuming a lifetime of 2.3 ps. The dashed gray line corresponds to a
simulation with a lifetime of 2.9 ps.

. . . . . . . . . . . . . . . . . . . . 110

Figure 4.18: Partial level scheme of 20O.

. . . . . . . . . . . . . . . . . . . . . . . . . 111

Figure 4.19: The ata, bta, dta and yta distributions of the 20O reaction product in
the three-foil settings. Data is shown as a solid black line and a scaled
simulation is shown as a dashed red line. . . . . . . . . . . . . . . . . . . 112

Figure 4.20: Doppler-corrected energy spectra of gamma rays from 20O in the three-
foil, large-separation setting. The data shows contributions at 1674-keV
from the decay of the 2+
1 state. The
contributions of the populations of the 2+
1 state are esti-
mated with the simulated responses depicted with the dashed green and
the dotted purple lines respectively. The solid gray line is an exponen-
tial background and the solid red line is the sum of the two simulation
components and the background.

1 state and at 1898-keV from the 4+

1 state and the 4+

. . . . . . . . . . . . . . . . . . . . . . 113

Figure 4.21: Doppler-corrected energy spectra of gamma rays from 20O in the three-
foil, short-separation setting. A gate has been placed on the gamma-ray
emission angle θ < 50◦. The dashed green line and the solid red line
are from G4Lifetime simulations assuming a lifetime of 100 and 30 ps
respectively for the 4+
1 state in 20O. The solid gray line is an exponential
background.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 114

xvi

Figure 4.22: Doppler-corrected energy spectra of gamma rays from 20O in each of the
the three-foil, short-separation settings. A gate has been placed on the
gamma-ray emission angle θ < 50◦. The solid red lines are the results of a
G4Lifetime simulation assuming a lifetime of τ (2+
1 ) = 16 ps and no feeding
from the 4+
1 state. The solid gray lines are exponential backgrounds fit to
a large energy region.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . 115

Figure 4.23: Doppler-corrected gamma-ray energy spectra are shown for data (blue cir-
cles), background from the scaled response of reaction products excluding
the Mg isotopes (black lines), and the sums of the simulated and back-
ground responses (red lines). 31Mg is shown (a) in the upper spectra and
the constraint that the 244-keV transition only interacts within r = 20 mm
of its first interaction is included in the lower, filled spectrum (scaled by
0.5). 32Mg is shown (b) with the r = 20 mm gate applied to the 170-keV
gamma ray in the upper spectrum (scaled by 0.6) and with additional
gates requiring proper identification of the interaction points of the 885-
keV gamma ray in the lower, filled spectrum. Insets show the number of
decays near (250 to 525 mm), at mid-distance (525 to 700 mm) and far
from the target (700 to 1000 mm). The data (open circles) is compared to
simulations assuming lifetimes of 5 ns (red solid line), 15 ns (blue dashed
line), and 45 ns (black dotted line). . . . . . . . . . . . . . . . . . . . . . 118

Figure 4.24: Possible values for partial cross section and lifetime of the 0+

2 state of
32Mg within 1 σ are plotted (gray band). From the work of Ref. [60], the
cross section upper limit of 0.10 mb is included (red line), and the overlap
of that work and this work is highlighted (red hatched region).

. . . . . 122

Figure 4.25: Distribution of the angle af p and the position xf p for 32Mg reaction prod-
ucts in the target-only setting with the TRIPLEX device. A portion of
the 32Mg reaction products are lost at the edges of the distribution with
extreme values of both af p and xf p.

. . . . . . . . . . . . . . . . . . . . 127

xvii

Figure 4.26: Distribution of the angle af p for 32Mg reaction products in the target-only
setting with different xf p ranges. In the top spectrum, central values are
chosen (−50 mm < xf p < 50 mm) for which all angles af p are accepted
by the S800 spectrograph. A solid red line represents a Gaussian fit to
the distribution with the height A, mean µ, and width σ as free parame-
ters. The bottom spectrum shows the af p distribution for for non-central
position (−300 mm < xf p < −150 mm). The red solid line in the bottom
spectrum represents a Gaussian fit with only the height A as a free pa-
rameter and the mean µ and width σ fixed to the values obtained from the
fit to the distribution in the top spectrum. The gap between the solid red
line and the data in the bottom spectrum represents the reaction products
that were lost due to limited af p acceptance of the S800 for this range of
xf p values.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 128

Figure 4.27: The parallel-momentum distribution of 32Mg reaction products in the
target-only setting. The counts are corrected for the loss of angular accep-
tance which occurs mostly at small and large momenta. The solid red line
depicts a skew-Gaussian fit [103] to the data from 10.32 to 10.84 GeV/c.
The difference between the fit and the data at the edges of the momentum
distribution represents the loss of 32Mg reaction products due to the finite
momentum acceptance of the S800 spectrograph.

. . . . . . . . . . . . . 129

Figure 5.1: The B(E2; 2+

1 → 0+

1 ) value in 32Mg reported in experiments over the
past 25 years. The present RDM measurement is shown with a red square.
The Coulomb-excitation measurements from Refs. [13,37,38,40] are shown
with open circles. The open triangles are the results from Refs. [36,39].
The upward-pointing triangles reflect the B(E2) values without feeding
corrections while the corresponding downward-pointing triangles take into
account feeding corrections. The saltire symbol (×) represents the fast
timing measurement of Ref. [41].

. . . . . . . . . . . . . . . . . . . . . . 132

Figure 5.2: The B(E2; 2+

1 → 0+

1 ) values in the Mg isotopic chain from 28Mg to 40Mg.
The 32Mg result from this experiment is shown with a filled black cir-
cle. The experimental values from other results are shown with black
saltires (×) and are from Refs. [104,96,105,106]. Theoretical calculations
are also shown for SDPF-M [32] (red line and open diamonds), SDPF-U-
MIX [21] (blue line and open inverted triangles), AMPGCM [34,35] (pur-
ple line and open circles), EKK [22] (green line and open squares), and
USDA [107] (brown line and open triangles).

. . . . . . . . . . . . . . . 133

Figure 5.3: The 0+, 2+, and 4+ states and B(E2) values in 32Mg from the SDPF-U-
MIX calculation with fixed npnh configuration. Figure is from Ref. [21].
The numbers between the states represent the B(E2) value of that tran-
sition in e2fm4.

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 136

xviii

Chapter 1

Introduction

1.1 The Atomic Nucleus

A nucleus sits at the core of every atom. The atom has a radius of 10−10 m, while the
nucleus is merely 10−14 to 10−15 m in radius. However, the atomic nucleus is very dense

and contains 99% of the atom’s mass. The atomic nucleus is composed of protons and

neutrons, together called nucleons, which are bound together by the strong nuclear force.

The number of nucleons in a nucleus can vary from a few to a couple of hundred, although

not all combinations of proton and neutron number will form a bound system. The number

of protons Z determines the element of the nucleus and different isotopes of a given element

are determined by the number of neutrons N . The total number of nucleons, also known

as the mass number, is A = Z + N . Commonly, a particular nuclide, the nuclear species

determined by a specific proton and neutron number, is designated with the notation AZ

where Z is written as the elemental abbreviation. For example, the nuclide with six protons

Z = 6 and seven neutrons N = 7 is the element carbon and would be written as 13C.

Thousands of different nuclides are known to exist and can be depicted in a Segr´e chart

such as the chart shown in Fig. 1.1 [1]. Each nuclide is represented by a square and arranged

with increasing N along the x-axis and increasing Z along the y-axis. Figure 1.1 labels the

isotopes by their mode of radioactive decay. The variety of radioactive decay modes is but

1

Figure 1.1: A Segr´e chart depicting the known nuclides. The nuclides are arranged by
increasing N along the x-axis and increasing Z along the y-axis. The decay mode of each
nuclide is represented by the color. The blue and black narrow rectangles represent the
proton and neutron magic numbers. Magic number 2 is not labeled in this figure. The figure
is adapted from Ref. [1]

one example of the breadth of phenomena that exists in nuclear structure.

Nucleons are not point particles, but have internal structure made up of quarks. The

proton is composed of two up quarks and one down quark while the neutron has one up quark

and two down quarks. The sum of the charges of these quarks, +2/3e for each up quark and
−1/3e for each down quark, results in the net charges of the nucleons: +e for protons, 0 for

neutrons. The strong interaction confines the quarks together within the nucleon while the

weak interaction is responsible for the beta decay process which is the most common decay

mode for unstable nuclides. At the scale of the nucleus, instead of dealing with the strong

and weak interactions directly, it is more practical to use the effective nuclear force.

The exact form of the nuclear force is not known. At very short distances, less than the

2

radius of a nucleon which is roughly r = 10−15 m, the nuclear force is strongly repulsive. The
nuclear force between nucleons is strongly attractive for distances between about r = 10−15 m
and r = 2 × 10−15 m. At larger distances the strength of the attractive nuclear force dies

rapidly away and approaches zero.

Due to the positive charge of the protons, the Coulomb interaction plays an important

role in the nuclear system. The Coulomb repulsion felt by protons in a typical nucleus is

smaller than the net nuclear attraction, so the protons remain bound. Figure 1.1 shows that

the smallest stable nuclides have roughly equal numbers of protons and neutrons. However

for larger stable nuclides, the nucleus contains more neutrons than protons. This is a result

of the Coulomb interaction causing the protons to be less bound than the neutrons.

Each nucleon is a spin-1/2 particle and has orbital angular momentum l. The intrinsic

spin and orbital angular momentum of a nucleon couple to the total angular momentum j.

The angular momenta of all nucleons add vectorially to the total angular momentum of the

nucleus, J or I, often called the spin of the nucleus. The total energy of the nucleus is the

sum of the energy contributions of the nucleons including nuclear correlations. Many states

of a given nucleus are possible, corresponding to different combinations of the nucleons’

single-particle states. The lowest-energy state of a nucleus is referred to as the ground state.

The higher-energy states are usually referred to by their excitation energy Ex relative to the

ground state. The spins and excitation energies of nuclear states are crucial observables that

can assist our understanding of nuclear phenomena and the underlying nuclear interaction.

3

1.2 The Shell Model

Abundant experimental evidence suggests that at particular numbers of protons or neutrons

certain properties emerge. Figure 1.2 shows the excitation energy of the the first excited

states in even-even nuclides [2]. The figure shows that at particular numbers of protons or

Figure 1.2: The energy of the first excited states in even-even nuclides. The large excitation
energies at proton or neutron number of 8, 20, 28, 40, 50, 82, and 126 are evidence for closed
shells at these magic numbers. The figure is adapted from Ref. [2].

neutrons the energy of the first excited state is dramatically larger than for surrounding

nuclides. These numbers are called magic numbers and they are 2, 8, 20, 28, 50, 82, and

126 (2 is less evident from Fig. 1.2 but other evidence supports its magicity). Large energy

gaps between the single-particle states for individual nucleons would require large energy to

excite a nucleon to the next available state. This explains the increase in excitation energy

that is observed at magic numbers of nucleons. What is fascinating is the regularity with

which the shell gaps emerge for many nuclides with a magic number of protons or neutrons.

4

The phenomenon of magic numbers can be explained with the nuclear shell model. The

nuclear shell model is similar to the atomic shell model which describes the behavior of

atomic electrons surrounding a nucleus. These shell models assume independent particle

motion for the particles involved, and they assume that each particle is affected by a central

potential, V (r). Nucleons are densely bound together so it is not obvious that they could

be independent of one another. However, nucleons obey the Pauli exclusion rule, meaning

that no two nucleons can share the same quantum numbers. This implies that very little

space in the nucleus is shared by the nucleons so they do not often collide [3]. Each nucleon

is affected by the strong force caused by each other nucleon. While the nuclear interaction

has a rich structure and is not simple a central force, the net force felt by a nucleon is

sufficiently centralized for the shell model description to be valid [3]. Protons and neutrons

are distinguishable particles so they do not Pauli-block each other from filling single-particle

levels. Thus, there will be separate shells for both protons and neutrons. However, the

nuclear force is nearly the same for both protons and neutrons [2] and one result of this is

that protons and neutrons have the same set of magic numbers. The Coulomb repulsion

between protons causes a shift in their energy levels, but no change in the resulting magic

numbers.

The three-dimensional harmonic oscillator potential is a good starting point for param-

eterizing the effective one-nucleon potential of the nuclear shell model. The harmonic oscil-

lator potential has the form

VHO(r) =

mω2r2

1
2

(1.1)

5

and the nucleons occupy single-particle levels with energy


 = (N + 3/2)¯hω

(1.2)

where the N is the harmonic oscillator quantum number and N ≥ 0. Each energy level has

an occupancy of

DN = (N + 1)(N + 2).

(1.3)

The levels of the harmonic oscillator potential are shown in the left column of Fig. 1.3.

When the number of nucleons in a level meets the occupancy, that level is filled and the

next nucleon must occupy a level that is higher in energy by ¯hω. This causes a natural

gap between the energies of a closed-shell system and the system with one more nucleon.

However, for the harmonic oscillator these gaps are at nucleon numbers 2, 8, 20, 40, 70,

112, and 168 as shown in the left column of Fig. 1.3. The harmonic oscillator alone is not

sufficient to reproduce all the magic numbers greater than 20.

A more appropriate central potential choice is the Woods-Saxon potential [4] depicted in

Fig. 1.4. The Woods-Saxon central potential has the form

VW S(r) =

V0

1 + e(r−R0)/a0

(1.4)

where V0 is the depth of the central potential, R0 is the radius, and a0 is the diffuseness which

controls how suddenly the potential well returns to zero [2]. One advantage of the Woods-

Saxon potential is that the center of the potential is flatter than the harmonic oscillator. This

properly accounts for the saturation of the nuclear density [3]. The Woods-Saxon potential

also goes to zero at large r, while the harmonic oscillator increases to infinity. The trend

6

Figure 1.3: The single-particle orbits for neutrons in the 208Pb nucleus, calculated for the
harmonic oscillator potential, Woods-Saxon potential, and Woods-Saxon potential with spin-
orbit term. The figure is adapted from Ref. [2]. The number in brackets is the occupancy
of that orbit and is followed by the cumulative occupancy in that orbit and all lower-energy
orbits. The orbits are labeled by the harmonic oscillator quantum number N , the orbital
angular momentum quantum number l = s, p, d, f, g, h, i meaning l = 0, 1, 2, 3, 4, 5, 6, and
twice the total angular momentum quantum number 2j.

7

Figure 1.4: The forms of the central potentials discussed in the text. The harmonic oscillator
potential is shown as a blue dashed line and the Woods-Saxon potential is shown as a solid
red line. The Woods-Saxon potential is shifted vertically by one energy unit to align with
the minimum of the harmonic oscillator potential.

toward zero more accurately reflects the fact that the nuclear force is a short range effect

which dies out at large distances. The shape of the Woods-Saxon potential causes a splitting

of the energy levels compared to the harmonic oscillator. The nucleons with larger orbital

angular momentum l are lower in energy than states with lower l in the same harmonic

oscillator N shell. In Fig. 1.3, the orbital angular momentum of a state is labeled according

to the convention used in spectroscopic notation, s = 0, p = 1, d = 2, f = 3, g = 4, h = 5,

i = 6, etc.

The final adjustment to the central potential that results in the proper magic numbers

is the inclusion of a spin-orbit term [5, 6]. This term depends on the coupling of intrinsic

spin s and orbital angular momentum l and causes the state with s oriented parallel to l

to be more bound than the antiparallel arrangement. For example, in Fig. 1.3 the 0p3/2

orbit is lower than the 0p1/2 for the Woods-Saxon plus spin-orbit potential. This effect

is large enough that an orbital from the next higher harmonic oscillator quantum number

N + 1 can be lowered to intrude on a lower harmonic oscillator level N . One example is the

0g9/2 state which has harmonic oscillator quantum number N = 4 but lies close in energy

8

to the N = 3 states for the Woods-Saxon plus spin-orbit potential in Fig. 1.3. This effect

results in the low-energy negative-parity states seen in many medium-mass and heavy nuclei

since a nucleon can occupy a state with parity opposite that of the other states available at

low energy. With the spin-orbit term, the observed magic numbers are now reproduced, as

indicated by the large numbers in the gaps between levels in the Woods-Saxon plus spin-orbit

potential in Fig. 1.3.

The shell model is a valuable theoretical tool that can explain many properties of nuclides

at and near closed shells. As already discussed, the shell model explains the large excitation

energy of the first excited state for magic nuclides. For a nucleus with one more or one fewer

nucleon than a doubly closed-shell nucleus, the shell model can suggest a straightforward

interpretation. For example, 49Sc has Z = 21 and N = 28, one more proton than 48Ca.

The 49Sc ground state can be imagined as an inert 48Ca core with a single valence proton in

the next available proton orbital, which is 0f7/2. This model predicts that the ground-state
spin and parity is 7/2− (− parity corresponding to the odd value for l = f = 3), which

turns out to reproduce observations. For full calculations, the shell model must account for

the energies of the inert core, the valence nucleons, the interactions between each of the

nucleons and the core, and the interactions among the valence nucleons. This is no small

task, but leads to accurate results and is computationally much simpler compared to treating

the interactions among all A nucleons in a nucleus.

1.3 Nuclear Collectivity

Far from closed shells the number of nucleons outside of a closed shell becomes large. In

these regions it becomes impractical to describe the behavior of a nucleus with one or a few

9

nucleons outside of an inert core which causes shell model approaches to struggle. Instead,

there are collective models that describe the nucleons with the collective motion of many

nucleons [7].

One collective mode in nuclear systems is vibrational motion, in which the nucleons

vibrate relative to a spherical shape. The vibrational modes are often described as phonons

which have a fixed energy and spin 2. A standard vibrational band of states is built on the

0+ ground state. The first vibrational state has one phonon, so by coupling to the ground

state it can produce only a 2+ state. The two-phonon vibrational mode couples a second

phonon to the first and the possible states from the angular momentum coupling are 0+, 2+,

and 4+. The three-phonon mode again couples the angular momentum of the third phonon

to any of the two-phonon states which results in states 0+, 2+, 3+, 4+, and 6+ [3]. Since

each additional phonon adds the same energy, the two-phonon states are at about twice the

excitation energy of the one-phonon state, and the three-phonon states have about thrice

the excitation energy of the one-phonon state. This pattern continues, with the n-phonon

states at an excitation energy of about

En = nE(2+
1 ).

(1.5)

The regular spacing of these excited state energies is a clear signature of vibrational behavior.

A rigidly deformed nucleus will rotate about an axis perpendicular to the axis of symme-

try. This rotational behavior is another collective phenomenon that exists in many nuclides.

The energy of a rotational band built on a 0+ ground state increases with spin,

Erot(J) = J(J + 1)

¯h2
2I

10

(1.6)

where I is the moment of inertia of the nucleus and the spin J is even [7]. The states of a

typical rotational band are spaced with increasing energy separation at increasing spin. The

shape of a nucleus can be characterized by the quadrupole deformation β and the asymmetry

of the deformation γ. β is positive and represents the size of the deformation along the axis
of symmetry, with β = 0 representing a spherical nucleus. For β > 0 and γ = 0◦, the nucleus

has a prolate shape, that is, the nucleus is deformed but maintains an axis of symmetry and

it is longer along the axis of symmetry than it is for a perpendicular axis (for instance, like
an American football). For β > 0 and γ = 60◦, the shape is oblate, where the nucleus is

shorter along the symmetry axis than a perpendicular axis (similar to the shape of a curling
stone). For values of γ between 0◦ and 60◦, the nucleus has no axis of symmetry and is

therefore triaxially deformed.

The simple pictures of vibrational and rotational nuclei are often insufficient to describe

the more complicated cases that arise in real nuclei. The collective modes can be coupled to

an excited single-particle state, for instance. The vibrational and rotational modes can be

coupled together as well, and the deformation of a rotating nucleus may not be rigid, but

changing throughout the rotational band. Nevertheless, the systematic changes in energy

of vibrational and rotational states provide a paradigm to which we can compare observed

states. The ratio of the excited energies of the 4+ and the 2+ state E(4+)/E(2+) is often

used. This ratio is E(4+)/E(2+) = 2.0 for vibrational bands from Eq. 1.5. For rotational

bands, the ratio is E(4+)/E(2+) = 3.33 from Eq. 1.6.

11

1.4 Electromagnetic Transitions

Not only are the states of a nucleus interesting, but the transitions between two nuclear states

hold enlightening physical information as well. The atomic nucleus can undergo spontaneous

radioactive decay from one state to another. The random decay of a nuclear state can be

characterized by its lifetime. The lifetime τ of a state is proportional to the state’s half-life,

τ =

t1/2
Ln(2)

(1.7)

where the half-life t1/2 is the time for half of a sample of N nuclei in the state of interest to

decay. Radioactive decay reduces the energy of the nucleus while emitting radiation. The

inverse process causes the nucleus to transition from a lower-energy state to a higher-energy

state in the presence of either real or virtual radiation. An example of this is a Coulomb-

excitation reaction where the nucleus is excited to a higher-energy state due to the electric

field of another nucleus.

Most of the transitions between states of the same nucleus are electromagnetic transitions

mediated by the exchange of a photon called a gamma ray [2]. Photons are spin-1 bosons

that carry energy Eγ proportional to their frequency ν. A nucleus in initial state i which

decays to a final state f will emit a gamma ray with angular momentum λ such that the

angular momentum of the system is conserved. This requires

|Jf − Ji| ≤ λ ≤ Jf + Ji

(1.8)

where Ji and Jf are the spins of the initial and final nuclear states i and f . Since photons

are spin-1, transitions between two nuclear states of spin Ji = Jf = 0 cannot occur via

12

gamma decay. Instead the transition occurs via a different decay process such as electron

capture or pair production.

The strong interaction conserves parity so all nuclear states have a well-defined parity
π = ±1 [2]. For gamma decay between states with good parity, the gamma ray also has a
well-defined character which is either electric (E) when πf πi(−1)l = 1 or magnetic (M ) when
πf πi(−1)l = −1. Transitions between states are often referred to by their electromagnetic

multipolarity πλ, for instance E0, E1, M 1, E2, etc.

When decays of more than one multipolarity compete between two states, the multipo-

larity with the lower λ usually dominates. Formally, the decay rate between an initial state
|ψi > and a final state |ψf > is

(cid:18) 8π(λ + 1)

(cid:88)

(cid:19)(cid:18) k2λ+1

(cid:19)| < ψf|| ˆO(πλ)||ψi > |2

W =

λ[(2λ + 1)!!]2

¯h

π,λ

(2Ji + 1)

(1.9)

where | < ψf|| ˆO(πλ)||ψi > | is a reduced matrix element of the one-body operator ˆO , π is

E or M , and λ is the angular momentum of the transition [2]. The quantity k is the wave

number of the gamma ray, k = Eγ/¯hc. From Eq. 1.9 the reduced transition rate is defined

as

B(πλ) =

| < ψf||O(πλ)||ψi > |2

(2Ji + 1)

.

(1.10)

The reduced transition rate contains interesting physical properties of the nuclear system:
the wavefunctions |ψi > and |ψf >, and the transition operator O(πλ). Reduced transition

rates are often used to compare the observed properties of a nuclear transition with either

theoretical calculations or with other nuclides. Since the reduced transition rate depends on

Ji, the reduced transition rate for the decay of state i to state f can be different from the

13

transition from f to i and requires the following conversion,

B(f → i) =

2Ji + 1
2Jf + 1

B(i → f )

(1.11)

To determine the reduced transition rate, the experimenter must measure the gamma-ray

energy Eγ, the multipolarity πλ, and the half-life t1/2 which is inversely proportional to the

decay rate. For the most common electromagnetic transitions the following equations are

often used as a shorthand [2]:

B(E1) =

B(M 1) =

0.435
γt1/2,p

E3

39.4
γt1/2,p

E3

e2fm2MeV3fs,

µ2
N fm2MeV3fs,

B(E2) =

564
γt1/2,p

E5

e2fm4MeV5ps.

(1.12)

(1.13)

(1.14)

and

The subscript p signifies that t1/2,p is the partial half-life of the state. The quantity t1/2,p =

t1/2/b for cases in which a state may decay via multiple branches where b is the branching

ratio. Reduced transition rates are often compared with the Weisskopf estimate [8]. The

Weisskopf estimate for a reduced transition rate is based on the transition of a single nucleon.

For measured reduced transition rates that are much larger than the Weisskopf estimate, the

transition is considered to involve the collective motion of many nucleons. The Weisskopf

(cid:18) 1

(cid:19)(cid:20)

(cid:21)2

3

4π

(3 + λ)

14

estimate is

BW (Eλ) =

(1.2A1/3)2λe2fm2λ

(1.15)

for electric transitions, and

BW (M λ) =

for magnetic transitions.

(cid:18)10

(cid:19)(cid:20)

3

(cid:21)2

π

(3 + λ)

(1.2A1/3)2λ−2µ2

N fm2λ−2

(1.16)

Reduced transition rates can probe the nuclear structure effects that have been discussed

so far such as shell closures and collective modes. The B(E2) value for the 0+

in even-even nuclei shows the existence of closed shells. Figure 1.5 shows the B(E2; 0+

1 transition
1 → 2+
1 )
value for argon isotopes from 32Ar to 48Ar [9]. The B(E2) value is minimized for the closed

1 to 2+

Figure 1.5: B(E2; 0+
1 ) values for argon isotopes, adapted from Ref. [9] . The B(E2)
value is minimized for isotopes with a conventional magic number of neutrons, N = 20 or
N = 28.

1 → 2+

shells of neutrons at N = 20 and N = 28. For these magic nuclides, neutron excitation

requires the breaking of a closed shell so only proton excitations dominate the 0+

1 → 2+

1

transition. As a result, the transition rate is hindered. Between the closed shells, the B(E2)

value is maximized (see Fig. 1.5). Thus, a large B(E2) value is a signature of collective
behavior away from closed-shell structures. The ratio of B(E2) values for the 4+ → 2+
transition and the 2+ → 0+ transition also probes the collective nature of the band of
those states. B(E2; 4+ → 2+)/B(E2; 2+ → 0+) = 2.0 is expected for a vibrational band

15

and B(E2; 4+ → 2+)/B(E2; 2+ → 0+) = 1.43 is expected for a symmetrically-deformed

rotational band [3].

1.5

Islands of Inversion

The magic numbers were first introduced to explain observations about nuclei near stabil-

ity [10]. As experiments have expanded the library of nuclear physics data to more unstable

nuclides, it has been seen that the magic numbers do not remain static for all nuclides [11].

Near stability, the magic numbers indicate closed shell which support spherical shapes, en-

hanced stability, and first excited states with large energies and low transition rates in even-

even nuclides. Regions with a conventional magic number where the nuclides do not exhibit

the effects of a shell closure but instead behave more like mid-shell collective nuclides are

called islands of inversion. Islands of inversion have been proposed in neutron-rich regions

for many magic numbers such as N = 8, 20, 28, 40, [12–19] and more recently at 50 [20]. Gen-

erally, the islands of inversion occur when the large shell gap that exists for magic nuclides

near stability is suddenly reduced or is overcome by proton-neutron correlations in nuclides

with extreme excesses of protons or neutrons. The physical mechanisms underlying these

changes have been suggested to be the tensor component of the proton-neutron interaction

and the three-body effective forces [11, 21–23].

The existence of an island of inversion was first postulated to explain the unexpected

excess binding energy for neutron-rich isotopes near the N = 20 magic number [24–26]. The

evidence for the N = 20 island of inversion can also be seen in the energy trend of 2+

1 excited

states in even-even nuclides as shown in Fig. 1.6. For N = 20 nuclides close to stability, the

2+
1 energy is larger compared to the neighboring even-even isotopes. This trend is broken

16

Figure 1.6: The energy of the first 2+ states in even-even nuclides. For many nuclides with
the magic number N = 20 or N = 28 the first 2+ energy is larger than the neighboring
even-even isotopes. This is not the case for 32Mg at N = 20 and for 40Mg and 42Si at
N = 28.

for 32Mg, the most neutron-rich N = 20 isotone shown in Fig. 1.6. The breakdown of the

N = 28 magic number can also be seen in this plot at 40Mg and 42Si.

