Quadratic Zeeman Spectral Diusion of Thulium Ion Population in a Yttrium Gallium Garnet Crystal Jacob H. Davidson1Antariksha Das1yNir Alfasi1Rufus L. Cone2Charles W. Thiel2and Wolfgang Tittel1 3 4

2025-05-02 0 0 5.21MB 12 页 10玖币
侵权投诉
Quadratic Zeeman Spectral Diffusion of Thulium Ion Population in a Yttrium
Gallium Garnet Crystal
Jacob H. Davidson,1, Antariksha Das,1, Nir Alfasi,1Rufus L. Cone,2Charles W. Thiel,2and Wolfgang Tittel1, 3, 4
1QuTech and Kavli Institute of Nanoscience, Delft University of Technology, 2628 CJ Delft, The Netherlands
2Department of Physics, Montana State University, Bozeman, Montana 59717, USA
3Department of Applied Physics, University of Geneva, 1211 Geneva 4, Switzerland
4Schaffhausen Institute of Technology - SIT, 1211 Geneva 4, Switzerland
(Dated: October 12, 2022)
The creation of well understood structures using spectral hole burning is an important task in the
use of technologies based on rare earth ion doped crystals. We apply a series of different techniques
to model and improve the frequency dependent population change in the atomic level structure of
Thulium Yttrium Gallium Garnet (Tm:YGG). In particular we demonstrate that at zero applied
magnetic field, numerical solutions to frequency dependent three-level rate equations show good
agreement with spectral hole burning results. This allows predicting spectral structures given a
specific hole burning sequence, the underpinning spectroscopic material properties, and the relevant
laser parameters. This enables us to largely eliminate power dependent hole broadening through the
use of adiabatic hole-burning pulses. Though this system of rate equations shows good agreement
at zero field, the addition of a magnetic field results in unexpected spectral diffusion proportional to
the induced Tm ion magnetic dipole moment and average magnetic field strength, which, through
the quadratic Zeeman effect, dominates the optical spectrum over long time scales. Our results
allow optimization of the preparation process for spectral structures in a large variety of rare earth
ion doped materials for quantum memories and other applications.
I. INTRODUCTION
Rare-earth ion doped crystals (REICs) are interesting
materials due to their long-lived excited states and their
exceptionally long optical coherence times at cryogenic
temperature [1, 2]. In particular, along with the pos-
sibility for spectral tailoring of their inhomogeneously
broadened 4fN-4fNtransitions, this makes them prime
candidates for a number of applications in classical and
quantum optics. Examples include laser stabilization,
RF spectrum analysis, narrow band spectral filtering,
and quantum information storage and processing [2–7].
Thulium-doped Yttrium Gallium Garnet (Y3Ga5O12,
Tm:YGG) is one such material. Its 3H63H4transition
at 795 nm wavelength features an optical coherence time
of more than 1 ms [8–10], which is one of the longest
among all studied REICs. In combination with the ac-
cessibility of this transition—within the range of com-
mercial diode lasers—this makes it a natural candidate
for applications.
The quality of created features and the resulting con-
sequences for associated applications, are dependent on
the spectroscopic properties of the dopant ions and their
numerous interactions with other atomic components in
their local crystalline environment [11, 12], the details of
the optical pumping process, and the spectral and tempo-
ral profile of the applied laser pulses[13, 14]. Deep under-
standing of the relation between spectroscopic properties,
These authors contributed equally to this work. Present Address:
National Institute of Standards and Technology (NIST), Boulder,
Colorado 80305, USA
These authors contributed equally to this work.
optical control fields, and spectral diffusion dynamics has
resulted in improvements of this process in a number
of other rare-earth-doped materials including Tm:YAG,
Eu:YSO, and Pr:YSO [15–17]. However, this important
connection has thus far not been made for Tm:YGG.
In this paper we track the evolution of population
within the electronic levels of Tm3+ ions in YGG (see
Fig. 2 for simplified level scheme) by semi-continuous
monitoring of spectral holes for many sequences of ap-
plied spectral hole burning pulses. The characteristic
shapes and sizes of these spectral features are matched
to a rate equation model that encompasses the ground
(3H6), excited ( 3H4), and bottleneck (3F4) levels in this
material with associated lifetimes and branching ratios.