The conventional N = 20 magic number is a result of the large shell gap between the

neutron sd shell and the neutron pf shell. In this region it is useful to consider the number of

neutrons that occupy levels in either the sd shell or the pf shell. The shell gap supports the

normal neutron configuration in which zero neutrons are promoted to the pf shell, leaving

zero holes in the sd shell. This is also called the zero-particle-zero-hole or 0p0h configuration.

The intruder configurations are those in which pairs of neutrons are excited from the sd shell

to the pf shell. These are the two-particle-two-hole (2p2h) configurations, four-particle-four-

hole (4p4h) configurations, etc. These intruder configurations are not energetically favored

17

near stability due to the large energy needed to cross the sd-pf shell gap. Moving from

34Si to 32Mg, two protons are removed from the proton d5/2 orbital. This causes a sudden
reduction in the contribution from the proton-neutron πd5/2− νd3/2 interaction, reducing in

the sd-pf shell gap [11]. This allows intruder configurations to become energetically favored

in 32Mg.

1.6 Collectivity in 32Mg

The 32Mg nuclide has Z = 12 and N = 20. According to the magic numbers near stability,

32Mg should be magic as well. However, 32Mg is neutron rich and within the N = 20

island of inversion which was first identified from the mass measurements of neutron-rich Na

isotopes [24]. Evidence for the inclusion of 32Mg in the island of inversion also came from a

mass measurement [27] which noted the same signature of excess binding energy as observed

in the neutron-rich Na isotopes. Additionally, the low energy of the 2+

1 state of 32Mg at

885 keV suggested a deformed shape [15].

The unexpected binding energy in the N = 20 island of inversion was reproduced in

Hartree-Fock calculations which suggested the deformation in this region was due to the

filling of negative parity orbitals from the neutron 0f7/2 level [28]. Shell model calculations

performed in only the sd shell could not reproduce the experimental results for the neutron-

rich Na and Mg isotopes [26]. By including orbitals from both the sd shell and the pf shell,

the shell model calculations were more successful [25, 29–31]. Shell model calculations using

the sd and pf shells have been extended to make predictions across the the entire N = 20

island of inversion [21, 22, 32, 33]. Beyond mean field calculations have successfully described

this region as well [34, 35].

18

Several experiments have probed the possible deformation of 32Mg by measuring the re-

duced E2 transition rate between the 0+

1 and 2+

1 states. An intermediate-energy Coulomb

excitation reaction was used to make the first measurement which found the result of

B(E2; 0+

1 → 2+

1 ) = 454(78) e2fm4 [13]. This result was consistent with the available shell

model calculations that include both sd and pf shells. The measurement was apparently

confirmed by Ref. [36] which found B(E2; 0+

1 → 2+

1 ) = 440(55) e2fm4 and interpreted 32Mg

as being dominated by 2¯hω configurations, i.e., configurations with two neutrons across the

sd-pf shell gap. Shortly after, two additional measurements were made of the reduced E2

transition rate, one reporting a larger value of B(E2; 0+

1 ) = 622(90) e2fm4 [37], and
1 ) = 449(53) e2fm4 [38] in very good agreement with the first

1 → 2+

one reporting B(E2; 0+

1 → 2+

two measurements [13, 36]. More measurements of the reduced transition rate have been

made with Coulomb excitation reactions [39, 40], and a fast-timing lifetime measurement

has also been reported [41] which was used to deduce the B(E2) value. None of these later

results agree with the large B(E2) value from Ref. [37]. The Coulomb excitation results of

Refs. [36, 39] have discussed the role of feeding contributions from higher-lying states and

report a feeding-corrected B(E2; 0+

1 ) value near 350 e2fm4. Despite the variation in
the measurement of the B(E2) value they all agree that the B(E2) is larger in 32Mg than

1 → 2+

in the neighboring even-even 30Mg nuclide, indicating the onset of deformation in 32Mg.

Nevertheless, the large range of the B(E2; 0+

1 → 2+

1 ) results leaves uncertainty in the degree

of the deformation in 32Mg. To understand the deformation in the ground-state band of

32Mg, it is necessary to find a conclusive B(E2; 0+

1 → 2+

1 ) value that sufficiently accounts

for feeding effects. It would be useful to have a model-independent measurement of the 2+
1

lifetime that is precise enough to resolve the different B(E2; 0+

1 → 2+

1 ) results.

The 4+

1 state of 32Mg is also important to consider. The decay of this state was first

19

observed in a beta-decay experiment [42] but it was not until the later work of Ref.

[43]

that assigned it to 32Mg in coincidence with the 885-keV 2+

1 → 0+

1 transition. Later, a

proton inelastic scattering experiment established the 4+

1 assignment [44]. Along with the

reported measurement of the 6+

deformed rotational band [45]. So far no B(E2) value has been reported for the 4+

1 state, the ground-state band has been suggested to be a
1 →
2+
1 transition, but this measurement could provide crucial evidence for the deformed rotor

interpretation of 32Mg. Using the expected ratio of reduced E2 transition rates in a deformed

band (B(E2; 4+

1 → 2+

1 )/B(E2; 2+

1 → 0+

1 ) = 1.43) and a B(E2; 2+

1 → 0+

1 ) estimated to

be 90 e2fm4 from the measurements discussed in the previous paragraph, the symmetric

deformed rotor interpretation suggests B(E2; 4+

1 → 2+

1 ) = 129 e2fm4. This corresponds to

a 4+

1 lifetime estimate of τ = 1.0 ps. Alternatively, for a vibrational band the expected ratio
1 ) = 180 e2fm4 and

1 ) = 2.0, leading to a B(E2; 4+

1 )/B(E2; 2+

1 → 2+

1 → 0+

is B(E2; 4+

1 → 2+

a lifetime of τ = 0.7 ps. Therefore, the lifetime of the 4+

1 state can be used to distinguish

between collective modes in the ground-state band of 32Mg.

The properties of 32Mg discussed so far are consistent with an inversion between the

intruder 2p2h configurations and the normal 0p0h configuration. In this picture, the ground-

state band is dominated by 2p2h configurations and a band built on a higher-lying 0+

state should be dominated by the 0p0h configuration. The 0+

2 state was first observed

ten years ago at 1058 keV in a (t, p) two neutron transfer reaction [14]. This experiment

reconstructed the 32Mg reaction product energy to distinguish the excited state at 1058 keV.

The angular distribution of recoiling protons was used to deduce the spin-parity assignment

of 0+. Gamma rays in coincidence with the population of the 0+

2 state indicated that the 0+
2

state decays to the 2+

1 state with a transition energy of 172 keV. However, the low intensity

of these gamma rays indicated that the 0+

2 state has a long lifetime (τ > 10 ns). Initially,

20

the properties of the 0+

2 state were considered to be additional evidence for the mixing of

the 0p0h and 2p2h configurations. However, later analyses indicated that the intruder 4p4h

configurations were necessary to explain the energy, the long lifetime, and the (t, p) reaction

cross section of the 0+

2 state [14, 46, 47]. This is in agreement with the importance of the 4p4h

configurations suggested in recent shell-model studies [21, 22]. Since the first observation,

the 0+

2 state has not been confirmed and its lifetime remains largely unconstrained. The

importance of the 0+

2 state in understanding the contribution of intruder configurations such

as the 4p4h configurations merits additional measurements of the state energy and lifetime.

The work presented in this dissertation has tried to characterize the collective properties

driven by intruder configurations in 32Mg with new experimental observations. First, with a

model-independent lifetime measurement of the 2+

1 state, the B(E2) discrepancy from past

measurements can be resolved. This is also an important quantity for discriminating between

theoretical calculations. Second, the lifetime measurement of the 4+

1 is another goal since

this can provide additional insight about the collective nature of the ground-state band.

Finally, the 0+

2 state should be explored again to confirm the existence of this state and

provide unique insights on the contribution from intruder configurations on the collectivity

in 32Mg. Ultimately, with these new measurements, a clearer picture of the collectivity in

32Mg should be apparent.

21

Chapter 2

Experimental Techniques

The observations presented in this dissertation are from an experiment studying the in-flight

decays of excited states in 32Mg. The experiment has the following brief outline. A rare-

isotope beam of 34Si is produced from a stable beam of 48Ca ion with fragmentation reactions

on a 9Be target. The 34Si rare-isotope beam is delivered to the secondary 9Be target where it

undergoes two-proton removal reactions resulting in 32Mg. The 32Mg reaction products are

produced in excited states and continue to move at relativistic velocities along the beamline.

Gamma rays are emitted from the 32Mg ions as they decay in-flight to lower-energy states.

Ultimately, each positively identified 32Mg reaction product is associated with the gamma

rays observed in coincidence.

This chapter describes the details of several of the techniques used in this experiment.

First, this chapter provides a discussion of gamma-ray spectroscopy since the observation of

gamma rays emitted from in-flight decays is the basis for the lifetime measurement techniques

which are used in this dissertation. Then, the lifetime measurement techniques used in this

work are described, including the Doppler-shift Attenuation Method, the Recoil-distance

Method, and the novel Cascade Doppler-shift Method. The last section of this chapter

discusses the properties of the direct reaction mechanism that is used in this experiment to

produce 32Mg.

22

2.1 Gamma-ray Spectroscopy

The gamma decay of a nucleus causes a change of energy in the nucleus which is equal to

the energy of the emitted gamma ray1. By measuring the energy of the emitted gamma ray,

the excitation energy of the nuclear states can be reconstructed.

2.1.1 Gamma-ray Interactions in Matter

The three most relevant interactions between gamma rays and matter for this experiment are

photoelectric absorption, Compton scattering and pair production. Each of these interactions

results in the transfer of energy from the gamma ray to the absorbing material. If all energy

is transferred from the gamma ray it vanishes. If the gamma ray survives with a portion of

its original energy and momentum, it is deflected in a new direction and can interact again

at another location in the material. The greater the mass attenuation coefficient µ/ρ for a

given interaction, material, and gamma-ray energy, the more likely it is for the interaction to

take place. The mass attenuation coefficients for gamma rays in sodium iodide, a material

used in many gamma-ray detectors, are shown in Fig. 2.1 for various interactions [48]. For

gamma rays with less than about 250 keV of energy, photoelectric absorption is the most

likely interaction. Between energies of 250 keV and 7 MeV, Compton scattering is most

likely. For high-energy gamma rays above 7 MeV, pair production is most likely. A common

scenario for a gamma ray detected in the present experiment is a series of Compton-scatter

interactions until the remaining gamma-ray energy is low enough that a final photoelectric

absorption takes place.

1The nucleus also recoils and carries away part of the energy difference between the two states as kinetic
energy, but the kinetic energy of the nucleus is small compared to the energy of the gamma ray for the
systems discussed in this work.

23

Figure 2.1: The mass attenuation coefficients for gamma-ray interactions in sodium iodide.
Photoelectric absorption, Compton scattering and pair production interactions account for
nearly all of the attenuation. The figure is from Ref. [48].

In photoelectric absorption, the gamma ray transfers all of its energy to an absorber

atom [49]. The energy deposited into the atom removes an electron from binding. Any

energy from the gamma ray that is in excess of the electron binding energy provides the

electron with initial kinetic energy. The gamma rays of interest in this experiment were all

above 100 keV in energy so the ejected electron carries away most of the gamma-ray energy

as kinetic energy, and the energy lost to unbind the electron is negligible. The likelihood

of photoelectric absorption increases with proton number by Zn where n is between 4 and

5. High-Z materials are often used in gamma-ray detectors to enhance the likelihood of

24

photoelectric absorption.

In Compton scattering, the gamma ray scatters off an electron in the material [49]. The

gamma ray transfers only a portion of its energy and is deflected by the angle θ relative to

its incoming trajectory. The remaining energy of the gamma ray is

E(cid:48) =

E

1 + E

m0c2 (1 − cos θ)

(2.1)

where E is the initial energy of the gamma ray, and m0 is the electron rest mass. Eq. 2.1 is

true for gamma rays that Compton scatter with a free electron. For the more likely case of

Compton scattering with an electron bound in an atom, the binding energy of the electron

must be included, but this is small enough that Eq. 2.1 can be reliably used [49].

Pair production can only occur for gamma rays with energy above 1.02 MeV. In this

interaction the gamma ray is fully absorbed to create an electron-positron pair. Any gamma-

ray energy that is greater than 1.02 MeV contributes to the kinetic energy of the pair.

The positron can then annihilate with another electron in the material which generates a

subsequent pair of gamma rays with energy of 511 keV.

When these gamma-ray interactions occur in a detector, they result in free electrons inside

the material. These electrons cause an electric signal which is proportional to the energy

deposited by gamma rays. Through the observation of this electric signal, the gamma-ray

energy can be measured.

2.1.2 Relativistic Doppler Shift

Gamma rays which are emitted from a source that is moving in-flight relative to a stationary

observer experience Doppler shift. This is a result of the fundamental wave-like nature of

25

photons and special relativity. Two observers which are moving relative to one another will

disagree on the measured frequency of a gamma ray, and thus the energy. The laboratory

itself including the gamma-ray detectors is considered stationary compared to the nucleus

which is moving at relativistic speeds. Any gamma rays emitted from the in-flight nucleus

will be emitted with an energy that is Doppler-shifted in the laboratory frame.

The Doppler-shifted energy observed in the laboratory frame of reference is

ES =

EU
γ

1

1 − β cos θ

(2.2)

where EU is the unshifted energy of the gamma ray in the frame of the ion, β = v/c is the

velocity of the ion relative to the speed of light c, and θ is the angle between the trajectory

of the ion and the direction of emission of the gamma ray. γ is the Lorentz factor

1(cid:112)

1 − β2

.

γ =

(2.3)

The ion-frame energy EU is often the quantity of greatest interest since it is used to find

the excitation energy of the nuclear states. To determine EU from the observed lab-frame

energy ES, a Doppler-shift correction must be applied which is the inverse of the Doppler

shift in 2.2,

EU = ESγ(1 − β cos θ).

(2.4)

To obtain the energy between two states of an isotope moving at relativistic speeds, it is

necessary to measure not only ES, but also β and θ. The instruments used to measure these

quantities are described in Chapter 3.

The resolution of ES varies depending on the gamma-ray detector being used and often

26

comes at a trade-off with efficiency. Excellent gamma-ray energy resolution can be obtained

by using high-purity germanium detectors [50]. For gamma rays emitted in-flight, the un-

certainties in β and θ lead to an error in EU which is larger than the error in ES and is

known as Doppler broadening. The angle θ can have several different sources of error. The

position uncertainty of the gamma-ray interaction is a source of error in θ. Since θ depends

on the position and trajectory of the ion when the gamma ray is emitted, the uncertainty

in θ is also affected by any uncertainty in the ion position and trajectory. The lifetime of

the excited state contributes to the error in θ since the reaction products will travel different

distances before decaying. Sources of error in β come from the momentum distribution of

the reaction product, and additional momentum broadening can be caused as the product

passes through a material. β can be measured by detectors downstream of the gamma-ray

emission.

While the Doppler-shift of gamma rays emitted in-flight is often a challenge for detection,

the effect can actually be a valuable property for many lifetime measurement techniques. As

mentioned in the previous paragraph, the lifetime of an excited state causes in-flight reaction

products to decay at various positions along the beamline, changing the angle θ in Eq. 2.2

event-by-event. The gamma-ray energy spectrum can be sensitive to the lifetime of the

excited state based on the distribution of these decay positions. This method of analyzing

the lifetime is called the lineshape method and more details can be found in Ref. [51]. Other

lifetime measurement techniques cause the parameter β in Eq. 2.2 to depend on the lifetime,

for example by placing degrader or stopper foils in the path of the recoiling ion. The ion

slows as it moves through the material so that in-flight decays occur at a β that depends on

the amount of material the ion has passed through, and hence depends on the lifetime. This

is the premise of methods such as the Doppler-shift Attenuation Method and the Recoil-

27

distance Method [52] which are discussed in Sections 2.2.1 and 2.2.2 respectively.

2.2 Lifetime Measurement Techniques

The excited states of a nucleus can spontaneously decay, transitioning to another state. A

sample of N0 nuclei will decay according to the law of radioactive decay

N (t) = N0e−t/τ

(2.5)

where τ is the lifetime of the state. Two different nuclear states are connected by an operator

ˆO which is related to the lifetime of the initial state

∝ (cid:107)(cid:104)ψf(cid:107) ˆO(cid:107)ψi(cid:105)(cid:107)2.

1
τi

(2.6)

Therefore, lifetimes are connected to the fundamental radioactive nature of nuclear states,

and can provide an understanding of the transition strength between two states. The tran-

sition strengths in turn can illuminate profound nuclear structure properties such as collec-

tivity, configuration mixing, and shape coexistence, discussed in part in Section 1.3.

The lifetime of an excited state can span orders of magnitude from 1 attosecond to

thousands of years. Many sensitive techniques shown in Fig. 2.2 have been developed to

measure lifetimes throughout this range [52]. It is necessary to use different techniques for

measuring different ranges of lifetimes. When possible, the direct methods shown in Fig. 2.2

are preferred because they do not rely on external inputs from theoretical calculations.

The details of lifetime measurement techniques used in this dissertation are discussed

in the following sections. The discussions cover the Doppler-Shift Attenuation Method, the

28

Figure 2.2: Several methods used to measure the lifetimes of excited states. The figure is
from Ref. [52].

Recoil-distance Method, and the Cascade Doppler-shift Method which is a novel technique

designed to study isomers decaying in-flight.

2.2.1 Doppler-shift Attenuation Method

The Doppler-shift Attenuation Method (DSAM) can be applied to measure lifetimes ranging

from 1 fs to a few ps. DSAM relies on slowing the nucleus as it travels through a material

on the same time scale as the lifetime of the state of interest. The moving ion emits gamma

rays that are Doppler-shifted according to Eq. 2.2. As the ion travels through a material

the ion loses energy and slows. Often, a stopper or backing is applied to the target to

increase the energy loss of the reaction product shortly after it is produced. As the ion

29

slows, the Doppler-shifted energy of the emitted gamma ray changes. If the lifetime of the

excited state is on the same order as the time the ion spends slowing through the material,

the Doppler-shifted gamma-ray energy results in a spectral shape that is sensitive to the

lifetime.

The analysis of DSAM data relies on understanding the slowing of the ion as it moves

through the target and backing material [52]. The experimental average velocity ¯v is found

from the observed shift in the gamma-ray energy centroid

¯Es = E0(1 + ¯β cos θγ)

(2.7)

at various angles θγ, where ¯β = ¯v/c. The experimental velocity is expressed as F = ¯v/v0

and compared to a theoretical function F (τ ) which is sensitive to the lifetime. To sufficiently

slow the ion in the backing foil, it is often practical to use a high-Z material such as tantalum

or gold since the stopping power is larger for High-Z materials. One must have a precise

understanding of the stopping power of the nucleus as it travels through the material to

extract a precise lifetime from an observed DSAM gamma-ray energy lineshape. The stop-

ping power is a result of numerous interactions including electron stopping with the atomic

electrons of the material, and nuclear stopping with the nuclei of the material itself. This

is a largely statistical process and many models of stopping power are available. Thus, the

lifetime from a DSAM measurement is ultimately dependent on the choice of stopping power

model which are based on experimental measurements of stopping powers. By varying the

assumptions made in the stopping power model, a statistical distribution of the lifetime re-

sults can be found and used to evaluate the model-based uncertainty of the final result. In

practice, these stopping power models can be included in a simulation that incorporates the

30

other experimental conditions. The gamma-ray energy spectrum from the simulation can be

directly compared to the experimental spectrum to determine the lifetime.

2.2.2 Recoil-distance Method

For lifetimes on the order of 1 ps to a few ns, the Recoil-Distance Method (RDM) can be an

appropriate method, depending on the experimental setup. While in DSAM measurements

a foil is placed immediately after the target, in RDM measurements, the second foil is

positioned with a well-measured space between it and the target. Most decays in RDM

measurements occur either between the target and second foil, or after the second foil where

the ion is at a lower velocity. This method is practical for measuring lifetimes that are on

the same order as the time it takes for the ion to traverse the gap between the two foils. The

additional foil can either be used as a stopper where the ions come to rest, or a degrader where

the ions are slowed but then continue along the beamline. For the experiment presented in

this dissertation a degrader foil was used.

Often the detection of gamma rays is used to determine whether the decay occurred before

or after reaching the second foil. Figure 2.3 shows a device called a plunger which is used to

arrange the foils for an RDM measurement [53]. In this setup, the gamma rays are emitted

with energy that is Doppler-shifted according to Eq. 2.2. The ion is moving at a greater

velocity before reaching the second foil so gamma rays emitted in this region are emitted

with greater lab-frame energy for detectors at forward angles. After reaching the second foil

the ion velocity is reduced so the lab-frame energy for detectors at forward angles is also

reduced. As a result, the gamma-ray energy spectrum has two components corresponding

to the decays of fast-moving ions and slow-moving ions. This provides a measurement of the

number of decays occurring before and after reaching the second foil.

31

Figure 2.3: A schematic of a generic plunger device used in Recoil-distance Method (RDM)
measurements. The image is from Ref. [53].

The mathematical formalism for the RDM is outlined in Ref. [52]. The number of decays

that occur while the ion is moving at the faster velocity before reaching the degrader is

If = N0(1 − e−d/vτ )

(2.8)

where d is the distance the ions recoil and v is the velocity of the ion before the second foil.

The number of decays at the slower velocity is

Is = N0e−d/vτ .

(2.9)

32

Then, we can find the ratio

R =

Is

If + Is

= e−d/vτ .

(2.10)

The quantities v and d are be experimentally accessible, allowing the lifetime τ to be ex-

tracted.

There are numerous experimental complications that are not expressed in the above

equations. For instance, the velocity is not the the same for each ion. The trajectory of the

ion is not always a straight line through the center of the foils but makes an angle relative

to the central beam axis. The gamma-ray energy spectrum also becomes complicated by

in-flight decays occurring within the foils themselves and by the presence of neutron-induced

background. In practice, these complications can be included in a simulation that reproduces

the experimental configuration and makes an assumption about the lifetime of the excited

state. By systematically modifying the lifetime assumption in the simulation, the physical

value can be extracted by comparing the simulated gamma-ray energy spectrum to the

experimentally observed spectrum.

The feeding of the excited state of interest provides a complication to any lifetime mea-

surement. In the presence of strong feeding, the time for a state to decay does not reflect

that state’s lifetime, but instead reflects the total time for that state to decay, plus the time

for each of the higher-lying feeding states. To correct for the feeding effect, one must know

the amount of feeding present in the experiment, and the branching ratio and lifetime of

any feeding state. These can be used in a system of equations that properly portray the

population of the excited state as it is directly fed from the reaction, indirectly fed by the

feeding states, and depopulated by its own decay. Equivalently, the simulation can be used

to reproduce the feeding from these higher lying states if their feeding is well-understood.

33

2.2.3 Cascade Doppler-shift Method

For longer lifetimes (τ > 100 ps) the decay does not occur near the reaction target. Therefore,

an alternative approach is required to apply Doppler-shift corrections to emitted gamma rays.

The novel Cascade Doppler-shift Method (CDM) described here has been developed to make

Doppler-shift corrections and to measure lifetimes as well. The essential problem is to find

another way to track the emission point of the gamma ray without relying on the assumption

that the emission occurs near the target. When the gamma ray is a part of a cascade of
gamma rays that are emitted nearly simultaneously (within t ≤ 100 ps) it is possible to use

the timing information of the cascade to locate the common decay position [54]. Another

novel approach [55] is to use the energy information of the gamma-ray cascade to find the

common decay position, which is now feasible using detectors with high gamma-ray energy

and position resolution such as GRETINA [50].

In practice, we have a cascade of gamma rays γ1 and γ2 such as those shown in Fig. 2.4. In

Figure 2.4: A generic level scheme in which the highest state C has a long lifetime and is
followed by a cascade of gamma rays γ1 and γ2 involving a short-lived intermediate state B
and a ground state A.

an in-flight experiment, these gamma rays are each Doppler-shifted in the laboratory frame

according to their own emission angle, θ1 or θ2, as shown in Fig. 2.5. However, both gamma

rays are emitted at a common velocity and from a common position. If the ion-frame energy

34

Figure 2.5: A generic experimental setup in which two gamma rays γ1 and γ2 are emitted
at angles θ1 and θ2 from an in-flight ion and are detected. The gamma-ray detector array
GRETINA is depicted here [50].

EU of γ1 is known, then the emission angle θ1 can be found from the measured Doppler-

shifted energy ES and the ion velocity β using Eq. 2.2. Using the gamma-ray position

information of a high-resolution detector such as GRETINA [50], the emission point of γ1

can be determined. Since the state between γ1 and γ2 is short-lived (state B in Fig. 2.4),

it can be assumed that γ2 is emitted from the same location as γ1. Then, it is possible to

make a Doppler-shift correction to the observed lab-frame energy of γ2 using the emission

angle θ2 and Eq. 2.4.

One key assumption in a CDM measurement is that both gamma rays are emitted from

the same location. However, due to the non-zero lifetime of the state between the emission

of the gamma rays (state B in Fig. 2.5), the ion does travel through a distance between the

emissions of γ1 and γ2. This complication can be neglected if the lifetime of state B is short

enough (< 10 ps). For longer lifetimes, the spectral lineshape of the peak corresponding the

γ2 is affected.

A practical difficulty of this method is that the gamma-ray cascade can be emitted at

any position after the target, allowing θ1 and θ2 to take on any value. This causes the

35

lab-frame energies of γ1 and γ2 observed over many events to span a large range of values.

This leads to a significant background contribution since a priori, every observed energy in

the Doppler-shifted range of γ1 must be treated.

The distribution of the common decay points for many cascade events provides a probe

of the lifetime of the isomer (state C in Fig. 2.4). Longer lifetimes will result in a greater

proportion of the cascade decays occurring far from the target while for shorter lifetimes,

the decays points accumulate close to the target. For many applications, the observed

distribution of decays along the beamline is coupled with the efficiency of the gamma-ray

detector which is not uniform along the beamline. To account for this, precise information of

the detector efficiency must be known at several points along the beam path. If a simulation

that incorporates the true detector efficiency as a function of position is developed, it can

be compared with the observed decay trend to extract the lifetime.

2.3 Direct Reactions

This section describes the properties of direct reactions that are relevant to the main reaction

in this work, 9Be(34Si, 32Mg)X. Direct reactions are nuclear reactions which excite very few

of the degrees of freedom of the system [56, 57]. One example of a direct reaction is the two-

nucleon knockout reaction depicted in Fig. 2.6. The dynamics of the nucleus which losses two

nucleons can be approximately treated by considering only the two removed nucleons and

the remaining nucleus A. The complex behavior of the several individual nucleons within A

can be neglected in a direct reaction.

The direct reaction is a valuable experimental tool because the reaction mechanism is

sensitive to the wave functions of the few nucleons involved. For example, one-nucleon

36

Figure 2.6: A depiction of a two-nucleon knockout reaction. The relevant angular momentum
couplings are shown. This figure is from Ref. [57].

removal reaction experiments have led to a more general understanding of nuclear halo

systems [58]. A direct knockout reaction was used in the discovery of the proton bubble in

34Si [59].

It has been shown that two-nucleon knockout is also a direct reaction [60], indicating

the sensitivity of this reaction to the wavefunctions of the two removed nucleons. In the

present experiment, two-proton knockout is used to populate neutron-rich 32Mg from the

parent nucleus 34Si. The two-proton knockout reaction should primarily affect the two

removed protons, leaving the remaining nucleons mostly unaffected. The ground state of

34Si is considered to have a closed shell structure [32, 61] which makes direct reactions

from 34Si suitable to probe the wavefunction of states in 32Mg. The final states in 32Mg

with the largest overlap with the closed-shell ground state of 34Si are expected to have the

greatest cross section. This provides a criteria to determine the relative amount of normal

and intruder configurations in states of 32Mg.

37

Chapter 3

Setup and Calibration of

Experimental Devices

The results in this dissertation are from an experiment performed at the National Supercon-

ducting Cyclotron Laboratory (NSCL) in 2017 to study collectivity in 32Mg by measuring

the lifetimes of excited states. The excited states of interest include the yrast 2+

1 and 4+
1

states which have short lifetimes (τ < 100 ps) and the isomeric 0+

2 state which has a long

lifetime (τ > 10 ns [14]). Due to the large difference in the lifetimes involved in this study,

it was necessary to use two experimental setups at the target chamber that were each opti-

mized for a different lifetime range. The experimental devices present in the beamline before

and after the target chamber were the same between the two experimental setups. In this

chapter, the common experimental devices are described, then the different target chamber

configurations are explained. The relevant calibrations of the experimental devices is also

explained in this chapter.

The experiment in question involves a direct reaction with a rare-isotope beam. There

are only a handful of facilities available in the world that can support such experiments and

the NSCL is one of them. Several different devices are required to work in tandem to result

in a successful experiment. For this experiment, this included the following devices: the

Superconducting Source for Ions (SuSI) [62], the coupled cyclotrons K500 and K1200 [63, 64],

38

the A1900 fragment separator [65], and the S800 spectrograph [66]. The measurement of

gamma rays is crucial for the measurement of excited state lifetimes and in this experiment

was managed with GRETINA [50, 67] which is discussed in a later section of this chapter.

For the measurement of the lifetimes of the yrast states, the TRIPLEX device was used.

3.1 Superconducting Source for Ions (SuSI)

The Superconducting Source for Ions (SuSI) [62] was used to ionize atoms of 48Ca to be

used in the primary beam of this experiment. A metallic source of 48Ca was placed in an

oven and heated until gaseous 48Ca atoms could escape and can be used in the ion source.

SuSI ionizes atoms under the principle of electron cyclotron resonance [68]. The ions and

free electrons are trapped within a magnetic field. Both the ions and the electrons oscillate

at their respective cyclotron frequencies

ωc =

qB
m

(3.1)

where B is the strength of the transverse magnetic field and q and m are the charge and

mass of the particle. Since an electron has far less mass than an ion, its cyclotron frequency

is much greater. A resonance frequency (RF) field is applied to the gas at the cyclotron

frequency of the electrons to increase their temperature without directly heating the ions

which would cause more ions to escape the trap. The increased temperature of the electrons

leads to more collisions between the ions and the electrons and causes more electrons to be

ionized from the ion. Ions eventually escape the trap with a moderate charge state (48Ca8+

in this experiment) and an electric field sweeps them to the K500 cyclotron.

39

3.2 The Coupled Cyclotrons, K500 and K1200

The cyclotron accelerators [69] K500 and K1200 are used to accelerate the ions produced by

the ion source to velocities of up to 50% the speed of light for the primary fragmentation

reaction. A cyclotron has a vertical magnetic field that is transverse to the direction of the

ion motion. This sets the ions in a circular trajectory. The ions pass through electrodes

called dees that each alternate between positive and negative charge over time at the RF of

the ions. The ions experience steady acceleration from the dees throughout their motion.

With increasing speed, the ion cycles at larger radii until it reaches the limit of the cyclotron

and is guided out.

The K500 cyclotron receives partially ionized atoms from the ion source and accelerates

them to roughly 10% the speed of light. Once emitted from the K500 cyclotron, the ions

encounter a stripper foil which removes the remaining electrons from the ions. Now in their

greatest charge state, the ions enter the K1200 and are accelerated further to 50% the speed

of light. In this experiment, the 48Ca primary beam reaches 140 MeV/nucleon. The primary

beam is delivered to the 9Be production target where it undergoes a fragmentation reaction.

3.3 A1900 Fragment Separator

The A1900 fragment separator [65] accepts the many species of reaction products from the

primary fragmentation reaction and separates the secondary beam of interest from the rest.

There are twenty-four superconducting quadrupole magnets and four superconducting dipole

magnets in the A1900 fragment separator. The dipole magnets are tuned to the magnetic

40

rigidity of the desired secondary beam. The magnetic rigidity is

Bρ =

mvγ

q

(3.2)

where m, v, and q are the mass, velocity, and charge of the ions which pass along the central

1(cid:113)

1−β2

trajectory of the fragment separator, γ =

is the Lorentz factor, B is the strength of

the magnetic field created by the dipole magnets, and ρ is the bending radius.

In the first half of the fragment separator, the magnetic rigidity in Eq. 3.2 depends only

on the mass and charge of the ions. The velocity v is the same for all the fragmentation

products since the primary beam ions exit the K1200 cyclotron with the same velocity, and

velocity is approximately unchanged by the fragmentation reaction. Slits are placed at the

exit of the first half of the fragment separator to select the desired mass to charge ratio A/Z

for the secondary beam.

An aluminum wedge divides the first half of the fragment separator from the second and

acts as a degrader. The ions pass through the wedge and lose energy proportional to the

square of their atomic number. These ions now have different velocities in the second set

of magnets. As a result, the magnetic rigidity of the ions is related to the mass and proton
number of the ions by roughly Bρ ∝ A2.5/Z1.5 [70]. This allows an isotopic selection in the

second half of the A1900 fragment separator that is complementary to the selection made

by the first half of the fragment separator. Another set of slits at the exit of these magnets

is used to select the separated secondary beam of interest, which for this experiment was

34Si. The secondary beam is then guided to the experimental hall where the experiment is

carried out.

41

3.4 S800 Spectrograph

The S800 spectrograph shown in Fig. 3.1 is a superconducting device capable of analyzing

the secondary beam and secondary reaction products at the NSCL coupled-cyclotron facil-

ity [66, 71]. The S800 consists of an analysis line before the secondary target position, and a

Figure 3.1: A schematic of the S800 spectrograph. The figure is from Ref. [71].

spectrograph line after the secondary target. The secondary reaction occurs at the secondary

target position, and reaction products continue into the spectrograph line where the reaction

products are focused by two quadrupole magnets. The trajectory of the reaction products

are then bent along the dispersive axis by two dipole magnets and reach the focal plane of

the S800.

At the focal plane, several detectors shown in Fig. 3.2 are used to identify the nucleus and

measure its trajectory and momentum [71, 72]. The E1 scintillator is a plastic scintillator at

the S800 focal plane used to measure the time-of-flight (TOF) of the nucleus. The position

and angular trajectory of the nucleus are measured at the focal plane using a pair of cathode

readout drift chambers (CRDC’s). An ionization chamber is used to identify the Z of the

42

Figure 3.2: The S800 focal plane. The cathode readout drift chambers (CRDC’s), plastic
scintillator (E1 scintillator), ionization chamber, and IsoTagger hodoscope are depicted. The
image is from Ref. [71].

nucleus based on the energy loss of the nucleus through the chamber. Finally, a CsI(Na) ho-

doscope, called IsoTagger, is positioned at the end of the focal plane and, in this experiment,

is used to observe gamma rays from isomers implanted in an aluminum stopper [73, 74].

The S800 supports two modes of operation: the focus mode which was used in this

experiment, and the dispersion-matching mode. The focus mode is designed to deliver the
secondary beam to the secondary target chamber with a small spatial focus (±0.5 cm) and
allow a large momentum acceptance (∆p/p = ±2%) in the spectrograph. In focus mode the

ion energy resolution is limited to roughly 1 part in 1000. In dispersion-matching mode the

beam is momentum-dispersed at the target position leading to a large position distribution.

43

As a result the momentum acceptance is reduced to ∆p/p = 0.5% but the energy resolution

is improved to 1 part in 5000. The angular acceptance for secondary reactions at the target

position is 100 mrad for both operation modes.

3.4.1 Cathode Readout Drift Chambers

Each of the two Cathode Readout Drift Chambers (CRDC’s) shown in Fig. 3.3 is composed

of cathode pads and an anode wire to measure the dispersive and non-dispersive positions

of the reaction product. The volume of each CRDC is filled with gas consisting of 80% CF4

and 20% C4H10. As a nucleus passes through a CRDC it ionizes atoms in the gas and the

ionized electrons drift to the anode wire, while the positive charges drift to the cathode pads.

Charge is collected on several of the pads. Instead of choosing the pad with the greatest

Figure 3.3: A schematic of a CRDC detector used in the S800 focal plane. The beam axis
z, dispersive axis x, and non-dispersive axis y are shown. The image is from Ref. [71].

charge deposit as the location of the reaction product, the entire distribution is used to

44

provide sub-pad resolution along the dispersive direction. The drift time of the electrons to

the anode wire is sensitive to the position along the non-dispersive direction. In this way,

the first CRDC is used to obtain the x and y positions of the ion at the focal plane (xf p and

yf p) relative to the central z-axis. The second CRDC is located 1 m past the first CRDC

and is used to find the angles af p and bf p of the trajectory of the ion relative to the z-axis in

the x and y directions, respectively. These coordinates are depicted in Fig. 3.4. The CRDC’s

Figure 3.4: A schematic of the CRDC detectors used in the S800 focal plane and the position
and angular information of the ion. The beam axis z, dispersive axis x, and non-dispersive
axis y are shown. The first CRDC provides a measurement of the x and y position at the
focal plane, xf p and yf p. The angle information af p and bf p comes from the difference in
the x and y positions between the two CRDC’s.

are limited to a rate of about 5000 ions per second, mostly due to the long drift time of the

positive ions.

The CRDC position and angle information is used to reconstruct the trajectory of the

nucleus at the target position. This is done with the ion optics code COSY INFINITY to

determine the transfer matrix S [75]. S represents the transformation of the ion parameters

as it travels from the target position to the focal plane due to the S800 magnets and includes
magnetic field aberrations up to the 5th order. The inverse matrix, S−1, is used to map

45

the parameters measured at the focal plane back to the parameters at the target position.

The parameters determined by the CRDC’s at the focal plane are denoted xf p, yf p, af p,

bf p. Because the dipole magnets bend the beam in the x-direction, the xf p parameter is

used to determine the kinetic energy of the ion. As a result, the x-position of the ion at the

target position is not known, but the kinetic energy is known. The parameters at the target

position are the y-position yta, the angle of the trajectory in the x-direction ata, the angle in

the y-direction bta, and the kinetic energy of the ion relative to the reference kinetic energy

that would pass through the center of the S800 dipoles dta.

Each CRDC consists of 224 pads and each pad carries an electric signal. The signals of

all the pads of a CRDC must be matched to one another by adjusting the voltage gain and

offset. Three different isotope species that illuminate all the pads are used to calibrate all

the gains and offsets. When they are properly calibrated, the CRDC pad energy for a single

isotope appears as a straight line across all of the pads. The process is iterative, typically

requiring fewer than five iterations to converge on final offset and gain values. An example

of the uncalibrated and calibrated CRDC signals for one isotope is shown in Fig. 3.5.

The cathode pad voltage and drift time information of each CRDC must be related to

absolute values of x and y, respectively. This is done by placing an aluminum plate with

holes at known positions in front of each detector. This aluminum plate is called a mask.

The absolute position of the holes in each mask is known. During the experiment, a mask

is placed in front of each CRDC one at a time and the secondary beam is delivered to the

CRDC, illuminating only the positions on the pad where a hole in the mask is located.

Figure 3.6 shows a typical depiction of a CRDC illuminated by ions passing through the

holes of a mask. The known positions of these holes can then be associated with signals

from the CRDC. A linear calibration consisting of a slope and an offset is used to relate

46

Figure 3.5: The 224 channels of each CRDC must have their signals matched so that there
is no bias for a particular channel due to the variations in the gain of the electronics. The
signal from the channels before matching is shown in the top spectrum for the first CRDC,
and the signal from the same channels after matching is shown in the bottom spectrum. The
events are for the incoming 34Si secondary beam and 32Mg reaction product observed at the
focal plane.

the y (non-dispersive) position to the drift time of the signal. This calibration depends on

the properties of the gas inside the CRDC’s and can therefore change over the course of the

experiment. An example of the calibration done for the non-dispersive direction is shown in

Fig. 3.6. A mask calibration was performed just before lifetime data was taken with each of

the target chamber setups. The x (dispersive) position is constrained by the known absolute

position of the pads in the CRDC’s, so the mask calibration is only used to confirm the

linear calibration from the physical arrangement of the pads.

47

Figure 3.6: The top section shows the uncalibrated x and y positions of the first CRDC with
the mask placed in front. The holes in the mask allow ions to pass through to the CRDC.
The lower section shows the linear calibration between the raw position values ch and the
calibrated positions y. The line is defined by y = −0.156248ch + 107.28

3.4.2 Ionization Chamber

The ionization chamber at the focal plane of the S800 measures the energy lost by the ion

as it travels through the chamber. The stopping power is the energy lost per unit length by

an ion passing through a medium and is described by the Bethe formula [49],

dE
dx

=

4πe4z2
m0v2 N B

48

(3.3)

where

(cid:20)

B ≡ Z

(cid:19)

(cid:18)
1 − v2
c2

− v2
c2

(cid:21)

.

ln

2m0v2

I

− ln

(3.4)

In this equation the properties of the heavy ion are the proton number z and the the velocity

v, while the atomic properties of the absorbing medium are described by the number density

N , the atomic number of the atoms Z, and the average excitation and ionization energy of

the electrons I. e is the fundamental charge, m0 is the electron mass, and c is the speed

of light. The Bethe formula shows the energy loss of a heavy ion is proportional to z2/v2.

Thus the energy loss can be effectively used to identify the element of the ion.

The stopping medium used in the S800 ionization chamber is P10 gas (90% argon and

10% methane). The chamber is divided into sixteen flat segments stacked parallel to one

another and perpendicular to the beam axis. Averaging over the sixteen segments reduces

the statistical fluctuation that would result from using only a single chamber. For elements

with higher proton number, the relative difference in stopping power decreases, but the

ionization chamber is capable of resolving elements up to Z = 50.

The sixteen pads of the ionization chamber each deliver a voltage signal which must be

matched to one another. Every pad has its energy adjusted by a slope and offset parameter

to match the value of a chosen reference pad. To perform the calibration, at least three

different reaction products are selected. For each reaction product, the mean energy loss in

each of the sixteen pads is found. One pad is chosen as the reference pad for each other pad

to be matched to. Then, for each other pad, the mean energy loss values are matched to the

mean energy loss of the chosen reference pad by a least squares fit method.

The energy loss in the ionization chamber is sensitive to the element of the ion but it

also depends on the path length of the ion’s trajectory as it moves through the chamber.

49

Since the ions in the spectrograph are bent in the dispersive x direction, the path length

in the ionization chamber depends mostly on xf p. This effect causes the energy loss for a

particular element to lose resolution, hindering the particle identification. A correction for

this is made by selecting one element and viewing the ionization chamber energy loss versus

xf p for all those ions. An example of the trend is shown in Fig. 3.7. The trend is fit with

Figure 3.7: The top section shows the uncalibrated energy loss measured by the ionization
chamber as a function of xf p. An exponential line is fit to the data and shown in red. The
exponential fit very nearly resembles a straight line in this case. The bottom section shows
the ionization chamber energy loss calibrated for xf p.

50

an exponential function

dE(x) = dE0e−p(x0−x)

(3.5)

where dE(x) is the energy loss at a position x along the dispersive direction, and dE0, p, and

x0 are parameters to be fit, with x0 representing the reference x value, and dE0 representing

the energy loss at x0. Then, the effect that xf p has on the path length is corrected by
applying the factor ep(x0−x) to each event.

3.4.3 E1 Timing Scintillator

The E1 scintillator is a plastic scintillator at the S800 focal plane used to measure the time-

of-flight (TOF) of the incoming nucleus. This is done by comparing the E1 scintillator time

with the either the S800 object scintillator (OBJ), the A1900 extended focal place scintillator

(XFP), or the timing provided by the cyclotron resonance frequency (RF).

The TOF is a critical component of identifying the nucleus in the S800. As the ion passes

through the S800 dipoles it experiences uniform circular motion under the force caused by

the magnetic fields. The uniform circular motion is described by

F
m

=

v2
r

where the force causing the circular motion is

F = Bqv

(3.6)

(3.7)

and B is the strength of the magnetic field, q is the charge of the ion, v is the ion’s velocity,

m or A is the ion’s mass, and r often denoted ρ is the radius of the circular motion. Then,

51

the mass to charge ratio can be related to the TOF t

A
q

=

Bρ
v

=

Bρ
d

t

(3.8)

where d is the distance traveled by the ion. In practice, the quantity Bρ

d is fixed so A

q is

proportional to t. The E1 scintillator also serves as the trigger for readout for the S800. It

can withstand a rate of up to 106 particles per second and has a timing resolution of 100 ps.

The rate however is more constrained by the other focal plane detectors, for instance, the

CRDC’s which have a limit of about 5000 particles per second.

In this experiment, the TOF was determined using the E1 scintillator and the OBJ

scintillator. A signal from the E1 scintillator was used as the start signal and the OBJ

scintillator signal was used as a stop signal. The time difference between the two signals is

digitized by three different devices in order to provide redundant backups of this important

quantity. A Phillips 7186 time-to-digital converter (TDC), an ORTEC 566 time-to-analog

converter (TAC), and a Mesytec 32-channel multi-hit TDC are available. The Phillips TDC

malfunctioned during this experiment. The Mesytec multi-hit TDC was used in the data

analysis due to its slightly better resolution than the ORTEC TAC.

The TOF is sensitive to the mass-to-charge ratio of ions, allowing the experimenter to

differentiate between isotopes of the same element. However, the TOF also depends on both

the x-position and the angle a that the ion trajectory makes in the x-direction. This is

because, for instance, an ion with a large xf p and af p will travel through the S800 spectro-

graph in a broader arc than an ion with a smaller xf p and af p, causing it to take a longer

time to reach the E1 scintillator. The xf p and af p dependence hinders the good separation

between different isotopes and should be corrected for. The correction is an adjustment

52

to the time-of-flight which is linear with both xf p and af p. Figure 3.8 demonstrates how

the time-of-flight depends on xf p and af p and how the correction reduces this dependence.

This correction to the TOF and the above corrections to the energy loss measured by the

Figure 3.8: The uncalibrated time-of-flight (TOF) is shown with respect to af p and xf p.
The calibrated TOF is also shown, presenting a more uniform TOF distribution which allows
for a more accurate particle identification with the S800 spectrograph.

ionization chamber were sufficient to identify the isotopes of interest in this experiment.

53

3.4.4 IsoTagger Hodoscope

Located at the end of the beam line is a CsI(Na) hodoscope called the IsoTagger. The

IsoTagger consists of 32 sodium-doped cesium iodide crystals arranged into eight rows of four

detectors as depicted in Fig. 3.9. The IsoTagger hodoscope can be used either to measure

Figure 3.9: A diagram of the IsoTagger hodoscope. The 32 sodium-doped cesium iodide
crystals are arranged into eight rows of four detectors each. The beam is stopped in an
aluminum plate and gamma rays emitted from the decay of isomers are detected by the
IsoTagger hodoscope.

the total kinetic energy of the ion to identify charge states, or to measure the energy of

gamma rays emitted from isomers that survive to the end of the S800 focal plane. In the

present experiment, the IsoTagger was used to identify isomers. In this setting an aluminum

plate is placed in front of the IsoTagger causing ions to come to a stop just before the device.

Gamma rays that are emitted following beta decay or the decay of an isomeric excited state

54

are observed by IsoTagger. The trigger from the E1 scintillator is directed to the IsoTagger

to read out the signals when a reaction product reaches the focal plane.

The energy calibration for the IsoTagger can be accomplished using a known isomer

among the reaction products. The reaction product 32Al reaches the aluminum stopper

in front of the IsoTagger and has a strongly-populated isomer with an excitation energy

of 956.6 keV and a lifetime of 290(30) ns [76]. The decay of this isomer in the aluminum

stopper will produce a cascade of two gamma rays with energies of 221.9(3) and 734.6(3) keV

which are detected in IsoTagger. These known gamma-ray energies were used to calibrate

the IsoTagger CsI(Na) detectors. Figure 3.10 shows the calibrated energy spectrum from

IsoTagger gated on 32Al reaction products detected at the focal plane. A timing gate is used

to select events in the IsoTagger energy spectra that occur shortly after the trigger in the

E1 scintillator. To use the IsoTagger to measure the lifetime of isomers, the total time for

the ion to travel from the target to the focal plane in addition to the time for the readout

trigger from the E1 scintillator to reach the IsoTagger must be determined. Since the isomer

in 32Al is already known, it was used to estimate the target reaction-to-IsoTagger readout

time as 420 ns. While the IsoTagger hodoscope is effective at observing sufficiently long-lived

isomers such as the isomer in 32Al used in the calibration, it becomes less effective as the

isomer lifetime becomes shorter (τ < 100 ps).

3.4.5 Run-by-run Corrections

This experiment spanned about one week in time. Over such a time span, it may be necessary

to consider changes to the calibrations of the S800 detectors over the course of the experiment.

A single run of the experiment lasts one hour and the properties of the detector systems can

change over the course of these runs. Hence run-by-run corrections to the detectors can

55

Figure 3.10: An IsoTagger spectrum gated on 32Al ions at the focal plane of the S800
spectrograph. The gamma-ray transitions at 221.9 and 734.6 keV which follow the decay of
the isomer at 956.6 keV are observed [76].

provide an important improvement of the experimental resolution.

The y-positions measured by the CRDC’s comes from the drift time of the positive

ionization charges, which in turn depends on the thermodynamic properties of the gas inside

the CRDC’s. The pressure and temperature of the gas in the CRDC’s can change over the

course of the experiment, resulting in the need for run-by-run corrections to the CRDC y

calibration. Figure 3.11 shows the yf p parameter for 32Mg reaction products before and after

run-by-run corrections were applied over all of the runs with the plunger setup. Without

56

Figure 3.11: yf p distributions, gated on 32Mg reaction products. The distribution before
run-by-run corrections is shown in black, while the distribution after run-by-run corrections
is shown in red. The correction causes the distribution to be centered in the CRDC and
slightly narrower.

the run-by-run corrections, the yf p distribution for 32Mg is centered slightly away from

zero. The incoming beam was checked with a viewer at the beginning of the experiment

and was observed to be centered on the nominal z-axis. The off-center yf p can be caused

by the changing characteristics of the CRDC gas over time and by shifts in the beam spot.

To correct for this, the association between raw and calibrated y-values determined by the

mask calibration above is adjusted. The offset value of the mask calibration is adjusted run-

by-run to re-center the 32Mg yf p distribution. The final distribution is shown in Fig. 3.11 in

red and is now centered and slightly narrower than the yf p distribution before run-by-run

corrections.

57

3.5 Gamma-Ray Energy Tracking In-beam Nuclear Ar-

ray (GRETINA)

The Gamma-Ray Energy Tracking In-beam Nuclear Array (GRETINA) is an array of high-

purity germanium gamma-ray detectors that combines high energy resolution, high position

resolution, and large efficiency. At the time of this experiment, GRETINA consisted of ten

cryostat modules, each of which house four high-purity germanium detectors. This section

will describe the mechanical composition including the crystal shape and segmentation, and

the technical properties including the methods for signal decomposition and gamma-ray

tracking. While a full gamma-ray tracking algorithm was not implemented in the analysis,

its discussion is included in this section for completeness. Following the description of the

general characteristics of GRETINA, a description of the setup and calibration which are

specific to this experiment are provided.

3.5.1 GRETINA Description

A single GRETINA module contains four individual high-purity germanium crystals as

shown in Fig. 3.12 [67]. The crystals have a hexagonal face pointed to the center of

GRETINA. The face is concave so that the center-facing edge of the crystal is equidistant

from the center. Moving outward from the hexagonal face, the crystal smoothly becomes

larger to compactly rest next to the adjacent crystals. In each module there are two crystals

of type A and two of type B, with slightly different geometries. Each crystal is electrically

segmented into 36 segments, as shown in the top-right part of Fig. 3.12. Each crystal also

has a central core contact which reports the total energy from the entire crystal. The central

signal provides the best energy resolution, while the signals from the segments provide more

58

Figure 3.12: Diagram of a GRETINA module. The full module is shown in the bottom
section. The four high-purity germanium crystals in a single module are depicted in the
top-left. The segmentation of a single crystal is depicted in the top-right. The figure is
adapted from Ref. [67].

position information [50]. Each segment signal is digitized and processed on a dedicated

digitizer board at a rate of 100 MHz [77, 78]. Each of the modules are held in the same

aluminum frame which is aligned in front of the the S800 spectrograph [50].

GRETINA incorporates technical advances in gamma-ray spectroscopy including signal

decomposition and gamma-ray tracking. Signal decomposition refers to the practice of an-

alyzing the digitized waveforms of the segment signals to determine sub-segment position

information of the gamma-ray interaction [67]. A single incoming gamma ray can interact

multiple times with the germanium crystal, including in the same segment. Therefore, one

segment signal represent the superposition of multiple interactions with different positions

59

and energy deposits. The segment containing the gamma-ray interaction is sensitive to the

energy, but the neighboring segment signals caused by image charges are critical for under-

standing the position of the interaction. The digitized waveforms from the segments are

compared to a library of waveforms that correspond to gamma-ray interactions at various

known points in the segment. In this way the observed waveform is decomposed into the

components that best match the possible expected waveforms and contain the energy and

position information of the one or several interactions in each segment. Ultimately, the signal

decomposition can achieve a position resolution of 2 mm.

After signal decomposition, the event can contain multiple interaction points including

Compton scattering and photoelectric interactions, and can comprise the total energy of a

real gamma ray, or the partial energy due to a backscatter event. It is the role of gamma-ray

tracking algorithms to group together the interactions caused by a common gamma ray, put

them in the correct order, and exclude interactions caused by other gamma rays. A typical

gamma-ray tracking routine follows several steps to find the best arrangement of the interac-

tion points according to a calculated figure-of-merit (FoM). First, several nearby interaction

points are grouped together and assumed to be likely from the same gamma ray. Then,

the most likely order of the interactions is found assuming the set of interactions involves

a single photoelectric interaction preceded by one or several Compton scatter interactions.

This is done by trying all possible orders of the interactions and using the measured energy

and positions information to estimate how nearly it matches the expected values according

to the Compton-scatter formula (Eq. 2.1). One possible representation of the FoM is

(3.9)

(cid:88) (cos Θi

F oM =

en − cos Θi

vec)2

wi

60

where the sum is over all Compton-scatter interactions i, Θi

en is the scattering angle deter-

mined by the Compton-scatter formula and the measured energies, Θi

vec is the scattering

angle according to the measured positions of the interactions, and wi is a weight which is

often chosen as the number of interactions [50]. The order of interactions that results at the

smallest FoM is considered the most likely order. Gamma-ray tracking algorithms can then

attempt ungrouping and regrouping the sets of gamma-ray interactions to try to achieve a

smaller FoM.

For the present experiment, the signal decomposition approach described above was used

for both online and offline data analysis. The gamma-ray tracking routine above, including

the calculation of a FoM, was not implemented. Following the technique used in Ref. [50],

the interaction point with the largest energy deposit was chosen as the first interaction

point of one gamma ray. Then an add-back routine analogous to that used in Ref. [79] was

implemented to sum energies within an r = 80 mm sphere centered on the first interaction

point. Using the remaining interaction points, a second gamma ray was reconstructed with

the first interaction point and add-back energy found in the same manner. In general, if

interaction points still remain, the same add-back routine was repeated to define additional

gamma rays.

3.5.2 GRETINA Calibration

The electric signal from GRETINA must be calibrated to relate it to a real gamma-ray

energy. This is done by placing a radioactive source with a known decay information near

the detectors. For this experiment, the energy calibration made use of a 152Eu source and

was performed before the beginning of the experiment. To check the energy calibration, a

152Eu source was placed at the TRIPLEX device target position shortly before data was

61

taken.

The methods used in this work rely on precise knowledge of the efficiency of GRETINA

as well as the energy. For each circumstance where the efficiency of GRETINA must be

understood, a radioactive 152Eu source with a known activity was used. Figure 3.13 shows

the measured efficiency of GRETINA at several energies corresponding to intense gamma

radiations from the 152Eu source. The efficiency of GRETINA is used when determining

Figure 3.13: The efficiency of GRETINA measured using a 152Eu radioactive source. For this
measurement, the source was placed at the center of GRETINA. The efficiency of GRETINA
in the GEANT4 simulation used in this work is also shown with red cross (+) symbols. The
simulation efficiency is greater than the experimental efficiency by a factor of 1.13(2). The
simulation efficiency scaled down by 1.13 is shown with red saltire (×) symbols and matches
the data well.

the intensity of gamma-ray transitions from their observed peak area. The multiple foils

used in the RDM measurement also play a role in the effective efficiency of GRETINA. The

62

attenuation of a gamma ray passing through a material is

− µ
ρ λ

= e

I
I0

(3.10)

where I0 is the initial gamma-ray intensity, I is the gamma-ray intensity after the material,

ρ is the density, µ is the linear attenuation coefficient, λ is the areal density of the material.

The areal density is related to the density and the length l over which the attenuation takes

place by λ = ρl. The quantity µ/ρ is also referred to as the mass attenuation coefficient.

The degrader foils used in the TRIPLEX device were made from Tantalum. Tantalum has

mass attenuation coefficient for 1 MeV gamma rays of µ/ρ = 0.06567 cm2/g [80]. Then

for a 1 MeV gamma ray passing through 0.3 mm of tantalum, the remaining intensity is

I/I0 = 0.968. For lower-energy gamma rays, this attenuation plays an even greater role. For

0.1 MeV gamma rays, tantalum has a mass attenuation coefficient of µ/ρ = 4.302 cm2/g,

which leads to significant attenuation, leaving only I/I0 = 0.116 of the original gamma-ray

intensity after the tantalum. To understand the effect that the gamma-ray attenuation by

the TRIPLEX foils plays on gamma-ray efficiency, the 152Eu source was placed on the face

of each of the foils, one at a time, while efficiency calibration data was collected.

The RDM and CDM measurements both use a target located upstream of the center of

GRETINA. In the CDM measurement, the gamma rays can be emitted all along the beam

axis. Therefore it is necessary to understand how the efficiency of GRETINA changes for

gamma rays emitted from various positions along the beam axis. Efficiency calibration data

was taken with the 152Eu source at several different positions along the beam axis upstream

of the center of GRETINA. Figure 3.14 shows how the efficiency of GRETINA changes with

the distance of the gamma-ray emission point from the center of GRETINA, and compares

63

this with the same trend in the GEANT4 simulation package used in the analysis. The

Figure 3.14: The efficiency of GRETINA for the 779-keV gamma ray from a 152Eu radioactive
source placed at various distances upstream of the center of GRETINA. The measured
efficiency is shown with blue circles and error bars and the efficiency of GRETINA in the
GEANT4 simulation is shown as red saltire (×) symbols.

simulation follows the same trend as the data but with an efficiency that is 1.13(2) higher

than the true efficiency at all points. This is consistent with the enhancement in efficiency

in the simulation that is seen for all gamma-ray energies with a source at the GRETINA

center as shown in Fig. 3.13.

3.6 TRIple PLunger for EXotic Beams (TRIPLEX)

The TRIple PLunger for EXotic Beams (TRIPLEX) device [81] is the dedicated plunger

device at the NSCL for the in-flight measurement of short lifetimes in rare-isotope beam

experiments. Plunger devices have been a valuable tool for lifetime measurements for many

64

decades through RDM measurements [53, 82, 83]. The principle of a plunger device it to

hold a target foil and a degrader foil at a known distance so that nuclei recoiling from the

target travel with a known flight time between the two foils which is near the lifetime of

interest. Later, the Differential Recoil-distance Method [84] led to the design of plungers

with a total of three foils [81, 85]. The TRIPLEX device was designed to hold up to three

foils to allow the Recoil-distance Method and the Differential Recoil-distance Method [84].

In this section the details of the TRIPLEX device will be described, then the calibrations of

the TRIPLEX device components will be explained.

3.6.1 TRIPLEX Description

The TRIPLEX device shown in Fig. 3.15 has three main components: the bearing unit, the

foil system, and the support structure [81]. The bearing unit holds in place the immobile

parts of the device, while allowing for smooth motion of the components associated with

the moveable foils. This portion of the plunger consists of three concentric tubes and four

sliding bearings that allow the inner and outer tubes to move relative to the middle tube.

The middle tube is stationary and is the basis that all other stationary components are

attached to. The motors are attached to the middle tube and are connected to the mobile

inner and outer tubes by a small wire to drive these tubes with a minimum of force on the

sliding bearings. Other stationary components including the connectors for the electrical

components of the device, the micrometers, and the outer casing of the device which are

attached to the middle tube.

In the foil system, each foil is attached to a circular mounting frame which in turn attaches

to a ring on the appropriate tube of the bearing unit. The target connects to the inner-most

tube, the first degrader to the middle tube, and the second degrader to the outer tube. The

65

Figure 3.15: The TRIPLEX device used in this experiment. The schematic shows (A) the
outer casing, (B) a piezoelectric motor, (C) the outermost tube, (D) the middle stationary
tube, (E) the innermost tube, (F) the target positioned by the innermost tube E, (G) the
first degrader held stationary by the middle tube, and (H) the second degrader positioned
by the outermost tube. The figure is from Ref. [81].

target and the first degrader foils are each a 5 cm × 5 cm square. Each foil is attached

to a conical mounting frame that then is connected to the ring of the appropriate cylinder

by three screws. The target and first degrader frames and foils face one another and the

separation between these two foils can be seen by eye. The second degrader foil is mounted

on a frame that aligns concentrically with the first degrader frame and can extend inside the

first degrader frame. Though the first and second degrader foils can come into contact, the

first degrader frame blocks any visual check of the separation between the first and second

degrader at short distances. To accommodate the space within the first degrader frame, the

second degrader foil is circular in shape with a diameter of 4.8 cm. Each frame is electrically

isolated from the cylinder and the rest of the plunger structure to facilitate the voltage

66

capacitance and feedback functionality of the device. The first degrader is held rigidly by

three screws. The target and second degrader have springs along the screws between the

frames and the rings so that the springs are compressed as the screws are tightened. This

arrangement holds the foils so that the target and second degrader can be smoothly adjusted

to align with the rigid first degrader.

The support structure allows the TRIPLEX device to be positioned along the beamline

and is specifically designed to be used alongside the S800 spectrograph and either SeGA

(Segmented Germanium Array) or GRETINA. The main component of the support structure

is the beam pipe which is wider than the standard GRETINA beam pipe to allow the

TRIPLEX device to be positioned inside. The beam pipe attaches on its downstream end to

an S800 magnet by a flange that allows for rough adjustments to the alignment of the beam

pipe. Fine adjustments to the position of the TRIPLEX device are made within the beam

pipe using a set of aligning screws that connect the device itself to the supporting beam

pipe. The center of the foils can typically be aligned with the defined beam axis of S800

to within 1 mm. Electronic feedthroughs are located near the upstream end of the support

beam pipe and are used to connect the electrical components of the plunger device to the

control components outside the beam pipe while maintaining vacuum inside the beam pipe.

When installed in the support system, the plunger target and degrader foils rest about 13 cm

upstream of the center of GRETINA.

Several electrical components are used to set and measure the separation between the

foils of the TRIPLEX device. The motors are N-381 NEXACT linear actuators made by

Physik Instrumente (PI) [86] as shown in Fig. 3.16. The motors operate via linear actuators

depicted in Fig. 3.17. The motors provide the force on the inner and outer tubes that

move the target and second degrader foils relative to the first degrader foil. The motors also

67

Figure 3.16: A schematic of a motor used to drive a tube and foil in the TRIPLEX device.
The image is from Ref. [86].

provide a displacement measurement that is accurate to within 1 µm. Another measurement

of the displacement of each foil is made using a pair of micrometer GTL21 probes made by

the TESA group [87]. These are attached to the immobile middle tube and are oriented to

be compressed as the other two tubes move. The micrometers have a measurable range of

about 2 cm and are accurate to within 1 µm, however accuracy degraders near the edge of

their range. Nevertheless, when they are adjusted to the appropriate range, the micrometers

are a valuable complementary measurement of the displacement measured by the motors.

A third method to measure the distance between two foils is to use the capacitance

68

Figure 3.17: A cartoon depicting the operation of the linear actuators used within the motor
of the TRIPLEX device. The image is from Ref. [86].

between the foils when a pulse is applied to the first degrader. This is possible since the

three foils are electrically isolated from one another. When a voltage pulse is applied to

the first degrader, an induced signal can be measured from both the target and the second

degrader. A pair of foils behaves approximately like a parallel-plate capacitor, where the

capacitance is inversely proportional to the separation between the plates. Therefore, the

induced signal is most sensitive to short separations between the two foils because in this

range the signal size changes dramatically. The signal also gives a clear indication of when

the two foils are touching because electrical contact between two foils will cause a sudden

electrical short across the capacitor. ORTEC modules are used to manipulate these voltage

signals. A pulser sends a pulse directly to the first degrader. An amplifier module and a linear

gate and stretcher module are used to produce useful signals from each of the target and

second degrader induced responses. A multi-channel analyzer (MCA) receives the induced

responses and a trigger signal from the pulser after being sent through a gate and delay

module.

To operate the motors and monitor the signals from the motors, the micrometers, and

69

the induced pulser response, the Plunger Feedback Control software suite for Linux (PFCL)

is used [53]. The PFCL software was first developed to control the Cologne plunger [53]

and contains tools that were not used in this experiment and will be omitted from this

description. Three graphical user interfaces (GUI’s) are provided by the PFCL software and

are based on the LabVIEW programming language. The Motor GUI is used to control the

motion of the motors and monitor the relative displacement of the motors as it moves. The

Micrometer GUI provides a readout of the signal of the TESA micrometers. The MCA GUI

is responsible for receiving the pulser trigger and induced response signals and reports the

signal height of the response. All three of these programs are controlled by a PC located

near the TRIPLEX device in the experimental hall and can be operated remotely while the

experiment is running using a second PC. A diagram showing the connections among the

PC, the GUI’s, the electronic modules and the TRIPLEX components is shown in Fig. 3.18.

3.6.2 TRIPLEX Calibration

Using the TRIPLEX device and the associated hardware and software components, a precise

understanding of the separation between the foils can be achieved. One simple check of

the device is to compare the displacement reported by the motor with the displacement

according to the micrometer when it is within range. Studying this correspondence shows

that the motor and micrometer agree within 1 µm for both the target and the second degrader

systems. Deviation does occur near the edge of the micrometer range of the micrometer since

it loses precision at the limit of its range as mentioned earlier. Another deviation occurs

when the motor and the foil change direction. In this case, the motor and micrometer can

disagree by about 5 µm but come into sub-1 µm agreement again while the foil continues to

70

Figure 3.18: A diagram showing the connections between the TRIPLEX systems and the PC.
The PC uses the multi-channel analyzer (MCA) GUI, the Motor GUI, and the Micrometer
GUI to access the three systems represented here. The Pulser sends a trigger to the MCA and
a pulse to Degrader 1 which causes induced pulse responses on Target and Degrader 2 due to
the capacitance with these foils. NIM modules are used to manipulate the trigger signal and
the induced pulse responses. A motor controller operates and reads the displacement value
from each motor. The TESA controller reads the measurements from the TESA micrometers.

be driven in the same direction.

Much care is given to the alignment of the faces of the foils. This ensures that the ions

experience the same separation between the foils, regardless of their x and y positions along

the faces of the foils. For experiments involving small separations between the foils, the

alignment becomes a large component in the uncertainty in the absolute separation between

the foils.

The first foils to be aligned are the target and first degrader foils. The first degrader is

held rigidly in place with screws attaching it to the plunger device, while the target can be

adjusted by tightening or loosening its screws against the springs that hold the target steady.

71

First, the alignment is roughly performed by eye when the foils have about 0.1 − 0.5 mm

of separation. Then the voltage response of the target is used to optimize the alignment

beyond what can be accomplished by eye. The induced voltage signal increases as the foils

become closer, then saturates suddenly when they come into electrical contact. In practice,

the target foil is aligned to the degrader to achieve the maximum possible induced voltage

without the onset of electrical contact. The maximum induced voltage response signifies that

the target and first degrader are as parallel as possible and with a minimum of separation.

Any smaller voltage means that one of the edges of the target is misaligned further away

from the first degrader.

At the maximum induced voltage response, we have a theoretical zero-distance separation

between the two foils. This relies on a number of assumptions about the foils, including that

they are perfectly flat parallel plates, that there is no non-conductive material on the foil faces

causing them to touch without reaching electrical contact, and that the surface deformation

of the foils is negligible. To test the error of this zero-separation position, the target is

driven away from the first degrader and the height of the induced voltage signal is recorded

over a number of different separations between the foils. Due to the properties of parallel

plates, the inverse of the induced voltage is linear with the distance between the plates. We

compare the nominal zero separation from the alignment with the parallel-plate capacitor

trend extrapolated to where the separation becomes zero. The difference between these two

zero-separation values is associated with our error in the separation between the foils. The

difference depends on the quality of the foils but is typically be from tens of µm to 100 µm.

To align all three foils, the target and first degrader are aligned first, then the first degrader

is removed and the second degrader is aligned to the target, and finally the first degrader

is replaced to its original position and the alignment between the first degrader and second

72

degrader is checked for consistency.

The offset calibration of the TRIPLEX foils was performed before, during, and after the

experiment. Figure 3.19 shows the inverse induced voltage on the target at various nominal

separations from the first degrader which was used in the offset calibration check while the

TRIPLEX device was in the experimental beam pipe. The zero-separation offset values

Figure 3.19: The inverse of the induced voltage on the target foil at various separations from
the first degrader. The separation is measured relative to the location where the induced
voltage saturated, indicating electric contact between the foils. The change in separation
is measured by the motor displacement. The red dots are data points which are fit with a
linear function shown by a black dotted line. The fit matches the data well for separation
< 150 µm but deviates somewhat at larger separation. At large separation the assumption
that the foils behave like a parallel-plate capacitor breaks down and the induced voltage
value changes by small degrees that are difficult to resolve.

found throughout the experiment remained consistent for any pair of foils. The offset was

30 µm for the target-first degrader offset and 80 µm for the first degrader-second degrader

offset. The offsets were used in the estimation of the uncertainty of the separation between

two foils.

73

Another critical property of the TRIPLEX device is its absolute location relative to

GRETINA and S800. The position of GRETINA relative to the S800 spectrograph is main-

tained across all S800-GRETINA experiments through the use of alignments rails and a

mounting shell that holds the GRETINA modules in place. The position of the TRIPLEX

device is defined by the position of the first degrader relative to the center of GRETINA.

For every TRIPLEX-GRETINA-S800 experiment, the same support system is used includ-

ing the dedicated TRIPLEX beam pipe and flanges. The TRIPLEX device is inserted into

the beam pipe using a push stick that has a known length, positioning the bearing system

of the plunger at the same location in each experiment. The first degrader is occasion-

ally installed with spacers placed on each screw to ensure that the target, first degrader,

and second degrader can achieve the desired separation settings and the best range of the

micrometers can be used. The length of the spacers in this experiment was 2.0 mm and

caused a corresponding shift of the first degrader. Altogether, the upstream face of the

first degrader is 12.9 cm upstream of the center of GRETINA. The value is consistent with

previous TRIPLEX-GRETINA-S800 experiments that use a similar setup.

3.7 Experimental Setup to Study Short-lived States

As mentioned before, the TRIPLEX device was used only during the portion of this exper-

iment designed to study lifetimes in the 1 ps to 100 ps range alongside GRETINA and the

S800 spectrograph. For both portions of this experiment the primary beam was 48Ca at

140 MeV/nucleon and the secondary beam was 34Si. When the TRIPLEX device was in

place, the energy of the 34Si secondary beam was 60.0 MeV/nucleon with a rate of 8670 pps

per pnA (particle-nanoampere) of primary beam and a purity of 67.4% at the focal plane

74

of the A1900 fragment separator. The energy of the secondary beam in this portion of

the experiment is lower than in the experimental setup designed to study long-lived states

(86 MeV/nucleon) because a lower energy results in greater energy loss through the de-

graders. This increases the energy difference between the peak components in the RDM

analysis and increases the sensitivity of the peak shape in the DSAM analysis to short

lifetimes. The momentum acceptance of the secondary beam was ∆p

p = 1%.

The target for this part of the experiment was a 52.9 mg/cm2-thick 9Be foil. The beryl-

lium target in this part of the experiment is less thick than the target used in the experimental

setup to study long-lived states in order to reduce the energy straggling here. The first and

second degrader foils were made of tantalum and were 420 mg/cm2 and 427 mg/cm2 thick,

respectively. Using tantalum for the degraders increases the energy loss as the reaction

product passes through each degrader so that the energy difference between the target, first

degrader and second degrader peak components is large enough to be clearly identified. The

thicknesses of the foils were found by measuring the mass of each foil, then from the known

dimensions of the foils and density of the material, the thicknesses were then calculated.

Placing three foils on the TRIPLEX device in a rare-isotope beam results in numerous

parameters that must be constrained to arrive at useful lifetime results. The strategy to

do this is to start with a simple target configuration, constrain a quantity, then move to

progressively more complicated configurations. The settings discussed here are summarized

in Table 3.1. The first and simplest setting has no foils installed. This setting is used to

observe the 34Si secondary beam itself with the S800 spectrograph, measuring it energy,

position, and angular distributions. Then, the target is installed in the TRIPLEX device

and the degraded energy of the 34Si secondary beam is measured to understand the energy

loss through the target foil. Data is also taken for the target-only setting with the S800

75

Foils
none

T
T

T, 1D

T, 1D, 2D
T, 1D, 2D
T, 1D, 2D
T, 1D, 2D
T, 1D, 2D
T, 1D, 2D

S800 setting T-1D separation 1D-2D separation

quantity of interest

34Si
34Si
32Mg
34Si
34Si
34Mg
34Mg
34Mg
34Mg
34Mg

NA
NA
NA

10.0 mm
10.0 mm
25.0 mm
0.0 mm
0.0 mm
0.0 mm
0.0 mm

NA
NA
NA
NA

10.0 mm
22.0 mm
0.5 mm
0.7 mm
1.0 mm
2.0 mm

beam energy
T thickness

gamma-ray intensity

1D thickness
2D thickness
reaction ratio

lifetime
lifetime
lifetime
lifetime

Table 3.1: The separations between the foils in the TRIPLEX device. T refers to the target,
1D refers to the first degrader, and 2D refers to the second degrader. The S800 setting refers
to whether the S800 spectrograph was tuned to the unreacted 34Si secondary beam or the
32Mg reaction product.

spectrograph tuned to accept the 32Mg reaction product. This is used to constrain the

population of excited states in the reaction of interest. The first and second degraders are

then added one by one and the S800 spectrograph is tuned to the 34Si unreacted beam to

to understand the energy loss through each of the degrader foils as well.

Data was also collected with the foils placed at large separation so that the reaction

population of states in 32Mg could be understood. For this so-called large-separation set-

ting, the target-first degrader separation was 25 mm and the first degrader-second degrader

separation was 22 mm. With the large separations, the excited states in 32Mg that are

populated in reactions occurring on the target all decay to the ground state before the first

degrader is reached. Then, the first-degrader component is due only to reactions on the

first degrader and has no contribution from reactions on the target that survive long enough

to decay after the first degrader. Similarly, the second degrader component is due only to

reactions on the second degrader, not on the target or the first degrader. Thus, from the

large distance setting it is possible to quantify the relative reaction rate for the three foils.

76

This is a critical piece of information when analyzing the lifetime with the short distance

settings. In the short distance settings the peak components are due to reactions on the

corresponding foil as well as reactions on an upstream foil that decay past the relevant foil.

Having quantified the reaction ratio from the large distance setting, the peak components

can be fully understood and a sensitive lifetime can be extracted.

Finally, the settings with short separation between the foils can be used to determine

the lifetimes. One priority of this experiment was to measure the lifetime of the 2+

1 state

of 32Mg which has been determined from past Coulomb-excitation experiments to have a

lifetime of τ = 16(3) ps [88]. To be sensitive to this lifetime range, the second degrader was

positioned with a separation of 0.5 mm, 0.7 mm, 1.0 mm, and 2.0 mm from the first degrader

corresponding to approximate travel times of the nucleus between these two degraders of 5

to 20 ps. The lifetime of the 4+

1 state is another property of interest and has not been

measured in the past. The lifetime of the 4+

1 state was estimated to be near 1 ps and so

the separation between the target and first degrader was made to be as small as possible,

nominally 0.0 mm. As the separation between the first degrader and second degrader was

changed among the four settings above, the target and first degrader separation was left at

0.0 mm to accumulate as much data as possible at the smallest separation possible. This

setup takes advantage of the capabilities of the TRIPLEX device to simultaneously measure

lifetimes near 1 ps and 16 ps by setting each of the degraders at a sensitive separation.

During the experiment, a plastic 7 mg/cm2 polyethylene foil was added to the TRIPLEX

device to reduce the amount of different charge states seen by S800. The plastic foil was

attached to the downstream edge of the frame holding the second degrader foil. There was

2.2 cm of separation between the second degrader and the plastic foil. Different charge

states of the reaction products are created when the fully stripped ion gains one or several

77

electrons while passing through the target and degrader foils. The various charge states have

different mass-charge ratios which causes a portion of the desired 32Mg reaction products to

fail to reach the focal plane of the S800 spectrograph and remain undetected. The charge

states can also lead to unwanted contamination in the particle-identification spectrum or can

cause large numbers of unwanted reaction products to overwhelm the the CRDC’s which are

limited to a rate of 5000 Hz. The plastic foil is effective at stripping the electrons from ions,

returning them to a fully-stripped state. The additional energy loss caused by the plastic foil

required the S800 spectrograph to be retuned in order to center the 32Mg reaction products

at the focal plane.

3.8 Experimental Setup to Study Isomeric States

An alternative approach was necessary to observe decays from states such as the 0+

2 state of

32Mg which was observed only once previously with an estimated lifetime of τ > 10 ns [14].

The TRIPLEX device is not useful to observe this state since most decays will occur past the

second degrader, over a range on the order of 1 m at the relativistic beam velocities of this

experiment. It was necessary to use the Cascade Doppler-shift Method (CDM) discussed in

Section 2.2.3 in order to observe the Doppler-corrected energy of gamma rays emitted from

this state.

This portion of the experiment was also performed using a 34Si secondary beam but at
an energy 86 MeV/nucleon. The resultant beam had an intensity of 9 × 105 pps and a

purity of 82%. The 0+

2 state of 32Mg was populated in the 9Be(34Si, 32Mg)X reaction on a
0.57 g/cm2-thick 9Be target using the setup shown in Fig. 3.20. To validate our method, we

examined the 31Mg products created simultaneously from the same setup which populates

78

a (7/2−) isomer at 461 keV with a lifetime of τ = 15.1(12) ns [89]. Reaction products were

identified by TOF and energy-loss measurements from the S800 spectrograph [66].

Figure 3.20: The setup for the part of the experiment with the upstream target, designed
to measure the 0+
2 isomer
which emits a cascade of gamma rays γ1 and γ2 at angles θ1 and θ2 relative to the ion
trajectory.

2 isomer of 32Mg. The 9Be(34Si,32Mg)X reaction populates the 0+

Various alterations were made to the experimental setup to prioritize the application of

the CDM measurement of the 0+

2 state of 32Mg. Instead of using the TRIPLEX device, a
single Be target foil was used with a thickness of 0.57 g/cm2. Though a thicker target foil

increases the energy straggling of the products more than a thin one, this was offset by the

lack of degrader foils. With a thicker target, the yield of the product was directly increased.

In order for GRETINA to be as efficient to decays from the 0+

2 state as possible, the

target was placed 72 cm upstream. This brought more of the decays occurring over a range

of about 1 m closer to the center of GRETINA where the array has higher efficiency. A

specialized beam pipe was constructed to hold the target at this upstream location while

allowing space for GRETINA to close around the target and beam pipe. Due to the angular

spread of the reaction products, one concern is that reaction products of interest might not

79

be accepted by the angular coverage of S800. By placing the target further upstream, this

problem can be exacerbated because the products will spread further from the center of the

beam axis before they reach S800. In this experiment, the angular acceptance of S800 was

90(2)% for products created at the upstream target.

The background produced near the target position was attenuated by placing a hollow

cylindrical lead shield around the target. This shield was 19.05(5) cm long and 6.03(5) cm

in radius with a hollow bore through the center with a radius of 3.02(5) cm, allowing the

reaction products to pass through the center. The shield was held in place by a collar that

attached to the support shell of GRETINA and closed around the beam pipe. The target

was about 8 cm upstream of the upstream edge of the shield. In this arrangement, the shield

covered the full angular range of gamma rays emitted from the target that could otherwise

be detected by GRETINA.

The standard GRETINA setup at the time of this experiment had four detectors at

forward angles and the remaining six positioned at 90 degrees to the center of GRETINA

in order to have the greatest efficiency and resolution to gamma rays emitted from the

center of GRETINA. However in this experiment, the decays occur along the beamline for a

considerable distance and sensitivity to the lifetime depends on observing decays of the 0+
2

state at various points along the 1 m range. To accomplish this, GRETINA was arranged

to have four detectors at backwards angles, two at 90 degrees, and four at forward angles

relative to the beam axis measured from the center of GRETINA. This arrangement was

also used during the TRIPLEX device portion of the experiment.

80

3.9 Data Analysis Tools

The data analysis in nuclear structure experiments typically requires the handling of large

amounts of data. This subsection will introduce the main software suites that are used to

sort the data into structures that are useful for the experimenter, and simulation tools that

probe the details of the quantities of interest.

3.9.1 SpecTcl Online Data Analysis

During the experiment, the data is quickly analyzed to provide the experimenters with

timely information so that they can make informed decisions about the next steps of the

experiment. The main tool for this purpose is SpecTcl [90]. SpecTcl is a data analysis tool

used for many experiments at the NSCL. It is based on a C++ framework and allows the fast

processing of data streams from several sources into customizable graphical user interfaces

such as histograms.

3.9.2 GrROOT and ROOT Offline Data Analysis

The analysis of data after the experiment is an iterative process using the GrROOT package

based on the CERN ROOT toolkit [91]. GrROOT was developed to unpack data from

experiments using S800 and GRETINA. The original code can be modified to accommodate

additional data streams or alternative analysis methods, such as the CDM method used

in this experiment. GrROOT sorts the data in three individual steps: first the raw data

is unpacked from its hexadecimal form into ROOT trees of decimal values, then the raw

data is calibrated, and finally the ROOT trees are sorted into histograms. The calibration

and histogram-building steps are often repeated several times as additional calibrations and

81

corrections are applied to the data stream as they become understood by the experimenter.

In fact, most of the device calibration steps described in the sections above were carried out

using GrROOT. GrROOT allows the user to make two-dimensional graphical cuts to select

events corresponding to the incoming beam and outgoing reaction product of interest.

3.9.3 G4Lifetime Simulation

The data from the experiment can be compared to simulated properties as a systematic probe

of the true physics values. G4Lifetime is a simulation program based on GEANT4 [92],

a package effective at simulating the interaction of particles with matter, that has been

developed to extract lifetime results from TRIPLEX experiments [93]. All relevant properties

of the experiment can be handled within G4Lifetime including the incoming beam properties,

the TRIPLEX foil setup, the reaction process, the emission of decay particles and gamma

rays, and the detection of these radiations. G4Lifetime was also modified to allow for the

upstream experimental setup and the CDM measurement used to observe the 0+

2 state.

There are many aspects of G4Lifetime that must be constrained by physical measure-

ments to consistently reproduce a final lifetime result. The experiment and simulation are

designed so that the only free parameter in the simulation is the final lifetime result, and

every other physical property is measured.

The first G4Lifetime parameters to be constrained are the incoming beam parameters.

These are experimentally determined with the setting with no foils installed in the beam-

line. A comparison between the experimentally measured and the simulated incoming beam

properties are shown in Fig. 3.21. Using the CRDC’s of the S800 spectrograph the ion profile

distribution is known at the target position in terms of yta, ata, bta, and dta. The simulated

distribution of the incoming beam position, trajectory, and kinetic energy are all Gaussian

82

Figure 3.21: The ata, bta, dta and yta distributions of the 34Si secondary beam. Data is
shown as a black line and a scaled simulation is shown as a red line.

distributions where the centroid and width can be adjusted to fit the observed distribution.

The target and degrader foils are implemented in the code next, one by one. Each foil

is reproduced in the simulation by providing the thickness based on the mass measurement.

The default energy loss through the foils in the simulation does not perfectly reproduce the

the experimental energy which was measured by adding the foils one at a time and measuring

the reduced energy of the unreacted 34Si beam with the S800 spectrograph. A correction

to the energy loss of the foils in the simulation is made by scaling the density of the foil to

83

match the observed energy loss. In this experiment the scale density factors are 0.990, 0.983,

and 0.992 for the target, first degrader, and second degrader foils of the TRIPLEX device

respectively, and 1.000 for the target in the upstream configuration.

The nuclear reaction that results in the 32Mg product causes a change in the momentum of

the ion based on the reaction kinematics involved. A comparison between the simulated and

the experimentally observed ion profile distribution is shown in Fig. 3.22. In the simulation

Figure 3.22: The ata, bta, dta and yta distributions of the 32Mg reaction product with all
three foils installed in the TRIPLEX device. Data is shown as a black line and a scaled
simulation is shown as a red line.

84

each reaction event has its momentum scaled by the change in mass and by a parameter

called dp frac. Then the width of the outgoing momentum is adjusted by adding a vector

∆P to the outgoing momentum. The direction of ∆P is chosen at random. The magnitude

of ∆P is also random, but chosen from a Gaussian distribution centered about the parameter

dp with width controlled by the parameter dpFWHM. In practice, the centroid of the kinetic

energy distribution dta is sensitive to the dp frac parameter. The widths and tails of the dta,

ata, bta, and yta distributions of the reaction product are controlled by the dp and dpFWHM

parameters which are varied to match the observed distributions. The G4Lifetime simulation

can also control the edges of the reaction product distributions to match the acceptance of

S800. This effect can be seen as the sharp cutoff in the dta distribution of Fig. 3.22.

The excited state properties of the product nuclei are specified by listing each state,

excitation energy, gamma-ray transition energy, lifetime, and the fraction of reactions that

populate that state. When the reaction product is produced, the ion will continue to travel

along its trajectory and emit Doppler-shifted gamma rays with a random direction according

to the specified lifetime and energy of the excited states.

The gamma-ray efficiency in the simulation was compared to that observed with cali-

bration sources before and after the experiment. A comparison is shown in Fig. 3.13. The

simulated efficiency is systematically greater than the measured efficiency across all gamma-

ray energies by a factor of 1.13(2). This is believed to be caused by the fact that in the

simulation not all of the material between the ion and the GRETINA crystals is included.

This includes materials such as portions of the beam pipe, the TRIPLEX device, and the

housing for the GRETINA crystals. As a result, gamma rays are not attenuated in the simu-

lation as much as they are in reality. To correct for the difference in efficiency, the gamma-ray

yield in the simulation is simply scaled down by the enhancement factor. Since for part of

85

the experiment the target is 72 cm upstream, it was important to see how the efficiency of

GRETINA changes along the beam axis. Figure 3.14 shows that the efficiency enhancement

in the simulation remains consistent for gamma rays emitted from various distances from

the center of GRETINA.

G4Lifetime generates simulated gamma rays according to the provided excited state

parameters. One must determine the relative population of each of these excited states by

fitting the sum of the gamma-ray responses for each populated state plus an exponential

background to the data taken during the target-only setting. This is accomplished with

the target-only setting of the TRIPLEX experimental setup. The number of counts in each

gamma-ray energy peak can be more easily quantified with only the target in place than with

multiple foils installed in the TRIPLEX device. The results of this portion of the G4Lifetime

setup for the reaction products analyzed in this work are provided in Chapter 4.

Turning to the three-foil settings, the gamma-ray spectra become complicated by the fact

that each peak now has three components, each corresponding to decays occurring just past

each of the three foils. The relative height of these components contains information on the

lifetime for reactions in the target, but they also have contributions due to reactions on each

of the degraders. The relative number of reactions on each of the three foils is quantified

with the large-separation setting. In the large separation setting, the only contribution to

each component is due to reactions on the corresponding foil since the lifetimes of the excited

states are much shorter than the time it takes a nucleus to travel between the foils.

With the simulation prepared according to the above steps, the only unknown quantity

is the lifetime of the excited states. These lifetimes are determined following a chi-squared

analysis. A lifetime is assumed in the simulation, and the resulting gamma-ray energy

spectrum is fit to the experimental data and a chi-squared value is calculated. In the fitting

86

procedure, the simulation has an exponential background added to it. The background is fit

to the experimental data at energy regions far from the gamma-ray peaks. Then, the only

free parameter to fit the simulation is the scale of the simulated response. In cases with low

statistics, the number of counts in the bins of the experimental gamma-ray energy spectrum

can be quite small (< 20 counts). This requires the use of a Poisson chi-squared statistic [94]

(cid:88)

i

χ2 = 2

yi − xi + xi log

xi
yi

instead of the more common Gaussian chi-squared statistic

(cid:88)

χ2 =

(xi − yi)2

xi

i

(3.11)

(3.12)

where xi is the number of observed counts in a bin and yi is the predicted response. For

bins with large counts the Gaussian chi-squared statistic is approximately the same as the

Poisson chi-squared statistic and the 1σ uncertainty in the number of counts N in a bin is

√

√

approximately

N . For bins with few counts, the

N approximation does not hold and the

uncertainty in the observed number of counts becomes asymmetric with greater uncertainty

in the positive direction than the negative direction. This can be seen in the low-statistics

spectra shown in the next chapter where the error bars for bins with few counts have slightly

larger positive error than negative error.

87

Chapter 4

Data Analysis and Results

The lifetimes that were studied in 32Mg have very different time scales and required different

experimental techniques. The yrast 2+

1 and 4+

1 states have short lifetimes of roughly 16 ps

and 1 ps respectively. These are best measured with the TRIPLEX device setup which can

cover both lifetime ranges simultaneously. The results of this experimental setting will be

discussed first and will include lifetimes for the 2+

1 and 4+

1 states of 32Mg. Also presented

are the results of the analysis of byproduct data which yield the lifetimes of the 2+

1 state
1 state in 20O. The 30Mg and 20O results are valuable to confirm the

in 30Mg and the 2+

techniques against previously measured data sets.

Another goal of this research is to confirm the once-measured 0+

2 state of 32Mg [14] and
2 state of 32Mg has a long lifetime that cannot feasibly

constrain its lifetime. The isomeric 0+

be measured with the same RDM method used to study the 2+

1 and 4+

1 states and instead

requires the CDM method. The energy and lifetime of the 0+

2 state are reported later in this
chapter. The exclusive cross section that populates the 0+
2 state in the 9Be(34Si, 32Mg)X
reaction is also discussed. Additional byproduct data regarding the (7/2−) state of 31Mg is

presented, confirming the CDM analysis with a known physics case.

88

4.1

32Mg Short-lived States

The diminished excitation energy of the 2+

1 state in 32Mg was one of the first evidences that
the 32Mg nuclide belonged in the N = 20 island of inversion [15]. Several past measurements

1 → 2+

of the B(E2; 0+
1 ) value agree with the picture of a collective yrast band [13, 36–40],
although the results have varied considerably. The ratio B(E2; 4+ → 2+)/B(E2; 2+ → 0+)

remains undetermined since B(E2; 4+

1 → 2+

1 ) has never been reported. To improve our

understanding of collectivity in the ground-state band of 32Mg it is the goal here to resolve

the variation among the previous B(E2; 0+

1 → 2+

1 ) results and provide the first determination

of the B(E2; 4+

1 → 2+

1 ) value.

A portion of this experiment was designed to measure the lifetimes of the 2+

1 and 4+
1

states in 32Mg. The previous measurements of the B(E2; 0+

1 → 2+

1 ) value have led to an

estimate of the 2+

1 lifetime as τ (2+

1 ) = 16(3) ps [88]. A RDM measurement is appropriate

to study this lifetime range. For ion velocities of v = 0.3c, the 2+

1 state will travel about

d = 1.6 mm in one lifetime so it is appropriate to place a degrader roughly 1 mm away from

the target. Meanwhile, the 4+

1 state may have a lifetime in a very different range, roughly

τ (4+

1 ) = 1.0 ps. To study the lifetime near 1 ps with the TRIPLEX device a degrader is

placed directly after the target with zero separation. A RDM measurement of the 2+

1 lifetime

is still simultaneously possible with a sufficiently thin degrader. In fact, the first degrader

can act as an additional target for the measurement of the 2+

1 lifetime. Thus, the TRIPLEX

device allows for the simultaneous measurement of the rather different lifetimes near 16 ps

and 1 ps estimated for the 2+

1 and the 4+

1 states, respectively.

The secondary beam of 34Si was identified using the correlation between two time-of-

flight (TOF) values shown in Fig. 4.1. Along with 34Si, there was significant contamination

89

Figure 4.1: The particle identification (PID) for the 34Si secondary beam. The x-axis is
the TOF in arbitrary units between the E1 and OBJ scintillators of the S800 spectrograph.
The y-axis is the TOF in arbitrary units between the E1 scintillator of the S800 and the
XFP scintillator of the A1900 fragment separator. Both time values were digitized with the
Mesytec multi-hit TDC. The software gate used to select 34Si ions is shown as a solid black
line.

from 35P in the secondary beam. The desired secondary beam component was selected by

applying a software gate to the TOF-TOF correlation plot. The 34Si secondary beam was

observed in the focal plane of the S800 spectrograph during the setting with no foils installed

in the TRIPLEX device. The position, momentum, and angular distributions of the 34Si

secondary beam are shown in Fig. 3.21.

4.1.1 32Mg Target-only Setting

The target-only setting provides the simplest view of gamma-ray energies from the decay

of 32Mg excited states in this experiment. This setting includes only the 9Be target in the

TRIPLEX device. The purpose of the target-only settings is to observe the states that

90

are populated in this particular reaction. The energy and intensity of each gamma ray is

measured in this setting.

The particle identification (PID) spectrum for 34Si reaction products in the target-only

setting is shown in Fig. 4.2. A software gate is placed on each of the reaction products using

Figure 4.2: The PID for reaction products resulting from the 34Si secondary beam and the
9Be target in the target-only setting. The x-axis is the TOF in arbitrary units between the
E1 and OBJ scintillators of the S800 spectrograph. The y-axis is the energy loss of ions
traveling through the ionization chamber in arbitrary units. The software gates used to
select the 32Mg and 30Mg reaction products are shown with solid black lines.

the energy-loss and TOF information in the PID spectrum. The 32Mg reaction product is

well-separated from the others so the reaction product PID gate allows most of the reaction

product contaminants to be removed.

The momentum, position and angular trajectory distributions of the 32Mg reaction prod-

ucts in the target-only setting are shown in Fig. 4.3. The yta, ata and bta distributions are

all nearly Gaussian in shape. The dta distribution appears non-Gaussian because of the

91

Figure 4.3: The ata, bta, dta and yta distributions of 32Mg. Data is shown as a solid black
line and a scaled simulation is shown as a dashed red line.

significant loss of momentum acceptance at the edges of the dta range. The profile is made

by measuring the position and trajectory of the 32Mg ions at the S800 focal plane event-by-

event and reconstructing the parameters at the target position using an inverse map. The

details on this process are provided in Chapter 3.

Gamma rays were detected with GRETINA in coincidence with the 32Mg reaction prod-

ucts. The Doppler-shift corrected energies of the gamma rays observed in coincidence with

32Mg ions during the target-only setting are shown in Fig. 4.4. The Doppler-shift correction

92

Figure 4.4: Doppler-corrected energy spectra showing four transitions in 32Mg corresponding
to prompt decays observed in this experiment. Data is shown with points and error bars
while the simulation is shown with a solid red line.

was made assuming that the gamma rays were emitted from the back of the target foil. The

value of β for the correction corresponded to the velocity of a 32Mg reaction product that

would travel along the center of the S800 analysis line, but is corrected by the dta value for

each event. The yta position and the ata and bta angles for the 32Mg ion were also included

in the correction for each event.

Several gamma rays known to correspond to the decay of excited states in 32Mg can be

seen in Fig. 4.4. The decay of the 2+

1 and 4+

1 states are indicated by the 885 and 1437 keV

gamma rays, respectively. Higher-lying excited states are observed as well. The 1958 keV

transition feeds the 2+

1 state while the 2384 keV transition feeds the 4+

1 state [95]. These

transitions are shown in the level scheme in Fig. 4.5. Other gamma rays with intensities

below 7% can be observed as well. These gamma rays most likely correspond to the 2230-

keV transition populating the 2+

1 state from a (3−, 4+) state at Ex = 3117 keV, a 2603-keV

93

Figure 4.5: Partial level scheme of 32Mg.

transition populating the 2+

transition populating the 4+

1 state from a (1−, 2+) state at Ex = 3490 keV, and a 2882-keV
1 state from a (2+, 3−) state at Ex = 5203 keV [44]. There is

also a possible transition at 3264(16) keV that has not been reported before. Due to the low

intensities of these four transitions, they have been ignored in the lifetime analysis of the 2+
1

and 4+

1 states.

Table 4.1 shows the proportion of reactions that populate the four most relevant excited

states in this work, scaled so that their total sums to 100%. These intensities are used

in the G4Lifetime simulation for the target-only setting shown in Fig. 4.4 and are inputs

for the three-foil settings. The measurement of this intensity is vital to understanding the

contribution of feeding to the 2+

1 and 4+

1 states in order to obtain a lifetime result.

94

4.1.2 32Mg TRIPLEX Large-separation Setting

Data was collected with all three foils installed in the TRIPLEX device and set at a large

separation: 25.0 mm between the target and first degrader, 22.0 mm between the first

and second degraders. The purpose of this setting is to understand the relative number of

reactions producing the 32Mg excited states that occur on each foil. With three foils installed

in the TRIPLEX device, each gamma-ray peak is split into three separate components. For

large separations between the foils, each peak component depends only on the number of

reactions that occur on a single foil, and not on the lifetime of the excited states. This is

because the lifetimes of these states are all sufficiently short that an excited state populated

by a reaction on one foil does not survive to pass through another foil before decaying. Any

yield in the fast, reduced, and slow components of a peak are only due to reactions on the

target, first degrader, and second degrader respectively.

In this experiment, no slow component was observed for any peak in the large-separation

setting. This indicates that no reactions that take place on the second degrader were de-

tected.

It is almost certain that these reactions do occur, however the reaction products

from the second degrader have lower momentum than reaction products from any other foil.

This is because the 34Si secondary beam has higher Z than any of the reaction products

and therefore the ion loses more energy if it travels through the target and the first degrader

Ex (keV)
885.3(1)
2322.3(3)

Iπ
2+
4+

2843
4706

population (%)

60(5)
14(3)
13(2)
13(2)

Table 4.1: Prompt decays in 32Mg observed in this experiment. The populations observed
in this work are consistent with Ref. [60].

95

as 34Si than it would as the 32Mg reaction product. In general, the ratio of the number of

reactions between the first and second degraders would need to be considered, but it can be

ignored in this experiment since no reactions on the second degrader were detected. The

reaction ratio r will refer to the number of reactions on the target for every one reaction

on the first degrader. The reaction ratios r can vary among the excited states in a single

reaction product, so the ratios for the 2+

1 state the 4+

1 state, denoted r(2+

1 ) and r(4+

1 ), will

be considered independently.

Due to the energy loss of reactions products moving through the additional foils, the

velocity of the ions is lower than in the target-only setting. For the three-foil settings, a new

PID for reaction products must be made. The PID spectrum for the 34Si secondary beam

in the three-foil settings is shown in Fig. 4.6. Due to the lower velocity, the TOF of the

Figure 4.6: The PID for reaction products resulting from the 34Si secondary beam and the
9Be target in the three-foil settings. The x-axis is the TOF in arbitrary units between the
E1 and OBJ scintillators of the S800 spectrograph. The y-axis is the energy loss of ions
traveling through the ionization chamber in arbitrary units. The software gates used to
select the 32Mg, 30Mg, and 20O reaction products are shown with solid black lines.

96

reaction products changes as well.

The position, momentum, and angle distribution of 32Mg ions at the target position

during the three-foil settings are shown in Fig. 3.22. Every measurable parameter of the

32Mg reaction product profile is wider in the three-foil setting than in the target-only setting.

The additional foils cause the reaction product distributions to spread out as the ions pass

through the additional foils.

In the three-foil setting there is significant acceptance loss

visible at the edges of the dta distribution and the ata distribution.

The Doppler-shift corrected gamma-ray spectrum in Fig. 4.7 shows the two components of

the 885-keV peak in the large-separation setting. The Doppler-shift correction was optimized

Figure 4.7: Doppler-corrected energy spectra showing the two components of the 885-keV
peak in the large separation setting. Only gamma rays observed in the laboratory frame with
an emission angle of θ < 70◦ are shown. The solid red line depicts the result of a G4Lifetime
simulation that assumes a ratio of r(2+
1 ) = 3.2 for the number of reactions on the target
divided by the number of reactions on the first degrader that populate the 2+
1 state of 32Mg.
The dashed gray line is the background contribution from the neutron-induced 1014-keV
gamma ray from 27Al and it is scaled by ×5 in this image.

97

for decays occurring at the back of the target foil. However, due to the lifetime of the 2+
1

state, decays can occur a few millimeters past the back of the foil in which the state was

populated. This effect contributes to the width of the observed peak components. The
width of the peak components is larger for emission angles θ closer to 90◦, but the width

is smaller for forward-emitted gamma rays since the largest effect on the Doppler-shift for

forward angles is the velocity β. Therefore, to cleanly resolve the peak components in the

large-separation data, a cut was made on the gamma-ray emission angle which requires
θ < 70◦.

In Fig. 4.7, the fast component of the peak which corresponds to reactions on the target

foil appears at the nominal gamma-ray energy of 885 keV. The reduced component corre-

sponding to the reactions on the first degrader lies at roughly 860 keV. The slow component

does not appear in Fig. 4.7 due to the momentum of this reaction component lying outside

of the S800 momentum acceptance.

The ratio of the number of reactions between the target and the first degrader are input

into the G4Lifetime simulation. Several ratios were tested in the simulation and compared

with the observed spectrum. Figure 4.8 shows the distribution of the χ2 statistic for a

number of ratio values. The best ratio is r(2+

1 ) = 3.2(6) reactions on the target for every one

reaction on the first degrader. The χ2 distribution for r(2+

1 ) reveals another local minimum

at r(2+

1 ) = 2.6. Using the value r(2+

1 ) = 2.6 for the lifetime analysis changed the result by

< 0.1 ps which is much smaller than other sources of systematic uncertainty discussed later.

The simulation assuming r(2+

1 ) = 3.2 is depicted in Fig. 4.7 as a solid red line. The 2+

1 state

has significant feeding from the 4+

1 state which can affect the 2+

1 lifetime and ratio result

shown here. The simulation in Fig. 4.7 and the χ2 distribution in Fig. 4.8 include this

feeding with the lifetime and ratio values for the 4+

1 state that are determined from analyses

98

Figure 4.8: χ2 distribution for the ratio of the number of reaction on the target to those on
the first degrader for the 2+
1 state of 32Mg. The χ2 statistic obtains an absolute minimum
at r(2+

1 ) = 3.2(6). Another local minimum can is present at r(2+

1 ) = 2.6.

described later in this chapter.

Neutron-induced background can play a significant role in the Doppler-shift corrected

spectra when cuts are made on the emission angle θ [79]. This background occurs when

residual neutrons from the reactions on the foils induce additional reactions on the materials

in the laboratory. Most commonly, the neutrons encounter aluminum, such as in the beam

pipe components, and germanium in GRETINA. The neutrons are captured by or scatter

against these nuclei, populating excited states which emit gamma rays in the laboratory

frame. When Doppler-shift corrections that are tuned to the reaction products moving at

relativistic speeds are applied, the laboratory-frame neutron-induced background is spread

over a large energy range and has negligible effect on the corrected spectra. However, a cut

on the emission angle θ can result in the neutron-induced background becoming a significant

99

part of certain energy ranges of the Doppler-shift corrected spectra.

In Fig. 4.7, the neutron-induced background relevant to the measurement of the reac-

tion ratio for the 2+
1 state is shown as a dashed gray line and is scaled up by a factor of
×5 in order to be seen. This particular neutron-induced background component is from

the 1014.56-keV transition in 27Al. The intensity of this background was measured in the

laboratory-frame gamma-ray energy spectrum for the large separation setting where a to-

tal of 20(10) counts were observed. Then, the background was included in the G4Lifetime

simulation and scaled to match the observed intensity. Ultimately, the inclusion of this

background did not have a measurable effect on the reaction ratio obtained from Fig. 4.7.

No other neutron-induced background was observed that would affect the Doppler-shift cor-

rected energy range in Fig. 4.7.

The Doppler-shift corrected gamma-ray spectrum for the 4+

1 state in the large separation

setting is shown in Fig. 4.9. Like in Fig. 4.7, the Doppler-shift correction in Fig. 4.9 is

optimized for decays at the back of the target foil, and a gate is made on the gamma-ray
emission angle θ < 70◦. The fast and reduced peak components are visible, but the slow

component does not appear because reaction products produced in the second degrader are

not accepted by the S800 spectrograph. The reaction ratio that best fits the above spectrum

is r(4+

1 ) = 0.9(2), which is the value assumed in the simulation depicted with the solid red

line. Figure 4.10 shows the χ2 distribution for several different assumptions of this ratio

value. The reaction ratio is quite different between the 4+

1 and the 2+

1 state, indicating the

importance of measuring the reaction ratios for the excited states separately.

100

Figure 4.9: Doppler-corrected energy spectra showing the lineshape of the 1437-keV peak
in the large separation setting. Only gamma rays observed in the laboratory frame with an
emission angle of θ < 70◦ are shown. The solid red line depicts the result of a G4Lifetime
simulation that assumes a ratio of r(4+
1 ) = 0.9 for the number of reactions on the target
divided by the number of reactions on the first degrader that populate the 4+
1 state of 32Mg.
The dashed gray line assumes a ratio of r(4+

1 ) = 1.6.

4.1.3 32Mg TRIPLEX Short-separation Settings

To measure the lifetimes of the 2+

1 and 4+

1 states, the foils are placed so that the time it

takes the reaction products to travel between them is similar to the excited state lifetimes.

Previous measurements of the 2+

1 state suggest a lifetime of τ (2+
experiment, reaction products travel at roughly v = 0.3c so the 2+

1 ) = 16(3) ps [88]. In this

1 excited state will travel

approximately 1.6 mm after it is produced. The second degrader was set at different distances

from the first degrader to be sensitive to the 2+

1 lifetime, as summarized in Table 3.1. To be

sensitive to the shorter 4+

1 lifetime, the target and the first degrader were positioned with

zero separation. The number of observed counts of the 4+

1 state were much smaller than for

the 2+

1 state, so the target and first degrader were left in this one setting to accumulate the

101

Figure 4.10: χ2 distribution for the ratio of the number of reaction on the target to those on
the first degrader for the 4+

1 state of 32Mg. A clear minimum is visible at r(4+

1 ) = 0.9(2).

necessary statistics to be sensitive to the 4+

1 lifetime.

The foils installed in the TRIPLEX device for the short separation settings are the same

as for the large separation setting. Therefore, the reaction product PID is the same as that

shown in Fig. 4.6, and distributions of the position, momentum, and angular trajectory of

the 32Mg reaction products are the same as those shown in Fig. 3.22.

Figure 4.11 shows the 885-keV peak in the four settings used to determine the lifetime of

the 2+

1 state in 32Mg. The Doppler-shift correction is optimized for decays occurring after

the second degrader. The Doppler-shift effect is most significant at forward angles, so to

enhance the sensitivity to the Doppler-shift and therefore the lifetime, a gate is placed on
the gamma-ray emission angle, θ < 50◦. Two components of the 885-keV peak are clearly

visible: the reduced component corresponding to decays after the first degrader, and the

slow component corresponding to decays after the second degrader. The fast component is

102

Figure 4.11: Doppler-corrected energy spectra showing the two components of the 885-keV
peak. A gate is placed on the gamma-ray emission angle θ < 50◦. The red lines depict the
results of G4Lifetime simulations that assume a lifetime of 19 ps for the 2+
1 state of 32Mg.

small enough that it is hardly visible against the background. The lack of fast component is

expected since the previous 2+

1 lifetime results suggest a value that is long enough for most

32Mg reaction products to pass the target before decaying.

The lifetime of the 2+

1 state was determined by comparing the data to simulations that

assumed various values for the 2+

1 lifetime in 1-ps increments. The lifetime for each setting

was found from the minimum of the χ2 distribution, then the values from the four settings

103

with different first degrader-second degrader separations were averaged. The lifetime of the

2+
1 state when considering all four settings is τ (2+

1 ) = 18.7(1.2) ps, considering only the

statistical uncertainty. The solid red lines in Fig. 4.11 show the results of the G4Lifetime

simulation with τ (2+

1 ) = 19 ps assumed. Feeding from the 4+

1 and the Ex = 2843 keV states

is included in the simulation in the proper proportions according to the intensity results of

Table 4.1. In this feeding scheme, the ratio r(4+

1 )=0.9 is used as determined from the large

separation analysis and the lifetime of the 4+

1 state is set to the value determined later in

this section. The lifetime of the Ex = 2843 keV state could not be well-constrained in this

dataset. The 2843-keV state has been observed previously but no lifetime or spin-parity

assignment has been made [95]. In the simulated feeding of the 2+

1 state the lifetime of the

2843-keV state is assumed to be τ = 1.0 ps. Since feeding from the 2843 keV state is a small

proportion of the population of the 2+

1 state, the feeding state lifetime has a relatively small

impact on the 2+

1 result. For example, varying the Ex = 2843-keV state lifetime from 0.1 ps

to 3 ps causes the 2+

1 lifetime to change by only 1 ps. The systematic uncertainty in the

lifetime of the 2+

1 state is 1.1 ps where the largest contribution comes from the unknown

lifetime of the Ex = 2843-keV state. The final result of this measurement concludes that

the lifetime of the 2+

1 state is τ (2+

1 ) = 18.7 ± 1.2(stat.)±1.1(syst.) ps.

While τ (2+

1 ) = 18.7 ± 1.2(stat.)±1.1(syst.) ps is the final result for the data over all

settings, the individual settings randomly deviate from that value. This is clearest in the

0.7-mm setting where the slow component of the peak is larger than the corresponding

simulated response based on the lifetime of τ (2+

1 ) = 19 ps. The best match to the 0.7-mm
1 ) = 28+5−4 ps and is shown in Fig. 4.12. Figure 4.12 shows the Doppler-

setting alone is τ (2+

shift corrected gamma-ray energy with 14 keV/bin instead of 8 keV/bin which was used

in Fig. 4.11. The larger bin size reduces the impact of statistical fluctuations and makes

104

Figure 4.12: Doppler-corrected energy spectra showing the reduced and slow components
of the 885-keV transition in the 0.7-mm setting. The red solid line depicts the results of
G4LifeTime simulation that assumes a lifetime of 28 ps for the 2+

1 state of 32Mg.

the analysis of the 0.7-mm setting more sensitive to the height of the peak components and

therefore the 2+

1 lifetime. The Doppler-shift correction and emission angle θ gate used for

the analysis of Fig. 4.12 is the same as with all the other short-distance settings shown

in Fig. 4.11. The best-fit lifetime for each individual setting is shown in Fig. 4.13 along

with the weighted average result. Figure 4.13 highlights the difference between the 0.7-mm

setting and the other short-distance settings. The individual measurement from the 0.7-mm

separation setting has little bearing on the weighted average due to the large uncertainty

in the 0.7-mm setting. By excluding the 0.7 mm setting, the weighted average becomes

τ (2+

1 ) = 18.0(1.3) ps where the error is only from the statistical uncertainty. This result is

consistent with the original conclusion that includes all four short separation settings.

Figure 4.14 shows the lineshape of the 1437-keV peak over all runs in which the target and

first degrader were at 0.0 mm of separation from one another. The Doppler-shift correction

105

Figure 4.13: The best-fit lifetime for each of the four short-separation settings is shown with
the points and error bars. The weighted-average lifetime of τ (2+
shown as a solid red line.

1 ) = 18.7 ± 1.2(stat.) ps is

in Fig. 4.14 is optimized for decays occurring at the back of the target. A gate is placed to
show only the forward-emitted gamma rays, θ < 70◦. The gate on the emission angle θ is

larger than in the study of the 2+

1 lifetime in order to increase the counts for the 4+

1 peak.

In this case the lifetime of the 4+

1 state is short enough that only the fast and the reduced

components of the peak appear in the spectrum. Too few decays occur after the second

degrader for the slow component to be visible against the background. Instead of appearing

as two separate peaks, the fast and reduced components instead form a smooth lineshape

which is sensitive to the short lifetime.

The lifetime from this spectrum is τ (4+

1 ) = 0.9(2) ps where the error is from statistical

uncertainty alone. The solid red line shows the results of the G4Lifetime simulation when

a 0.9-ps lifetime is assumed, while the dashed gray line assumes a 1.6-ps lifetime. Feeding

from higher-lying excited states, such as the state at Ex = 4706 keV, may play a role

106

Figure 4.14: Doppler-corrected energy spectra showing the lineshape of the 1437-keV tran-
sition. The solid red line depicts the results of GEANT4 simulations that assume a lifetime
of 0.9 ps for the 2+
1 state of 32Mg, while the gray dashed line assumes a lifetime of 1.6 ps.

in this result. Since the lifetime of the 4706-keV feeding state is unknown, the lifetime

of the 4+

1 state reported here is an effective lifetime which includes the time it takes for

the higher-lying states to decay. Beside these feeding considerations, the major systematic

uncertainty contributions are from the target-first degrader zero-distance offset and from the

ratio r(4+

1 ) = 0.9(2). The target-first degrader offset was found to be 30 µm which leads

to a systematic uncertainty of +0.1 ps and the ratio r(4+
of ±0.1 ps. This contributes a systematic uncertainty of +0.1 ps. The final 4+

1 ) causes a systematic uncertainty

1 lifetime is

τ (4+

1 ) = 0.9 ± 0.2(stat.)±0.1(syst.) ps.

107

4.2

30Mg 2+

1 State

Byproduct data for 30Mg was also collected during the target-only and three-foil settings,

as can be seen in the reaction product PID plots shown in Fig. 4.2 and Fig. 4.6. The 2+
1

state of 30Mg at Ex = 1483 keV has a lifetime of 2.2(3) ps [96] as shown in the level scheme

in Fig. 4.15. The lifetime of this state can be measured using the same procedure that was

Figure 4.15: Partial level scheme of 30Mg.

used to measure the lifetime of the 4+

1 state in 32Mg. This known data set is valuable to

confirm the the new results presented in this work.

In the target-only setting there were too few counts to obtain the intensity of the gamma

rays in 30Mg. Instead, the three-foil large-separation setting was used without a gate on the

gamma-ray emission angle θ. Figure 4.16 shows the Doppler-corrected gamma-ray spectrum

for 30Mg in the three-foil large-separation setting. The 2+

1 peak can be seen at 1483 keV

108

Figure 4.16: Doppler-corrected energy spectra of gamma rays from 30Mg in the three-foil,
large-separation setting. The red solid line depicts the results of G4Lifetime simulations
with the best-fit intensity of the gamma rays. No gate was placed on the emission angle θ
in order to increase the statistics of the observed peaks.

and a second peak is observed near 1898 keV corresponding to the decay of the (4+) state at

Ex = 3381 keV to the 2+

1 state. From this plot it was determined that 67(7)% of the 30Mg

reaction products populate the 2+

1 state directly, while 33(7)% populate the 4+

1 state either

directly, or through the unobserved feeding transitions.

Only the 30Mg reaction products that were produced on the target were accepted by

the S800 spectrograph. The 30Mg reaction products that were produced on either the first

degrader or the second degrader had momentum that was too low and did not make it to

the S800 focal plane. As a result, the only component of any gamma-ray peak from 30Mg

that can be observed is the target component. Therefore, it is not necessary to determine

any of the ratios of reactions among the three foils.

Due to the short nature of the 2+

1 state lifetime (τ < 5 ps), it was best to analyze the

109

lifetime in the same manner the 4+

1 state of 32Mg was analyzed. All the short-separation set-

tings were summed together to gain large statistics for the Doppler-shift corrected lineshape

of the 1483-keV transition. The lineshape was the result of decays occurring through the

target and first degrader materials. Figure 4.17 shows the lineshape of the 1483-keV gamma

ray from 30Mg. The Doppler-shift correction was optimized for decays at the back of the

Figure 4.17: Doppler-corrected energy spectra of gamma rays from 30Mg in the three-foil,
short-separation setting. A gate was placed on the gamma-ray emission angle θ < 50◦. The
solid red line depicts the simulation results assuming a lifetime of 2.3 ps. The dashed gray
line corresponds to a simulation with a lifetime of 2.9 ps.

target. The gate on the gamma-ray emission angle is θ < 50◦. The lifetime of the 4+

1 state

is not known but was assumed to be τ = 0.5 ps. As in the measurement of the 4+

1 lifetime
1 state in 30Mg is also an effective lifetime. The main source

in 32Mg, the lifetime of the 2+

of systematic uncertainty is the zero-distance offset in the target-first degrader separation,

as it was in the measurement of the 4+

measurement in 30Mg is τ (2+

1 lifetime in 32Mg. The result from the 2+

1 lifetime
1 ) = 2.3 ± 0.3(stat.)±0.1(syst.) ps. This is consistent with the

110

previously measured value of 2.2(3) ps [96].

4.3

20O 2+

1 State

Another measurable lifetime in this dataset is the 2+

1 state of 20O which is depicted in the

level scheme in Fig. 4.18. Past measurements of the 2+

1 lifetime have varied and they include

Figure 4.18: Partial level scheme of 20O.

the results of RDM measurements of τ = 10.7(4) ps [97] and τ = 14.2(8) ps [98], and a

DSAM measurement of τ = 9.8(7) ps [99]. Whichever particular measurement is correct,

the lifetime is within the applicable range for a RDM measurement in this experiment.

The 20O reaction product was not observed in the target-only PID (Fig. 4.2). Therefore

all available data for 20O comes from the three-foil settings where the reaction product can

be identified with the PID shown in Fig. 4.6.

111

Figure 4.19 shows the momentum, position and angular trajectory distributions for 20O

reaction products. An unusual momentum distribution was observed for 20O, depicting two

Figure 4.19: The ata, bta, dta and yta distributions of the 20O reaction product in the three-
foil settings. Data is shown as a solid black line and a scaled simulation is shown as a dashed
red line.

Gaussian peaks. This was found to be due to the threshold setting for the CRDC’s. The

threshold was chosen to be useful for reaction products near 32Mg. For the reaction products

with lower Z, the charge deposited in the CRDC’s can be too low to reach the threshold and

be recorded. The electronic behavior for all the CRDC pads is not identical, but resulted

in the loss of more low-Z reaction products which impinged near the central pads of the

112

CRDC’s than those near the edges. As a result, two peaks appear in the distribution where

the pads accepted more of the low-Z reaction products. This was confirmed by checking

the other oxygen reaction products as well as fluorine and nitrogen isotopes where similar

momentum distributions were observed.

The Doppler-shift corrected gamma-ray energy spectrum for 20O in the three-foil large-

separation setting is shown in Fig. 4.20. No gate is placed on the emission angle θ in

Figure 4.20: Doppler-corrected energy spectra of gamma rays from 20O in the three-foil,
large-separation setting. The data shows contributions at 1674-keV from the decay of the
2+
1 state and at 1898-keV from the 4+
1 state. The contributions of the populations of the
2+
1 state and the 4+
1 state are estimated with the simulated responses depicted with the
dashed green and the dotted purple lines respectively. The solid gray line is an exponential
background and the solid red line is the sum of the two simulation components and the
background.

Fig. 4.20. The 1674-keV peak components correspond to the decay of the 2+

1 state while

the small contribution at 1898-keV is due to the 4+

proportion of reactions that populate the 2+

1 state which feeds the 2+

1 state. The
1 state is 80(9)% while the population of the 4+
1

113

state is 20(9)%.

The ratio of the number of reactions on the target to the first degrader was also obtained

from the three-foil large-separation dataset. The best-fit ratio to the large-separation data

is r = 1.0(2). The simulated response in Fig. 4.20 assumes a reaction ratio of r = 1.0 for

both the 2+

1 and the 4+
1 states.
1 feeding of the 2+

The 4+

1 state plays a significant role in the measurement of the 2+
1

lifetime. A measurement of the 4+

1 lifetime has never been reported. A rough estimation

of the 4+

1 lifetime can be obtained from the present experiment. Figure 4.21 shows the

Doppler-shift corrected energy spectrum for the 1898-keV transition from the 4+

1 state for

all the short-separation settings summed together. It was necessary to sum all the short-

Figure 4.21: Doppler-corrected energy spectra of gamma rays from 20O in the three-foil,
short-separation setting. A gate has been placed on the gamma-ray emission angle θ < 50◦.
The dashed green line and the solid red line are from G4Lifetime simulations assuming a
lifetime of 100 and 30 ps respectively for the 4+
1 state in 20O. The solid gray line is an
exponential background.

114

separation settings to result in sufficient statistics. Through comparison with simulations,

the lifetime of the 4+

1 state is estimated to be τ (4+

1 ) = 100(70) ps where the error is from

statistical uncertainty alone.

The Doppler-shift corrected gamma-ray energy spectra for the 2+

1 state are shown in

Fig. 4.22. The best-fit lifetime for the 2+

1 state of 20O is τ (2+

1 ) = 16(2) ps from statistical

Figure 4.22: Doppler-corrected energy spectra of gamma rays from 20O in each of the the
three-foil, short-separation settings. A gate has been placed on the gamma-ray emission
angle θ < 50◦. The solid red lines are the results of a G4Lifetime simulation assuming
a lifetime of τ (2+
1 state. The solid gray lines are
exponential backgrounds fit to a large energy region.

1 ) = 16 ps and no feeding from the 4+

uncertainty alone, when the feeding from the 4+

1 state is neglected. This result, reflecting

the effective lifetime of the 2+

1 state in this experiment, agrees with the previous result

115

from Ref. [98]. The lifetime of the 2+

1 was evaluated again with an estimation of the feeding

contribution from the 4+

1 state. The 4+

1 state was included assuming 20% direct population of

the 4+

1 state and and 80% of the 2+

1 state and a 4+

1 lifetime of 100 ps. Under these conditions,

the lifetime for the 2+

1 state became τ (2+

1 ) = 11.8(13) ps where the error corresponds only

to the statistical error in this data set. This feeding-corrected value is much closer to the

results of Refs. [99] and [97]. The feeding and lifetime of the 4+

1 is poorly constrained with

this dataset but play a major role on the 2+

1 lifetime, resulting in a significant systematic

uncertainty for the 2+

1 lifetime of 5 ps.

4.4

31Mg (7/2−) Isomer

The analysis of long-lived states requires a different approach than the RDM method used

to find the results in the previous sections. The CDM method is applicable for the isomers

measured in this work which include states in 32Mg and the byproduct 31Mg reaction prod-

uct. 31Mg will be described first since the lifetime of its isomer is well known and it can be
used as a demonstration of the novel CDM measurement. 31Mg has a (7/2−) isomeric state
at 461 keV with a lifetime of τ = 15.1(12) ns. The measurement of the (7/2−) lifetime using

the CDM will demonstrate the validity of the new method.

The decays we analyzed in 31Mg emit two gamma rays in cascade that both may interact

with GRETINA multiple times. As in the analysis of the short-lived states, the interaction

point with the largest energy deposit was chosen as the first interaction point of one gamma

ray. Then an add-back routine analogous to that used in Ref. [79] was implemented to

sum energies within an r = 80 mm sphere centered on the first interaction point. Using the

remaining interaction points, a second gamma ray was reconstructed with the first interaction

116

point and add-back energy found in the same manner.

In general, if interaction points

still remain, the same add-back routine can be repeated to define additional gamma rays.

However, to improve the signal-to-noise ratio for the CDM measurement, we selected events

with exactly two interaction spheres (gamma-ray multiplicity two). GRETINA was most
efficient for decays occurring 52 to 92 cm downstream of the target, corresponding to ±20 cm

from the center of GRETINA, so only events within this range were selected to reduce the

background contribution. According to the CDM technique, if one gamma ray γ1 is assigned

to the known transition, then the Doppler-corrected energy of the other gamma ray γ2 is

determined event by event based on the common decay position (see Fig. 2.5).

The CDM is applied to the 31Mg products to find the Doppler-shift corrected energy of

gamma rays from the isomeric decay and the result is shown in Fig. 4.23(a) together with a

partial level scheme. In this case, the 171-keV transition from the 221-keV excited state was

used as a reference to determine the decay location of the isomer at 461 keV. The 171-keV

transition plays the role of γ1 in Fig. 2.5, while the 240-keV transition is γ2. The peak in

Fig. 4.23(a) was fit with a Gaussian with a centroid energy of 244(5) keV which is consistent

with the literature value of 239.9(5) keV. The low energy of the 240-keV gamma ray suggests

that all interactions are likely to occur within a 20-mm sphere. Therefore, to improve the

sensitivity to events of interest, an additional gate requiring the 240-keV gamma ray to

deposit all energy within r = 20 mm of the first interaction point was implemented. This

gate results in the lower, filled spectrum of Fig. 4.23(a) which shows a reduced background

while retaining 66% of the peak counts. The scaled background spectrum in black in Fig. 4.23

was obtained by analyzing all 34Si reaction products excluding the Mg isotopes, which mostly

consists of 34Si and 33Al.

The energy spectrum of 31Mg was studied further with a G4Lifetime simulation that

117

Figure 4.23: Doppler-corrected gamma-ray energy spectra are shown for data (blue circles),
background from the scaled response of reaction products excluding the Mg isotopes (black
lines), and the sums of the simulated and background responses (red lines). 31Mg is shown
(a) in the upper spectra and the constraint that the 244-keV transition only interacts within
r = 20 mm of its first interaction is included in the lower, filled spectrum (scaled by 0.5).
32Mg is shown (b) with the r = 20 mm gate applied to the 170-keV gamma ray in the
upper spectrum (scaled by 0.6) and with additional gates requiring proper identification of
the interaction points of the 885-keV gamma ray in the lower, filled spectrum. Insets show
the number of decays near (250 to 525 mm), at mid-distance (525 to 700 mm) and far from
the target (700 to 1000 mm). The data (open circles) is compared to simulations assuming
lifetimes of 5 ns (red solid line), 15 ns (blue dashed line), and 45 ns (black dotted line).

118

included the isomeric 461-keV state, the lower-lying 221- and 50-keV states, and the gamma-

ray transitions at 240, 221, 171, and 50 keV, as shown in the level scheme of Fig. 4.23(a).

The simulated spectrum was added to the background and scaled to fit the measured peak

at 244 keV. The simulated energy of the transition from the isomer that best reproduced the

data was 239(1) keV, as is shown with the red line in Fig. 4.23(a). The difference between

the best-fit energy of 239 keV and the peak centroid energy of 244 keV was attributed to

the τ = 192 ps lifetime of the 221-keV state, corresponding to an average distance of 2 cm

between the emission points of the 171-keV and the 240-keV gamma rays.

The lifetime of the 461-keV state was studied using the distribution of decay points along

the beam line shown in the inset of Fig. 4.23(a). Although GRETINA is most efficient for

decays from 52 cm to 92 cm downstream of the target, the decay trend was analyzed in a

larger region from 25 to 100 cm past the target to improve the sensitivity to the lifetime. The

distribution of decay points does not change significantly for lifetimes greater than 10 ns so

we cannot place an upper limit on the lifetime of the 461-keV state. However, this data places

a 1 σ lower limit of 9 ns which is consistent with the known lifetime of τ = 15.1(12) ns [41].

4.5

32Mg 0+

2 Isomer

The CDM was also used to study the 0+

2 isomer in 32Mg at Ex = 1058 keV [14]. The only

previous observation of this state suggested a long lifetime of τ > 10 ps, making the CDM

necessary to properly correct the Doppler-shift of in-flight decays.

119

4.5.1 Energy Measurement and Distribution of Decay Points

The 32Mg result is shown in Fig. 4.23(b) where the CDM approach is used except now the

885-keV transition from the 2+

1 state of 32Mg is used as a reference to find the decay location.
That is to say, for 32Mg the 885-keV transition plays the role of γ1 in Fig. 2.5, while the

172-keV transition is γ2. Since the expected energy of the 0+

2 → 2+

1 transition (172 keV)

is low, similar to the 240-keV transition of 31Mg, the same condition was required that the

gamma ray deposits its energy within r = 20 mm of the first interaction point. The upper

spectrum of Fig. 4.23(b) shows a peak-like structure close to 170 keV corresponding to the

1 transition [14]. The background distribution is reproduced by analyzing the other

2 → 2+
0+
reaction products of the 34Si beam and scaling the result as shown with a solid black line. To

understand the significance of the peak-like structure at 170 keV the measured spectrum was

compared to a simulation including the 0+

2 isomer at 1058 keV and the cascade of gamma

rays with energies 172 and 885 keV. The 2+

1 state was also included in the simulation using

the lifetime value τ = 16(3) ps determined from B(E2) results [13, 36–40, 88]. The simulated

distribution was added to the background distribution and scaled to fit the peak at 170 keV

as shown by the red line.

The 170-keV peak was unambiguously confirmed by applying additional gates to the 885-

keV candidate. The result is shown in Fig. 4.23(b) as the lower, filled spectrum presenting

a clear signal with reduced background. In analogy with the r = 20-mm gate applied to the

170-keV transition in 32Mg, a gate was applied for the 885-keV gamma ray that requires

all interaction points to lie within 60 mm of the first interaction point. Additionally, we

used the interaction point information of the detected 885-keV gamma ray to test if it is

consistent with Compton scattering. The energies of the first interaction and remaining

120

interactions were used in the Compton scattering formula to obtain the scattering angle

(see Eq. (21) of Ref. [100]). If this angle agreed within 0.7 rad with the scattering angle

deduced from the decay position and interaction position information, the event passed the

gate. In this work the 0+

2 → 2+

1 energy is 165 ± 4(stat.) ± 2(syst.) keV which is included in

the simulated response in Fig. 4.23(b). The 7 keV difference between this measurement and

the previous measurement of 172(2) keV [14] is larger than the systematic uncertainty in

this measurement due to both the lifetime of the 2+

1 state (3 ps uncertainty, corresponding

to 0.6 keV uncertainty in energy) and the decay location calculation (5 mm uncertainty,

corresponding to 2 keV uncertainty in energy). The apparent discrepancy between the

observations may be due to the limited statistics of the elusive 0+

2 gamma decay in both

studies. For the rest of this work, the weighted average of 170(2) keV for the energy of the

1 transition is adopted.

2 → 2+
0+
The distribution of the decays of the 0+

2 state places an independent 1 σ lower limit on

the lifetime of τ > 8 ns, confirming the isomeric nature of this state [14]. The upper limit

could not be constrained, however, as shown in the inset of Fig. 4.23(b).

4.5.2 Constraints on the Cross Section and the 0+

2 Lifetime

The correlation between the lifetime and partial cross section populating the 0+

2 state (in-

cluding feeding from unobserved higher-lying states) can be studied from the yield of the

170-keV peak, shown as a gray band in Fig. 4.24. Since the gamma-ray efficiency in this

CDM measurement strongly depends on the lifetime, the possible cross section can be con-

strained for a given lifetime or vice versa. For example, if the assumed lifetime is 5 to

10 ns, the gamma-ray efficiency in this setup is maximized so the assumed cross section

must be minimized. The total error in this result includes 20% statistical uncertainty in the

121

Figure 4.24: Possible values for partial cross section and lifetime of the 0+
2 state of 32Mg
within 1 σ are plotted (gray band). From the work of Ref. [60], the cross section upper limit
of 0.10 mb is included (red line), and the overlap of that work and this work is highlighted
(red hatched region).

yield of the 170-keV peak. Another important source of uncertainty is the efficiency of the

gates. In order to keep this systematic uncertainty small, the spectrum including fewer gates

(Fig. 4.23(b), upper spectrum) was used in this portion of the analysis and contributes 13%

relative uncertainty, estimated by applying the same gates to calibration data taken with a

152Eu source. The statistical and systematic uncertainties were combined in quadrature to

a total relative uncertainty of 25% in the cross section at any given lifetime.

The lifetime can be constrained by including the upper limit on the partial cross section

of the 0+

2 state from a previous two-proton removal reaction experiment [60], shown with

a red horizontal line in Fig. 4.24. The previous and present removal reaction experiments

populated states in 32Mg using similar mid-target energies of 67 and 75 MeV/nucleon re-

spectively. Eikonal model calculations [57] based on the USDB [101] two-nucleon amplitudes

predict 5.11 and 5.23 mb for the inclusive cross section at 67 and 75 MeV/nucleon respec-

tively, suggesting that the results of the previous removal experiment with a slightly different

energy can be safely applied here. In the previous measurement, the 0+

2 state was not ob-

122

served but the 2+

1 and 4+

exclusive cross section populating the 2+

1 states were observed from their gamma decays. The measured
1 states accounted for 100(12)% of the 32Mg

1 and 4+

inclusive cross section, σinc = 0.76(10) mb. This suggests that at most only 12% of the reac-

tions populate the unobserved 0+

2 state, corresponding to an upper limit of 0.10 mb for the

partial cross section of the 0+

2 state. With this upper limit in conjunction with the present

result, the partial 0+

2 cross section and lifetime are constrained to 0.03 mb < σ < 0.10 mb

and 1.5 ns < τ < 38 ns, respectively, as shown by the red hatched region in Fig. 4.24. The

lifetime can be further constrained to 10 ns < τ < 38 ns from Ref. [14].

4.5.3 Independent Cross Section Measurement

The above assumption that the previously measured cross section at a slightly lower beam

energy [60] is consistent with the cross section in this experiment is likely robust, however

we do not need to rely on it entirely. The inclusive 32Mg cross section and the exclusive

cross section to the short-lived states can be obtained from the data collected during the

target-only setting of the TRIPLEX device. Then an upper limit on the exclusive cross

section of the 0+

2 state can again be found. The target-only setting had a mid-target energy

of 57 MeV/nucleon, lower than the mid-target energy of 75 MeV/nucleon used for the CDM

measurement. Nevertheless, the cross section measured using the TRIPLEX setup provides

another result for comparison from a setup that shares many experimental properties.

The inclusive cross section of 32Mg in the two-proton removal reaction from 34Si is

σinc =

N32Mg
N34SiNt

(4.1)

where N32Mg is the number of 32Mg reaction products, N34Si is the number of 34Si beam

123

ions, and Nt is the areal number density of 9Be nuclei in the target. The exclusive cross to

a particular state i of 32Mg is

σexc,i =

N32Mg,i
N34SiNt

.

(4.2)

Also, it must be true that the sum of the exclusive cross section over all states of 32Mg equals

the inclusive cross section,

(cid:88)

i

σexc,i = σinc.

(4.3)

The inclusive and exclusive cross sections are calculated using the same beam N34Si and the

same target Nt, but the number of products, N32Mg or N32Mg,i, are obtained in different
ways. N32Mg is determined from the number of 32Mg reaction products measured with the

S800 focal plane detectors. N32Mg,i, on the other hand, is based on the number of gamma

rays from the decay of state i that are observed with GRETINA. The experimental setup

designed for prompt decays is not sensitive to the exclusive cross section of the 0+

1 or 0+
2

states because these states do not promptly decay near GRETINA. However, the difference

between the inclusive cross section and the sum of the observed exclusive cross sections to

short-lived states, j, sets an upper limit on the exclusive cross section to the 0+

2 state,

< σinc −(cid:88)

j

σ

exc,0+
2

σexc,j.

(4.4)

The number of incoming 34Si beam nuclei, N34Si can be determined from the total number

of signals counted on the S800 OBJ scintillator upstream of the target position. This number

reflects the number of secondary beam nuclei of all isotopes that impinge on the scintillator,

including contaminants in the secondary beam. To find N34Si, we need to correct the number
of signals on the object scintillator Nobj by the transmission of 34Si ions from the scintillator

124

to the target tobt,t, and by the purity of the secondary beam at the object scintillator pobj.

Strictly speaking, tobj,t and pobj cannot be measured directly in our setup. To measure tobt,t

we would need the number of 34Si at the object but we have the number of secondary beam

ions of any species, including the contaminant secondary beams. To measure pobj we need

particle identification at the object, but we have this at the S800 focal plane which is not
necessarily the same. However, the product tobj,t × pobj can be found from the unreacted
beam setting for 34Si using the number of 34Si ions in the PID at the S800 focal plane divided

by Nobj. This assumes that the transmission of 34Si from the target to the S800 focal plane

is 100%, which is safe to assume since there appears to be no acceptance loss in the 34Si ion

profile in Fig. 3.21. Altogether, the proportion of OBJ scintillator signals representing 34Si

that impinged on the target was 0.792(9). The observed number of beam nuclei must also

be corrected for the efficiency of the object scintillator which was found to be 0.946(10)% for

the runs when the S800 spectrograph was tuned to the 32Mg reaction product. Ultimately,
the number of incident beam particles was N34Si = 5.91(15) × 109.

The areal number density of the target is found from the known thickness of the target

by

Nt =

tNA
ma

(4.5)

where the thickness of the target is t = 53(2) mg/cm2, NA is Avogadro’s number, and

the atomic mass of beryllium is ma = 9.012 g. This results in an areal number density of
Nt = 3.54(13)× 1021 cm−2 for the beryllium target nuclei. The target foil was manufactured

by Goodfellow [102] and is reported to have a purity of greater than 99.8%. The effect of

impurities in the target was neglected in this cross section calculation.

The number of 32Mg reaction products is measured with the energy loss versus TOF

125

PID spectrum at the S800 focal plane. Several corrections to the number of 32Mg products

must be taken into account as well. Signals from both of the CRDC’s is required to obtain

precise energy loss and TOF results. The combined efficiencies of the two S800 CRDC’s is

0.968(14) for the 32Mg fragment. The S800 data acquisition is triggered by signals from the

E1 scintillator which experiences deadtime when the scintillator is not ready to receive the

next signal. To estimate the deadtime, a 10 Hz clock is fed through the data acquisition

system. The raw signal of the clock is recorded as well as the live signal of the clock that

is only seen when the E1 scintillator is live and can receive a new event. The ratio of the

live clock to the raw clock represents the proportion of the time that the E1 scintillator was

alive, which was 0.93(3) for the setting accepting 32Mg reaction products.

Not all 32Mg reaction products are transmitted from the target to the S800 focal plane.

This loss must be accounted for in the cross section measurement. Figure 4.25 shows the

angle af p versus the position xf p for 32Mg ions that reached the S800 focal plane. Figure 4.25

shows sharp cutoffs at large af p and xf p indicating the loss of 32Mg ions that did not reach

the focal plane. For ions with xf p near the center of the focal plane, all angles af p are

accepted. A region at the center of xf p was selected and fit with a Gaussian function,

counts(a) = Ae−(a−µ)2/(2σ2).

(4.6)

Using the center µ and width σ of the Gaussian function for the central region as fixed

parameters, the height A of the function was fit to regions of xf p away from the center. A

distribution of af p for 32Mg reaction products for different selections of xf p are shown in

Fig. 4.26. For the non-central xf p regions, there is a gap between the data and the Gaussian

function at the edges of the af p distributions. The size of these gaps were quantified and

126

Figure 4.25: Distribution of the angle af p and the position xf p for 32Mg reaction products in
the target-only setting with the TRIPLEX device. A portion of the 32Mg reaction products
are lost at the edges of the distribution with extreme values of both af p and xf p.

associated with loss of angular acceptance of the S800. These steps were repeated with

different sizes of the central xf p region that was assumed to accept all af p angles. Using the

different central regions, consistent results were found and the average value for the S800

angular acceptance was 89(2)%.

Beside the loss due to af p angular acceptance, there is also loss from the momentum ac-

ceptance of the S800 spectrograph. The momentum acceptance is correlated with the angle

af p of the ion, in a similar way to how xf p acceptance is correlated with af p. Fig 4.27 shows

the momenta of 32Mg ions detected by the S800, corrected by the angular acceptance. The

momentum distribution shows a sharp drop at the edges of the acceptance. A skew-Gaussian

function [103] was fit to the distribution to estimate the amount of 32Mg that was produced

in the reaction but not observed due to limited momentum acceptance. The momentum

127

Figure 4.26: Distribution of the angle af p for 32Mg reaction products in the target-only
setting with different xf p ranges. In the top spectrum, central values are chosen (−50 mm <
xf p < 50 mm) for which all angles af p are accepted by the S800 spectrograph. A solid red
line represents a Gaussian fit to the distribution with the height A, mean µ, and width σ as
free parameters. The bottom spectrum shows the af p distribution for for non-central position
(−300 mm < xf p < −150 mm). The red solid line in the bottom spectrum represents a
Gaussian fit with only the height A as a free parameter and the mean µ and width σ fixed
to the values obtained from the fit to the distribution in the top spectrum. The gap between
the solid red line and the data in the bottom spectrum represents the reaction products that
were lost due to limited af p acceptance of the S800 for this range of xf p values.

acceptance for 32Mg ions was determined to be 81(5)%. Considering the momentum accep-

tance, angular acceptance, the detection efficiency and deadtime, the total number of 32Mg
produced in these reactions was N32Mg = 1.39(12) × 104.

128

Figure 4.27: The parallel-momentum distribution of 32Mg reaction products in the target-
only setting. The counts are corrected for the loss of angular acceptance which occurs mostly
at small and large momenta. The solid red line depicts a skew-Gaussian fit [103] to the data
from 10.32 to 10.84 GeV/c. The difference between the fit and the data at the edges of
the momentum distribution represents the loss of 32Mg reaction products due to the finite
momentum acceptance of the S800 spectrograph.

The number of reaction products that populate an excited state N32Mg,i is found from

the number of observed gamma rays emitted from the decay of state i. To find the exclusive

cross section of state i, the gamma rays emitted from the decay of state i must be quantified,

and the number of gamma rays feeding state i must be measured as well.

The inclusive cross section of 32Mg and the exclusive cross sections to the observed excited

states are listed in Table 4.2. The cross sections from this experiment σexp are compared with

the values obtained from the past experiment σprev [60] using the same reaction mechanism.

The inclusive cross section in both the present and the past works are within agreement. The

work in Ref. [60] only observed the 2+

1 and 4+

1 states in 32Mg and notes that they observed

an excess of gamma-ray energies above 1.4 MeV suggesting the existence of higher-lying

129

state

32Mg (inc.)

2+
1
4+
1

0+
2

Ex = 2843
Ex = 4706

σprev

0.76(10)
0.52(8)
0.24(8)

σexp
0.65(7)
0.33(5)
0.08(2)
0.07(2)
0.07(2)
< 0.14 < 0.10

—
—

Table 4.2: Two-proton removal cross sections in the 9Be(34Si,32Mg)X reaction are shown.
The column labeled σexp is from the present experiment and the column labeled σprev shows
the results of the work from Ref. [60]. The upper limit for the exclusive cross section of the
0+
2 state is based on the unobserved gamma-ray intensity as described in the main text.

states that could not be resolved in that work. In the present work, the higher-lying states

at Ex = 2843 keV and Ex = 4706 keV were clearly observed. The contributions of these

higher-lying states are subtracted from the cross sections of the 2+

1 and 4+

1 states in the

present work. As a result, the 2+

1 and 4+

1 cross sections from this experiment are lower than

the results from the past experiment where the contributions from higher-lying states were

not subtracted. By adding the cross sections from the higher-lying states to the 2+

1 and 4+
1

states according to the feeding indicated in Fig. 4.5, consistent results are obtained between

the present and past experiments for the cross sections of the 2+

1 and 4+

1 states. Due to the

agreement between the cross sections measured in the present experiment and the values

from Ref. [60], it is likely safe to apply the cross sections from Ref. [60] to the analysis of

the data from the upstream experimental setting at 80 MeV/nucleon, as was done in Section

4.5.2.

130

Chapter 5

Discussion of Results

The new lifetime results for excited states in 32Mg can give us a new understanding of the

structure of this isotope. The reduced E2 transition rates, B(E2), can be found from these

lifetimes. The B(E2) describes the interaction between the two associated states and can

be a powerful probe of collective behavior.

5.1 B(E2) Values in 32Mg and Even-even Mg Isotopes

The 2+

1 state in 32Mg has been studied numerous times. As a result there are many exper-

imental results for the B(E2; 2+

1 → 0+

ment from the present experiment is τ (2+

1 ) value as shown in Fig. 5.1. The lifetime measure-
1 ) = 18.7 ± 1.2(stat.)±1.1(syst.) ps and results in
1 ) = 80(7) e2fm4, which has the least error of any of the reported B(E2)

B(E2; 2+

1 → 0+

results. The present B(E2) value from a lifetime measurement accounts for feeding from

higher-lying states including the 4+

1 state at Ex = 2332 keV. The B(E2) value from the

present experiment is slightly lower but in good agreement with the weighted average of the

previous results B(E2; 2+

1 → 0+

1 ) = 91(13) e2fm4 [88].
1 → 0+

To interpret the meaning of this B(E2; 2+

1 ) value, it is compared with theoretical

predictions for 32Mg and for other even-even Mg isotopes [96, 104–106] in Fig. 5.2. The

USDA calculation is a shell model calculation that limits the valence orbits to the sd-shell

for both protons and neutrons [107]. The sd-shell orbitals are sufficient for describing many

131

1 → 0+

Figure 5.1: The B(E2; 2+
1 ) value in 32Mg reported in experiments over the past
25 years. The present RDM measurement is shown with a red square. The Coulomb-
excitation measurements from Refs. [13,37,38,40] are shown with open circles. The open
triangles are the results from Refs. [36,39]. The upward-pointing triangles reflect the B(E2)
values without feeding corrections while the corresponding downward-pointing triangles take
into account feeding corrections. The saltire symbol (×) represents the fast timing measure-
ment of Ref. [41].

nuclei with 8 to 20 of each nucleon since the valence particles will typically occupy the

sd-shell. However, for the members of the N = 20 island of inversion such as 32Mg, it is

understood that intruder configurations of neutrons occupying the 0f7/2 and 1p3/2 orbitals

dominate the low-energy states [11, 18]. As a result, the USDA calculation underestimates

the observed B(E2) strength in 32Mg.

It is important to include the 0f7/2 and 1p3/2 orbitals in an effective shell-model de-

scription of 32Mg. The SDPF-M [32] and SDPF-U-MIX [21] interactions both include these

intruder orbitals. Both the SDPF-M and SDPF-U-MIX implement the USD interaction [107]

for the sd -shell and base the pf -shell portion on the KB interaction [108]. They can express

132

1 → 0+

Figure 5.2: The B(E2; 2+
1 ) values in the Mg isotopic chain from 28Mg to 40Mg. The
32Mg result from this experiment is shown with a filled black circle. The experimental values
from other results are shown with black saltires (×) and are from Refs. [104,96,105,106].
Theoretical calculations are also shown for SDPF-M [32] (red line and open diamonds),
SDPF-U-MIX [21] (blue line and open inverted triangles), AMPGCM [34,35] (purple line
and open circles), EKK [22] (green line and open squares), and USDA [107] (brown line and
open triangles).

both normal configurations, with neutrons occupying the sd -shell, and intruder configura-

tions with neutrons in the pf -shell, and allow mixing of among these configurations. These

calculations are more successful than the USDA calculation at reproducing the increase in

B(E2) that occurs from 30Mg to 32Mg.

The angular-momentum-projected generator coordinate method (AMPGCM) [34, 35]

result is based on a mean-field calculation using the Gogny interaction [109]. The AMPGCM

identifies different deformed minima for the states in 32Mg and mixes these configurations.

The resulting wave functions for the 0+

1 and 2+

shown in Fig. 5.2. Ref. [35] suggests that the 0+

1 states were used to obtain the B(E2) results
1 state in 32Mg contains significant mixing

133

of both oblate and prolate configurations.

The most recent calculation shown in Fig. 5.2 is the extended Kuo-Krenciglowa (EKK)

method [22], based on the Kuo-Klenciglowa (KK) method [110–112]. The EKK method is

more closely tied to the microscopic theory of quantum chromodynamics than the other cal-

culations. However, parameters in the EKK method require phenomenological adjustments

based on observables in this region of nuclides. While the EKK calculation underpredicts

the B(E2) value for 32Mg, it does reproduce the general trend of increasing B(E2) as the

isotopic chain becomes more neutron-rich beyond 30Mg.
1 → 0+

For 32Mg, the present result for the B(E2; 2+

1 ) value is best reproduced by the

SDPF-U-MIX and AMPGCM calculations. All the theories depicted in Fig. 5.2 that include

configuration-mixing with pf -shell orbitals reproduce the increase in B(E2; 2+

1 → 0+
is a signature of this island of inversion. Predictions differ on whether the B(E2; 2+

1 ) which
1 → 0+
1 )
value is maximized at 34Mg or 36Mg, and the experimental results for 34Mg and 36Mg are

nearly the same. A reduction of the uncertainty for the B(E2; 2+

1 ) value for 34Mg
and 36Mg would shed light on the evolution of collectivity through the Mg isotopes. The

1 → 0+

investigation of B(E2) values in Mg ultimately leads to the intriguing case of 40Mg at N = 28

where the collapse of yet another magic number may play a significant role [113]. These

B(E2) values can inform the idea that the N = 20 and N = 28 islands of inversion merge

into a continuous region of inversion along the Mg isotopes according to the SDPF-U-MIX

calculation [21].

134

5.2 The 32Mg 4+

1 State

This dissertation reported the first lifetime measurement of the 4+
0.9 ± 0.2(stat.)±0.1(syst.) ps. The corresponding large value of B(E2; 4+

1 state of 32Mg as τ (4+

1 ) =

1 → 2+

1 ) = 150(30)

e2fm4 is a signature of the collectivity of this transition. This underscores the breaking

of the N = 20 magic number to allow the coherent mixing of many configurations into

a collective state. The spin-parity assignment of the 4+

through a proton-scattering experiment [44]. The B(E2; 4+

1 state was previously established
1 → 2+

1 ) result from the present

work is consistent with the 4+ assignment and the interpretation of this band as rotational

collectivity.

Figure 5.3 shows the results of an SDPF-U-MIX calculation of the lowest 0+, 2+, and

4+ states in 32Mg with fixed 0p0h, 2p2h, and 4p4h configurations [21]. The B(E2) is

shown by the numbers between the states in e2fm4. For higher-order intruder configura-

tions, the B(E2) value increases in Fig. 5.3. The 2p2h configuration result reproduces the

B(E2; 2+

1 → 0+

1 ) result from this experiment but underestimates the B(E2; 4+

1 → 2+

1 ). The

4p4h configuration result does agree with the observed B(E2; 4+

1 → 2+

1 ). This indicates the

significance of 4p4h intruder configurations in the ground-state band of 32Mg. The success

of the 2p2h calculation for the 2+

1 → 0+

1 and the 4p4h calculation for the 4+

1 → 2+

1 may

suggest that the relative contributions of the intruder configurations changes significantly

with spin.

The ratio of B(E2) values for the 4+

1 → 2+

1 and 2+

1 → 0+

1 transitions can now be reported

for the first time with the present lifetime measurement of the 4+

1 state. The B(E2) ratio is

a tool often used to characterize many different band structures. It is complementary to the

ratio of the excited energies of the 4+

1 and 2+

1 states. The energy and B(E2) ratios for 32Mg

135

Figure 5.3: The 0+, 2+, and 4+ states and B(E2) values in 32Mg from the SDPF-U-MIX
calculation with fixed npnh configuration. Figure is from Ref. [21]. The numbers between
the states represent the B(E2) value of that transition in e2fm4.

are shown in Table 5.1 along with the predicted values for different archetypes of collective

behavior. Both the energy and the B(E2) ratio indicate that the ground-state band exhibits

ratio
1 )/E(2+
1 )
1 →2+
1 )
1 →0+
1 )

E(4+
B(E2;4+
B(E2;2+

data
2.62

1.9(4)

vibrator

symmetric rotor EKK [22]

2

2

3.33

1.43

2.55

1.37

Table 5.1: Energy ratio and B(E2) ratio observed for 32Mg and from several predictions.
The EKK predictions are from Ref. [22].

collective behavior. The B(E2) ratio has a large uncertainty which comes mostly from the

4+
1 lifetime measurement so it is not useful here to characterize the nature of the ground-

state band further. The short lifetime of the 4+

1 state is near the limit of sensitivity for the

present experimental setup. A dedicated DSAM measurement at a lower beam energy would

likely be able to measure the 4+

1 lifetime with less uncertainty. As shown in Fig. 2.2 DSAM

136

experiments are often able to measure lifetimes as short as 10 fs.

5.3 The 32Mg 0+

2 State

Using the 0+
2 → 2+
0+
B(E2; 2+

2 lifetime result of 10 ns < τ < 38 ns and the weighted average energy of the
1 transition of 170(2) keV, the reduced E2 transition probability is 28 e2fm4 <
1 → 0+
2 ) < 122 e2fm4. For physically reasonable values of ρ2(E0) the E0 branch

is expected to be less than 1% [14, 114] therefore the 0+

2 → 0+

1 transition is assumed to

be negligible. Table 5.2 summarizes the reduced E2 transition rates to 0+ states in 32Mg

and neighboring even-even nuclei which can characterize the quadrupole collectivity in these

transitions. For 30Mg, a strong B(E2) is observed for the 2+

1 → 0+

1 transition [32, 116],

B(E2) (e2fm4)

1 → 0+
2+
1 → 0+
2+

1

2

30Mg

34Si

53(7) [96]

10.9(12) [96]

17(7) [115]
61(40) [61]

32Mg 1
80(7)
48+74−20

Table 5.2: B(E2) values in 32Mg and neighboring even-even isotopes.

whereas in 34Si a strong B(E2) appears for the 2+

1 → 0+

2 transition since both the 0+

2 and

2+
1 states are considered to be dominated by the collective 2p2h configurations [61]. The

present data for 32Mg indicate that the 0+

2 state is as collective as the 0+

1 state in contrast

with the sizeable difference in transition probabilities between 0+ states in both 30Mg and

34Si. Notably, the B(E2; 2+

2 ) value in 32Mg is comparable to the strong transitions
in 30Mg and 34Si. Indeed, the present result shows E2 strength that far exceeds the value

1 → 0+

calculated by the SDPF-U-MIX model [61] which predicts B(E2; 2+

The large collectivity in the 0+
1 → 0+

1central value of B(E2; 2+

2 ) = 15 e2fm4.
2 state of 32Mg indicates prominent intruder contributions to
2 ) = 48 e2fm4 corresponds to the central lifetime value of 24 ns

1 → 0+

137

the state. However, given the large experimental uncertainties, the current B(E2; 2+

1 → 0+
2 )

result does not allow for a stringent conclusion.

5.4 Cross Section to 32Mg 0+ States

Additional direct information about the intruder contributions can be obtained from the

partial cross section to populate the 0+

2 state in comparison with theoretical calculations

assuming a pure 0p0h configuration. The 34Si 0+

1 state is predominantly of 0p0h configu-

ration [21, 32] and the two-proton removal reaction cross section is sensitive to the wave-

function overlap between the incoming projectile and outgoing residual nucleus final state,

allowing the 0p0h occupancy in the 0+

2 state of 32Mg to be quantified. Reaction calculations

were performed following the method of Ref. [57] which applies the two-neutron amplitudes

(TNA) from shell-model calculations combined with eikonal, direct reaction theory. The sup-

pression factor R2n, the ratio of experimental to calculated inclusive two-nucleon removal

cross sections, is not well-known for the (34Si,32Mg) reaction. We use the value R2n = 0.5

seen for a number of less-exotic sd -shell nuclei [57]. Assuming the essentially pure 0p0h

32Mg ground state calculated with the USDB interaction [101] is the 32Mg 0+

2 state, the 0+
2

cross section is σ2n = 0.42 mb. This value is significantly larger than the experimental upper

limit of 0.10 mb indicating the physical 0+

2 state has a reduced 0p0h occupancy. The 3-level

mixing model of Ref. [46] predicts a smaller 0p0h occupancy with a probability α2 = 0.15

for the 0+

2 state due to the sizeable 2p2h and 4p4h contributions to this state which reduces

the overlap with the 34Si 0+

1 state. By scaling the cross section obtained from the USDB

pure 0p0h calculation with the 0p0h probability α2 of the 0+

2 state, the 3-level mixing model

results in a cross section of 0.06 mb. This result is consistent with the measured partial cross

138

section, strongly suggesting that the 0+

2 state contains strong admixtures of the 2p2h and

4p4h intruder configurations.

The present result raises the question, “Where does the 0p0h-dominant 0+ state exist

in 32Mg, if anywhere?” The 3-level mixing model [46] predicts the 0+

3 state at 2.22 MeV

with the 0p0h probability α2 = 0.81. This gives the partial cross section σ2n = 0.34 mb for

the 0+

3 state, although associated events have not been experimentally observed. Using the

USDB calculations the partial cross sections for the individual states are σ(0+) = 0.42 mb,

σ(2+) = 0.94 mb and σ(4+) = 1.26 mb for R2n = 0.50. The resulting inclusive cross section

of 2.62 mb is much larger than the experimental value of 0.76(10) mb [60] as is the case

for more exotic nuclei in this mass region [117, 118]. This discrepancy indicates either that

the R2n is strongly quenched in the (34Si,32Mg) reaction where the structure is thought to

change drastically between the two nuclei or that the 0p0h components in 32Mg are widely

spread or even fragmented above the neutron separation energy (Sn = 5.778 MeV), calling

for future investigation.

139

Chapter 6

Concluding Remarks

The lifetimes of excited states in 32Mg have been studied with established experimental tech-

niques and with the novel Cascade Doppler-shift Method for the study of isomers decaying

in-flight. The 2+

1 lifetime reported in this work is τ (2+

1 ) = 18.7 ± 1.2(stat.) ± 1.1(syst.) ps

which indicates the collectivity of the ground-state band and provides a resolution to the

discrepancy among previous experimental observations. The lifetime of the 4+
1 ) =
0.9 ± 0.2(stat.) ± 0.1(syst.) ps is the first lifetime measurement for this state. The result

1 state τ (4+

confirms the spin-parity assignment of 4+ and reinforces the evidence for collectivity in the

ground-state band of 32Mg along the traditional N = 20 magic number. The new CDM

method to study isomeric states decaying in-flight was used to observe the 0+

2 → 2+

1 transi-

tion at 170(2) keV and constrain the lifetime to 10 ns < τ (0+

2 ) < 38 ns. The result indicates

significant collectivity in the 0+

2 state.

The experimental results of this work can support interesting perspectives about the

32Mg nuclide. The precise B(E2; 2+

1 → 0+

1 ) result reported here is slightly lower than the

weighted average of the results of past experiments. This suggests that the collectivity in the

ground-state band is slightly less than it previously was believed to be. Given the evidence

for collectivity in the 0+

2 state, a new picture of 32Mg emerges: the collectivity once thought

to dominate only the ground-state band is actually well-mixed with the band beginning with

the 0+

2 state, which is more collective than once believed.

140

The available data suggests significant collectivity in both the 0+

1 and 0+

2 states of 32Mg.

It seems likely that the 2p2h and 4p4h configurations are both significant components of

the 0+

1 and 0+

2 states of 32Mg given the low cross section observed for these states in the
9Be(34Si,32Mg)X reaction. The SDPF-U-MIX calculation shown in Ref. [45] indicates that

the band built upon the 0+

1 state becomes more dominated by 4p4h than 2p2h configurations

as spin increases. The 0+

2 state could be the beginning of a band that begins with nearly

equal admixtures of 2p2h and 4p4h configurations but then evolves with increasing spin to

have more 2p2h component. Of course, it is possible for the band above the 0+

2 state to

become dominated by 0p0h configurations at higher spin. This would be reminiscent of the

original observation of the 0+

2 state which suggested the 0+

2 state is based on an almost pure

sd configuration [14]. The work presented here appears to rule out the dominance of the

0p0h configuration in the 0+

2 state but cannot support any claim about possible higher-lying

states in the 0+

2 band. States with prompt decays to the 0+

2 state would not be observed

in the CDM analysis of this experiment because they do not decay at the same location as

the 0+

2 state due to the long 0+

2 lifetime. Experimental data on the hypothetical 2+ and 4+

members of the band built upon the 0+

2 may provide crucial data on the configuration-mixing

of that band. For now however one must rely on shell-model calculations or other theoretical

approaches.

As mentioned earlier, the branching ratio of the 0+

2 → 0+

1 E0 transition in 32Mg is

expected to be less than 1% [14, 114]. The observation of this E0 transition rate would be

valuable to help understand the wave functions of the 0+

1 and 0+

2 states (see, for instance,

Ref. [119]).

The nature of 32Mg is an intriguing topic due to the significant admixture of the normal

0p0h configuration and the intruder 2p2h and 4p4h configurations [46]. For instance, within

141

a single band complex changes in the admixture of the intruder configurations may take

place as the nucleus increases in spin [45]. The complications in 32Mg may signal similar

effects in other island of inversion nuclei or neutron-rich regions.

In closing, this work reported the lifetimes of the 2+

2 states in 32Mg which

1 , 4+

1 and 0+

cover a wide range of values, requiring the use of three different analysis techniques. The

B(E2) results from the lifetime measurement indicate collectivity in the ground-state band

and the 0+

2 state. The novel CDM technique introduced here proved indispensable to observe

the 0+

2 state. As rare-isotope beams with high energies continue to be powerful experimental

tools, the CDM will be vital to extend the sensitive lifetime range of in-beam experiments.

142

BIBLIOGRAPHY

143

BIBLIOGRAPHY

[1] National Nuclear Data Center www.nndc.bnl.gov.

[2] B. A. Brown, Lecture Notes on Nuclear Structure (2017).

[3] R. F. Casten, Nuclear Structure from a Simple Perspective, 2nd ed. (Oxford University

Press, 2000).

[4] R. D. Woods and D. S. Saxon, Phys. Rev. 95, 577 (1954).

[5] M. G. Mayer, Phys. Rev. 75, 1969 (1949).

[6] O. Haxel, J. H. D. Jensen, and H. E. Suess, Phys. Rev. 75, 1766 (1949).

[7] A. Bohr and B. R. Mottelson, Nuclear Structure (W. A. Benjamin, 1969).

[8] V. F. Weisskopf, Phys. Rev. 83, 1073 (1951).

[9] B. Pritychenko, M. Mirch, B. Singh, and M. Horoi, Atomic Data and Nuclear Data

Tables 107, 1 (2016).

[10] M. G. Mayer, Phys. Rev. 74, 235 (1948).

[11] O. Sorlin and M.-G. Porquet, Prog. Part. Nucl. Phys. 61, 602 (2008).

[12] A. Navin, D. W. Anthony, T. Aumann, T. Baumann, D. Bazin, Y. Blumenfeld, B. A.
Brown, T. Glasmacher, P. G. Hansen, R. W. Ibbotson, P. A. Lofy, V. Maddalena,
K. Miller, T. Nakamura, B. V. Pritychenko, B. M. Sherrill, E. Spears, M. Steiner,
J. A. Tostevin, J. Yurkon, and A. Wagner, Phys. Rev. Lett. 86, 266 (2000).

[13] T. Motobayashi, Y. Ikeda, Y. Ando, K. Ieki, M. Inoue, N. Iwasa, T. Kikuchi,
M. Kurokawa, S. Moriya, S. Ogawa, H. Murakami, S. Shimoura, Y. Tanagisawa,
T. Nakamura, Y. Watanabe, M. Ishihara, T. Teranishi, H. Okuno, and R. F. Cas-
ten, Phys. Lett. B 346, 9 (1995).

[14] K. Wimmer, T. Kr¨oll, R. Kr¨ucken, V. Bildstein, R. Gernh¨auser, B. Bastin, N. Bree,
J. Diriken, P. V. Duppen, M. Huuyse, N. Patronis, P. Vermaelen, D. Voulot, J. V.
de Walle, F. Wenander, L. M. Fraile, R. Chapman, B. Hadinia, R. Orlandi, J. F. Smith,
R. Lutter, P. G. Thirolf, M. Labiche, A. Blazhev, M. Kalk´uhler, P. Reiter, M. Seidlitz,
N. Warr, A. O. Macchiavelli, H. B. Jeppesen, E. Fiori, G. Gerigiev, G. Schirieder, S. D.
Gupta, G. L. Bianco, S. Nardelli, J. Butterworth, J. Johansen, and K. Riisager, Phys.
Rev. Lett. 105, 252501 (2010).

[15] C. D´etraz, D. Guillemaud, G. Huber, R. Klapisch, M. Langevin, F. Naulin, C. Thibault,

and L. Carraz, Phys. Rev. C 19, 164 (1979).

144

[16] B. Bastin, S. Gr´evy, D. Sohler, O. Sorlin, Z. Dombr´adi, N. L. Achouri, J. C. Ang´elique,
F. Azaiez, D. Baiborodin, R. Borcea, C. Bourgeois, A. Buta, A. B¨urger, R. Chapman,
J. C. Dalouzy, Z. Dlouhy, A. Drouard, Z. Elekes, S. Franchoo, S. Iacob, B. Laurent,
M. Lazar, X. Liang, E. Li´enard, J. Mrazek, L. Nalpas, F. Negoita, N. A. Orr, Y. Pe-
nionzhkevich, Z. Podoly´ak, F. Pougheon, P. Roussel-Chomaz, M. G. Saint-Laurent,
M. Stanoiu, I. Stefan, F. Nowacki, and A. Poves, Phys. Rev. Lett. 99, 022503 (2007).

[17] A. Gade, R. V. F. Janssens, T. Baugher, D. Bazin, B. A. Brown, M. P. Carpenter, C. J.
Chiara, A. N. Deacon, S. J. Freeman, G. F. Grinyer, C. R. Hoffman, B. P. Kay, F. G.
Kondev, T. Lauritsen, S. McDaniel, K. Meierbachtol, A. Ratkiewicz, S. R. Stroberg,
K. A. Walsh, D. Weisshaar, R. Winkler, and S. Zhu, Phys. Rev. C 81, 051304(R)
(2010).

[18] B. A. Brown, Physics 3, 104 (2010).

[19] S. M. Lenzi, F. Nowacki, A. Poves, and K. Sieja, Phys. Rev. C 82, 054301 (2010).

[20] F. Nowacki, A. Poves, E. Caurier, and B. Bounthong, Phys. Rev. Lett. 117, 272501

(2016).

[21] E. Caurier, F. Nowacki, and A. Poves, Phys. Rev. C 90 (2014).

[22] N. Tsunoda, T. Otsuka, N. Shimizu, M. Hjorth-Jensen, K. Takayanagi, and T. Suzuki,

Phys. Rev. C 95, 021304(R) (2017).

[23] T. Otsuka, T. Suzuki, J. D. Holt, A. Schwenk, and Y. Akaishi, Phys. Rev. Lett. 105,

032501 (2010).

[24] C. Thibault, R. Klapisch, C. Rigaud, A. M. Poskanzer, R. Prieels, L. Lessard, and

W. Reisdorf, Phys. Rev. C 12, 644 (1975).

[25] E. K. Warburton, J. A. Becker, and B. A. Brown, Phys. Rev. C 41, 1147 (1990).

[26] B. H. Wildenthal and W. Chung, Phys. Rev. C 22, 2260 (1980).

[27] C. D´etraz, M. Langevin, M. C. Goffri-Kouassi, D. Guillemaud, M. Epherre, G. Audi,

C. Thibault, and F. Touchard, Nucl. Phys. A 394, 378 (1983).

[28] X. Campi, H. Flocard, A. Kerman, and S.Koonin, Nucl. Phys. A 251, 193 (1975).

[29] A. Poves and J. Retamosa, Phys. Lett. B 184, 311 (1987).

[30] A. Poves and J. Retamosa, Nucl. Phys. B 571, 221 (1994).

[31] N. Fukunishi, T. Otsuka, and T. Sebe, Phys. Lett. B 296, 279 (1992).

[32] Y. Utsuno, T. Otsuka, T. Mizusaki, and M. Honma, Phys. Rev. C 60, 054315 (1999).

[33] A. Poves, E. Caurier, F. Nowacki, and K. Sieja, Phys. Scr. T150, 014030 (2012).

145

[34] A. Rodr´ıguez-Guzm´an, J. L. Egido, and L. M. Robledo, Phys. Rev. C 62, 054319

(2000).

[35] A. Rodr´ıguez-Guzm´an, J. L. Egido, and L. M. Robledo, Nucl. Phys. A 709, 201 (2002).

[36] B. V. Pritychenko, T. Glasmacher, P. D. Cottle, M. Fauerback, R. W. Ibbotson, K. W.
Kemper, V. Maddalena, A. Navin, R. Ronningen, A. Sakharuk, H. Scheit, and V. G.
Zelevinsky, Phys. Lett. B 461, 322 (1999).

[37] V. Chist´e, A. Gillibert, A. Le´epine-Szily, N. Alamanos, F. Auger, J. Barrette, F. Braga,
M. D. Cortina-Gil, Z. Dlouhy, V. Lapoux, M. Lewitowicz, R. Lichtenth¨aler, R. L. Neto,
S. M. Lukyanov, M. MacCormick, F. Marie, W. Mittig, F. de Oliveira Santos, N. A.
Orr, A. N. Ostrowski, S. Ottini, A. Pakou, Y. E. Penionzhkevich, P. Roussel-Chomaz,
and J. L. Sida, Phys. Lett. B 514, 233 (2001).

[38] H. Iwasaki, T. Motobayashi, H. Sakurai, K. Yoneda, T. Gomi, N. Aoi, N. Fukuda,
Z. F¨ul¨op, U. Futakami, Z. Gacsi, Y. Higurashi, N. Imai, N. Iwasa, T. Kubo, M. Kunibu,
M. Kurokawa, Z. Liu, T. Minemura, A. Saito, M. Serata, S. Shimoura, S. Takeuchi,
Y. X. Watanabe, K. Yamada, Y. Yanagisawa, and M. Ishihara, Phys. Lett. B 522, 227
(2001).

[39] J. A. Church, C. M. Campbell, D.-C. Dinca, J. Enders, A. Gade, T. Glasmacher,
Z. Hu, R. V. F. Janssens, W. F. Mueller, H. Olliver, B. C. Perry, L. A. Riley, and K. L.
Yurkewicz, Phys. Rev. C 72, 054320 (2005).

[40] K. Li, Y. Ye, T. Motobayashi, H. Scheit, P. Doornenbal, S. Takeuchi, N. Aoi, M. Mat-

sushita, E. Takeshita, D. Pang, and H. Sakurai, Phys. Rev. C 92, 014608 (2015).

[41] H. Mach, L. M. Fraile, O. Tengblad, R. Boutami, C. Jollet, W. A. Pl´ociennik, D. T. Yor-
danov, M. Stanoiu, M. J. G. Borge, P. A. Butler, J. Cedark¨all, P. Fogelberg, H. Fynbo,
P. Hoff, A. Jokinen, A. Korgul, U. K¨oster, W. Kurcewicz, F. Marechal, T. Motobayashi,
J. Mrazek, G. Neyens, T. Nilsson, S. Pedersen, A. Poves, B. Rubio, E. Ruchowska, and
the ISOLDE Collaboration, Eur. Phys. J. A s01, 105 (2005).

[42] D. Guillemaud-Mueller, C. D´etraz, M. Langevin, F. Naulin, M. de Saint-Simon,

C. Thibault, and F. Touchard, Nucl. Phys. A 426, 37 (1984).

[43] G. Klotz, P. Baumann, M. Bounajma, A. Huck, A. Knipper, G. Walter, G. Marguier,

C. Richard-serre, A. Poves, and J. Retamosa, Phys. Rev. C 47, 2502 (1993).

[44] S. Takeuchi, N. Aoi, T. Motobayashi, S. Ota, E. Takeshita, H. Suzuki, H. Baba,
T. Fukui, Y. Hashimoto, K. Ieki, N. Imai, H. Iwasaki, S. Kanno, Y. Kondo, T. Kubo,
K. Kurita, T. Minemura, T. Nakabayashi, T. Nakamura, T. Okumura, T. K. Onishi,
H. Sakurai, S. Shimoura, R. Sugou, D. Suzuki, M. K. Suzuki, M. Takashina, M. Tamaki,
K. Tanaka, Y. Togano, and K. Yamada, Phys. Rev. C 79, 054319 (2009).

[45] H. L. Crawford, P. Fallon, A. O. Macchiavelli, A. Poves, V. M. Bader, D. Bazin,
M. Bowry, C. M. Campbell, M. P. Carpenter, R. M. Clark, M. Cromaz, A. Gade,

146

E. Ideguchi, H. Iwasaki, C. Langer, I. Y. Lee, C. Loelius, E. Lunderberg, C. Morse,
A. L. Richard, J. Rissanen, D. Smalley, S. R. Stroberg, D. Weisshaar, K. Whitmore,
A. Wiens, S. J. Williams, K. Wimmer, and T. Yamamato, Phys. Rev. C 93, 031303(R)
(2016).

[46] A. O. Macchiavelli, H. L. Crawford, C. M. Campbell, R. M. Clark, M. Cromaz, P. Fal-
lon, M. D. Jones, I. Y. Lee, M. Salathe, B. A. Brown, and A. Poves, Phys. Rev. C 94,
051303(R) (2016).

[47] A. O. Macchiavelli and H. L. Crawford, Phys. Scr. 92, 064001 (2017).

[48] R. D. Evans, The Atomic Nucleus, 1st ed. (Tata McGraw-Hill Publishing Company

Ltd., 1995).

[49] G. F. Knoll, Radiation Detection and Measurement, 4th ed. (Wiley, 2010).

[50] D. Weisshaar, D. Bazin, P. C. Bender, C. M. Campbell, F. Recchia, V. Bader,
T. Baugher, J. Belarge, M. P. Carpenter, H. L. Crawford, M. Cromaz, B. Elman,
P. Fallon, A. Forney, A. Gade, J. Harker, N. Kobayashi, C. Langer, T. Lauritsen,
I. Y. Lee, A. Lemasson, B. Longfellow, E. Lunderberg, A. O. Macchiavelli, K. Miki,
S. Momiyama, S. Noji, D. C. Radofrd, M. Scott, J. Sethi, S. R. Stroberg, Z. Sullivan,
R. Titus, A. Wiens, S. Williams, K. Wimmer, and S. Zhu, Nucl. Instrum Methods
Phys. Res., Sect. A 847, 187 (2017).

[51] A. Lemasson, H. Iwasaki, C. Morse, D. Bazin, T. Baugher, J. S. Berryman, A. Dewald,
C. Fransen, A. Gade, C. McDaniel, A. Nichols, A. Ratkiewicz, S. Stroberg, P. Voss,
R. Wadsworth, D. Weisshaar, K. Wimmer, and R. Winkler, Phys. Rev. C 85, 041303
(2012).

[52] P. J. Nolan and J. F. Sharpey-Schafer, Rep. Prog. Phys. 42, 1 (1979).

[53] A. Dewald, O. M¨oller, and P. Petkov, Prog. Part. Nucl. Phys. 67, 786 (2012).

[54] S. Shimoura, A. Saito, T. Minemura, Y. U. Matsuyama, H. Baba, H. Akiyoshi, N. Aoi,
T. Gomi, Y. Higurashi, K. Ieki, N. Imai, N. Iwasa, H. Iwasaki, S. Kanno, S. Kubono,
M. Kunibu, S. Michimasa, T. Motobayashi, T. Nakamura, H. Sakurai, M. Serata,
E. Takeshita, S. Takeuchi, T. Teranishi, K. Ue, K. Yamada, Y. Yanagisawa, M. Ishi-
hara, and N. Itagaki, Phys. Lett. B 560, 31 (2003).

[55] R. Elder, H. Iwasaki, J. Ash, D. Bazin, P. C. Bender, T. Braunroth, B. A. Brown,
C. M. Campbell, H. L. Crawford, B. Elman, A. Gade, M. Grinder, N. Kobayashi,
B. Longfellow, A. O. Macchiavelli, T. Mijatovi´c, J. Pereira, A. Revel, D. Rhodes, J. A.
Tostevin, and D. Weisshaar, Phys. Rev. C 100, 041301 (2019).

[56] P. G. Hansen and J. A. Tostevin, Annu. Rev. Nucl. Part. Sci. 53, 219 (2003).

[57] J. A. Tostevin and B. A. Brown, Phys. Rev. C 74, 064604 (2006).

[58] P. G. Hansen and B. M. Sherrill, Nucl. Phys. A 693, 133 (2001).

147

[59] A. Mutschler, A. Lemasson, O. Sorlin, D. Bazin, C. Borcea, R. Borcea, Z. Dombr´adi,
J.-P. Ebran, A. Gade, H. Iwasaki, E. Khan, A. Lepailleur, F. Recchia, T. Roger,
F. Rotaru, D. Sohler, M. Stanoiu, S. R. Stroberg, J. A. Tostevin, M. Vandebrouck,
D. Weisshaar, and K. Wimmer, Nature 13, 152 (2016).

[60] D. Bazin, B. A. Brown, C. M. Campbell, J. A. Church, D. C. Dinca, J. Enders, A. Gade,
T. Glasmacher, P. G. Hansen, W. F. Mueller, H. Olliver, B. C. Perry, B. M. Sherrill,
J. R. Terry, and J. A. Tostevin, Phys. Rev. Lett. 91, 012501 (2003).

[61] F. Rotaru, F. Negoita, S. Gr´evy, J. Mrazek, S. Lukyanov, F. Nowacki, A. Poves,
O. Sorlin, C. Borcea, R. Borcea, A. Buta, L. C´aceres, S. Calinescu, R. Chevrier,
Z. Dombr´adi, J. M. Daugas, D. Lebhertz, Y. Penionzhevich, C. Petrone, D. Sohler,
M. Stanoiu, and J. C. Thomas, Phys. Rev. Lett. 109, 092503 (2012).

[62] P. A. Zavodszky, B. Arend, D. Cole, J. DeKamp, G. Machinoane, F. Marti, P. Miller,
J. Moskalik, J. Ottarson, J. Vincent, and A. Zeller, Nucl. Instrum. and Methods B
241, 959 (2005).

[63] D. J. Morrissey, Nucl. Phys. A 616, 45c (1997).

[64] B.-M. Sherrill, Progress of Theoretical Physics Supplement 146, 60 (2002).

[65] D. J. Morrissey, B. M. Sherrill, M. Steiner, A. Stolz, and I. Wiedenhoever, Nucl.

Instrum. Methods Phys. Res., Sect. B 204, 90 (2003).

[66] D. Bazin, J. A. Caggiano, B. M. Sherrill, J. Yurkon, and A. Zeller, Nucl. Instrum.

Methods Phys. Res., Sect. B 204, 629 (2003).

[67] S. Paschalis, I.Y.Lee, A. Macchiavelli, C. M. Campbell, M. Cromaz, S. Gros, J. Pavan,
J. Qian, R. M. Clark, H. L. Crawford, D. Doering, P. Fallon, C. Lionberger, T. Loew,
M. Petri, T. Stezelberger, S. Zimmermann, D. C. Radford, K. Lagergren, D. Weisshaar,
R. Winkler, T. Glasmacher, J. T. Anderson, and C. W. Beausang, Nucl. Instrum.
Methods 709, 44 (2013).

[68] I. Brown, The Physics and Technology of Ion Sources, 2nd ed. (Wiley-VCH, 2004).

[69] E. O. Lawrence and M. S. Livingston, Phys. Rev. 40, 19 (1932).

[70] J. P. Dufour, R. D. Moral, H. Emmermann, F. Hubert, D. Jean, C. Poinot, M. S.

Pravikoff, and A. Fleury, Nucl. Instrum. and Methods Sect. A 248, 267 (1986).

[71] S800 Spectrograph wiki , https://wikihost.nscl.msu.edu/S800Doc/ (2020), ac-

cessed: 2020-05-09.

[72] J. Yurkon, D. Bazin, W. Benenson, D. J. Morrissey, B. M. Sherrill, D. Swan, and

R. Swanson, Nucl. Instrum. Methods, Sect. B 422, 291 (1999).

[73] K. Meierbachtol, D. Bazin, and D. J. Morrissey, Nucl. Instrum. Methods, Sect. A 652,

668 (2011).

148

[74] K. Wimmer, D. Barofsky, D. Bazin, L. M. Fraile, J. Lloyd, J. R. Tompkins, and S. J.

Williams, Nucl. Instrum. Methods Sect. A 769, 65 (2015).

[75] M. Berz, K. Joh, J. A. Nolen, B. M. Sherrill, and A. F. Zeller, Phys. Rev. C 47, 537

(1993).

[76] M. Robinson, P. Halse, W. Trinder, R. Anne, C. Borcea, M. Lewitowicz, S. Lukyanov,
M. Mirea, Y. Oganessian, N. A. Orr, Y. Peniozhkevich, M. G. Saint-Laurent, and
O. Tarasov, Phys. Rev. C 53, R1465 (1996).

[77] M. Cromaz, V. J. Riot, P. Fallon, S. Gros, B. Holmes, I. Y. Lee, A. O. Macchiavelli,

C. Vu, H. Yaver, and S. Zimmermann, Nucl. Instrum. Methods A 597, 233 (2008).

[78] J. Anderson, R. Brito, D. Doering, T. Hayden, B. Holmes, J. Joseph, H. Yaver, and

S. Zimmermann, IEEE Trans. Nucl. Sci. NS-56, 258 (2009).

[79] C. Loelius, H. Iwasaki, B. A. Brown, M. Honma, V. M. Bader, T. Baugher, D. Bazin,
J. S. Berryman, T. Braunroth, C. M. Campbell, A. Dewald, A. Gade, N. Kobayashi,
C. Langer, I. Y. Lee, A. Lemasson, E. Lunderberg, C. Morse, F. Recchia, D. Smalley,
S. R. Stroberg, R. Wadsworth, C. Walz, D. Weisshaar, A. Westerberg, K. Whitmore,
and K. Wimmer, Phys. Rev. C 94, 024340 (2016).

[80] National Institute of Standards and Technology X-ray Mass Attenuation Coeffi-
cients, https://physics.nist.gov/PhysRefData/XrayMassCoef/tab3.html (2004),
accessed: 2004-03-29.

[81] H. Iwasaki, A. Dewald, T. Braunroth, C. Fransen, D. Smalley, A. Lemasson, C. Morse,
K. Whitmore, and C. Loelius, Nucl. Instrum. Methods Phys. Res., Sect. A 806, 123
(2016).

[82] T. K. Alexander and K. W. Allen, Can. J. Phys. 43, 22 (1965).

[83] A. B. T. K. Alexander, Nucl. Instrum. Methods 81, 22 (1970).

[84] A. Dewald, S. Harissopulos, and P. von Brentano, Zeitschrift f¨ur Physik 334, 163

(1989).

[85] P. Bednarczyk, J. Gr¸ebosz, M. Kmiecik, A. Maj, W. M¸eczy´nski, S. Myalski, J. Stycze´n,
C. Domingo-Pardo, P. Doornenbal, J. Gerl, M. G´orska, H. J. Wollersheim, J. Jolie,
P. Reiter, A. Bracco, and F. Camera, Acta Physica Polonica B 41, 505 (2010).

[86] Physik

Instrumente, N-381 NEXACT Linear Actuator,

physikinstrumente.com/en/products/linear-actuators/,
13.

https://www.
2020-02-

accessed:

[87] TESA,

Wireless

probe,

https://tesatechnology.com/en-us/products/

measuring-probes-and-display-units/usb-or-wireless-measuring-probes/
wireless-probe_p40763.htm, accessed: 2020-02-13.

149

[88] C. Ouellet and B. Singh, Nucl. Data Sheets 112, 2199 (Data extracted from the ENSDF

online database June, 2019) (2011).

[89] H. Nishibata, T. Shimoda, A. Odahara, S. Morimoto, S. Kanaya, A. Yagi, H. Kanaoka,

M. R. Pearson, C. D. P. Levy, and M. Kimura, Phys. Lett. B 767, 81 (2017).

[90] R. Fox, SpecTcl Home page , http://docs.nscl.msu.edu/daq/spectcl/ (2003), ac-

cessed: 2020-04-04.

[91] K. Wimmer and E. Lunderberg, Nuclear Structure and Reactions Software, http://

nucl.phys.s.u-tokyo.ac.jp/wimmer/software.php (2016), accessed: 2020-04-04.

[92] S. Agostinelli, J. Allison, K. Amako, J. Apostolakis, H. Araujo, P. Arce, M. Asai,
D. Axen, S. Banerjee, G. Barrand, F. Behner, L. Bellagamba, J. Boudreau, L. Broglia,
A. Brunengo, H. Burkhardt, S. Chauvie, J. Chuma, R. .Chytracek, G. Cooperman,
G. Cosmo, P. Degtyarenko, A. Dell’Acqua, G. Depaola, D. Dietrich, R. Enami, A. Fe-
liciello, C. Ferguson, H. Fesefeldt, G. Folger, F. Foppiano, A. Forti, S. Garelli, S. Giani,
R. Giannitrapani, D. Gibin, J. J. G. Cadenas, I. Gonz´alez, G. G. Abril, G. Greeni-
aus, W. Greiner, V. Grichine, A. Grossheim, S. Guatelli, P. Gumplinger, R. Hamatsu,
K. Hashimoto, H. Hasui, A. Heikkinen, A. Howard, V. Ivanchenko, A. Johnson, F. W.
Jones, J. Kallenbach, N. Kanaya, M. Kawabata, Y. Kawabata, M. Kawaguti, S. Kelner,
P. Kent, A. Kimura, T. Kodama, R. Kokoulin, M. Kossov, H. Kurashige, E. Lamanna,
T. Lamp´en, V. Lara, V. Lefebure, F. Lei, M. Liendl, W. Lockman, F. Longo, S. Magni,
M. Maire, E. Medernach, K. Minamimoto, P. M. de Freitas, Y. Morita, K. Murakami,
M. Nagamatu, R. Nartallo, P. Nieminen, T. Nishimura, K. Ohtsubo, M. Okamura,
S. O’Neale, Y. Oohata, K. Paech, J. Perl, A. Pfeiffer, M. G. Pia, F. Ranjard, A. Ry-
bin, S. Sadilov, E. D. Salvo, G. Santin, T. Sasaki, N. Savvas, Y. Sawada, S. Scherer,
S. Sei, V. Sirotenko, D. Smith, N. Starkov, H. Stoecker, J. Sulkimo, M. Takahata,
S. Tanaka, E. Tcherniaev, E. S. Tehrani, M. Tropeano, P. Truscott, H. Uno, L. Urban,
P. Urban, M. Verderi, A. Walkden, W. Wander, H. Weber, J. P. Wellisch, T. Wenaus,
D. C. Williams, D. Wright, T. Yamada, H. Yoshida, and D. Zschiesche, Nucl. Instrum.
Methods Phys. Res., Sect. A 506, 250 (2003).

[93] NSCL lifetime group webpage, https://wikihost.nscl.msu.edu/lifetime/, ac-

cessed: 2020-04-04.

[94] S. Baker and R. D. Cousins, Nucl. Instrum. Methods 221, 437 (1984).

[95] K. Wimmer, Normal and Intruder Configurations in the Island of Inversion, presenta-

tion at the Direct Reactions with Exotic Beams conference (2016).

[96] M. S. Basunia, Nucl. Data Sheets 111, 2331 (Data extracted from the ENSDF online

database July, 2019 (2010).

[97] A. J. Rutten, A. Holthuizen, J. A. G. D. Raedt, W. A. Sterrenburg, and G. V. Mid-

delkoop, Nucl. Phys. A 344, 294 (1980).

150

[98] Z. Berant, C. Broude, G. Engler, M. Hass, R. Levy, and B. Richter, Nucl. Phys. A

243, 519 (1975).

[99] J. A. J. Hermans, G. A. P. Engelbertink, L. P. Ekstr¨om, H. H. Eggenhuisen, and

M. A. V. Driel, Nucl. Phys. A 284, 307 (1977).

[100] T. Lauritsen, A. Korichi, S. Zhu, A. N. Wilson, D. Weisshaar, J. Dudouet, A. D.
Ayangeakaa, M. P. Carpenter, C. M. Campbell, E. Cl´ement, H. L. Crawford, M. Cro-
maz, P. Fallon, J. P. Greene, R. V. F. Janssens, T. L. Khoo, N. Lalovi´c, I. Y. Lee,
A. O. Macchiavelli, R. M. Perez-Vidal, S. Pietri, D. C. Radford, D. Ralet, L. A. Riley,
D. Seweryniak, and O. Stezowski, Nucl. Instrum. Methods Phys. Res., Sect. A 836,
46 (2016).

[101] B. A. Brown and W. A. Richter, Phys. Rev. C 74, 034315 (2006).

[102] Goodfellow, http://www.goodfellow.com/ (2020), accessed: 2020-03-09.

[103] S. K. Ashour and M. A. Abdel-hameed, Journal of Advanced Research 1, 341 (2010).

[104] M. S. Basunia, Nucl. Data Sheets 114, 1189 (Data extracted from the ENSDF online

database June, 2019) (2013).

[105] N. Nica and B. Singh, Nucl. Data Sheets 113, 1563 (Data extracted from the ENSDF

online database June, 2019) (2012).

[106] P. Doornenbal, H. Scheit, S. Takeuchi, N. Aoi, K. Li, M. Matsushita, D. Steppenbeck,
H. Wang, H. Baba, E. Ideguchi, N. Kobayashi, Y. Konda, J. Lee, S. Michimasa,
T. Motobayashi, A. Poves, H. Sakurai, M. Takechi, Y. Togano, and K. Yoneda, Phys.
Rev. C 93, 044306 (2016).

[107] B. A. Brown and B. H. Wildenthal, Ann. Rev. Nucl. Part. Sci. 38, 29 (1988).

[108] T. T. S. Kuo and G. E. Brown, Nucl. Phys. A 114, 241 (1968).

[109] J. Decharg´e and D. Gogny, Phys. Rev. C 21, 1568 (1980).

[110] T. T. S. Kuo, S. Y. Lee, and K. F. Ratcliff, Nucl. Phys. A 176, 65 (1971).

[111] E. M. Krenciglowa and T. T. S. Kuo, Nucl. Phys. A 235, 171 (1974).

[112] E. M. Krenciglowa, T. T. S. Kuo, E. Osnes, and P. J. Ellis, Nucl. Phys. A 289, 381

(1977).

[113] H. L. Crawford, P. Fallon, A. O. Macchiavelli, P. Doornenbal, N. Aoi, F. Browne, C. M.
Campbell, S. Chen, R. M. Clark, M. L. Cort´es, M. Cromaz, E. Ideguchi, M. D. Jones,
R. Kanungo, M. MacCormick, S. Momiyama, I. Murray, M. Niikura, S. Paschalis,
M. Petri, H. Sakurai, M. Salathe, P. Schrock, D. Steppenbeck, S. Takeuchi, Y. K.
Tanaka, R. Taniuchi, H. Wang, and K. Wimmer, Phys. Rev. Lett. 122, 052501 (2019).

151

[114] T. Kib´edi, T. W. Burrows, M. B. Trzhaskovskaya, P. M. Davidson, and C. W. N. Jr.,

Nucl. Instrum. Methods Phys. Res., Sect. A 589, 202 (2008).

[115] R. W. Ibbotson, T. Glasmacher, B. A. Brown, L. Chen, M. J. Chromik, P. D. Cottle,
M. Fauerbach, K. W. Kemper, D. J. Morrissey, H. Scheit, and M. Thoennessen, Phys.
Rev. Lett. 80, 2081 (1998).

[116] O. Neidermaier, H. Scheit, V. Bildstein, H. Boie, J. Figging, R. von Hahn, F. K¨ock,
M. Lauer, U. K. Pal, H. Podlech, R. Repnow, D. Schwalm., C. Alvarez, F. Ames,
G. Bollen, S. Emhofer, D. Habs, O. Kester, R. Lutter, K. Rudolph, M. Pasini, P. G.
Thirolf, B. H. Wolf, J. Eberth, G. Gersch, H. Hess, P. Reiter, O. Thelen, N. Warr,
D. Weisshaar, F. Aksouh, P. V. den Bergh, P. V. Duppen, M. Huyse, O. Ivanov,
P. Mayet, J. V. de Walle, J. ¨Ayst¨o, P. A. Butler, J. Cederk¨all, P. Delahaye, H. O. U.
Fynbo, L. M. Fraile, O. Forstner, S. Franchoo, U. K¨oster, T. Nilsson, M. Oinonen,
T. Sieber, F. Wenander, M. Pantea, A. Richter, G. Schreider, H. Simon, T. Behrens,
R. Gernh¨auser, T. Kr¨oll, Kr¨ucken, M. M¨unch, T. Davison, J. Gerl, G. Huber, A. Hurst,
J. Iwanicki, B. Jonson, P. Lieb, L. Liljeby, A. S. nad A. Scherillo, P. Schmidt, and
G. Walter, Phys. Rev. Lett. 94, 172501 (2005).

[117] P. Fallon, E. Rodriguez-Vieitez, A. O. Macchiavelli, A. Gade, J. A. Tostevin, P. Adrich,
D. Bazin, M. Bowen, C. M. Campbell, R. M. Clark, J. M. Cook, M. Cromaz,
D. C. Dinca, T. Glasmacher, I. Y. Lee, S. McDaniel, W. F. Mueller, S. G. Prussin,
A. Ratkiewicz, K. Siwek, J. R. Terry, D. Weisshaar, M. Wiedeking, K. Yoneda, B. A.
Brown, T. Otsuka, and Y. Utsuno, Phys. Rev. C 81, 041302(R) (2010).

[118] I. Murray, M. MacCormick, D. Bazin, P. Doornenbal, N. Aoi, H. Baba, H. Craw-
ford, P. Fallon, K. Li, J. Lee, M. Matsushita, T. Motobayashi, T. Otsuka, H. Saku-
rai, H. Scheit, D. Steppenbeck, S. Takeuchi, J. A. Tostevin, N. Tsunoda, Y. Utsuno,
H. Wang, and K. Yoneda, Phys. Rev. C 99, 011302(R) (2019).

[119] B. P. Crider, C. J. Prokop, S. N. Liddick, M. Al-Shudifat, A. D. Ayangeakaa, M. P.
Carpenter, J. J. Carroll, J. Chen, C. J. Chiara, H. M. David, A. C. Dombos, S. Go,
R. Grzywacz, J. Harker, R. V. F. Janssens, N. Larson, T. Lauritsen, R. Lew, S. J.
Quinn, F. Recchia, A. Spyrou, W. B. Walters, and S. Zhu, Phys. Lett. B 763, 108
(2016).

152