At zero magnetic field we see good agreement between
our numerical model and measured results across many
different pump sequences of varying duration, power, and
spectral shape. With the addition of an external mag-
netic field the agreement with our numerical model dis-
appears as spectral diffusion from local host spins begins
to dominate the shape of all spectral features over long
timescales. We characterize the nature of this unexpected
behavior and expand our model accordingly by adding a
spectral diffusion term to account for a quadratic Zeeman
interaction with present noisy magnetic fields [18, 19].
The letter is structured as follows: In section II we
describe the experimental setup used to collect our mea-
surements. In section III we detail the atomic level struc-
ture in Tm:YGG and introduce spectral hole burning, the
workhorse of our investigations, to select a known set of
atomic population. In section IV we introduce and apply
a rate equation model which shows good agreement to the
measured spectral hole features. In section IV A we detail
the use of adiabatic pulses to shape spectral holes at zero
arXiv:2210.05005v1 [quant-ph] 10 Oct 2022
2
magnetic field with the goal of creating high-resolution
features. Section V shows un-controlled changes to cre-
ated spectral holes in the presence of magnetic fields and
connects these noise effects to the quadratic Zeeman ef-
fect. Section VI extends this quadratic Zeeman connec-
tion to the characterization of spectral diffusion results
that differentiates the measured results from those pre-
dicted by our model over longer timescales.
II. EXPERIMENTAL SETUP
To measure spectral holes, from population storage
in various atomic levels, over different timescales in
Tm:YGG we use the setup detailed in Fig. 1. A
CW diode laser tuned to the ion transition frequency at
795.325nm [20] is locked to a reference cavity resulting
in a linewidth of roughly 5kHz [21]. To craft short pulses
of high extinction ratio, its continuous wave emission is
directed to a free space AOM. The sinusoidal driving sig-
nal of the AOM is mixed with a signal modulated by an
arbitrary function generator, which allows programmable
control of the transmitted pulse amplitude for the first
order light.
After the amplitude control, the pulsed light is di-
rected to a fiber coupled phase modulator driven using
arbitrary waveforms for serrodyne frequency shifting and
more complex chirped pulse shapes as detailed in section
IV A. The optical signals are then sent through a polar-
ization controller to a 1% Tm:YGG crystal grown by Sci-
entific Materials Corp. and housed in a pulse tube cooled
cryostat at 500-700mK. A superconducting solenoid cen-
tered on the crystal applies a homogeneous magnetic field
from 0-2T(using about 1mA/mT of current) along the
crystal’s <111>axis. Signals transmitted through the
crystal are directed to a fiber-coupled photo diode and
recorded for subsequent analysis
Experimental control is handled on a number of differ-
ent time scales via custom Python scripts that ensures
signals are created at the correct moment [22]. For se-
quencing on timescales of longer than a second, the built-
in Python timing functions are used to adjust the ex-
periment. On all timescale shorter than seconds, timing
is handled by pre-programming a pulse generator that
produces correctly timed trigger signals for the various
devices. Waveforms for the arbitrary voltage signals are
generated by custom scripts and uploaded to the respec-
tive devices for arbitrary control of instantaneous pulse
frequencies and amplitudes.
III. TM:YGG SITE AND LEVEL STRUCTURE
Garnet crystals such as YGG have cubic crystal struc-
ture with O10
hspace group symmetry, which yields six
Tm3+ ion substitution sites, each with D2point group
symmetry[9, 23–25]. The magnetic and optical behav-
ior of Tm3+ ions in each of these sites is identical but
AWG1
trigger
Pulse Generator
AWG2
trigger
AOM
RF Synth
CW
29 dB
40 dB
0th
1st
PM
ECDL 795nm
QWP
HWP
+PDH
+FALC
QWP
PD
PD
Scope
trigger
PD
Mkr
T= 600mK, B = 0 - 2 T
Tm:YGG
FIG. 1. Schematic of the experimental setup. A PC programs
a sequence on a pulse generator (SpinCore Pulseblaster) with
nanosecond timing resolution that produces a set of trig-
ger pulses for all devices. Waveforms written to an arbi-
trary waveform generator (AWG1,Tektronix AWG 70002A)
voltage channel, are subsequently amplified (SHF S126 A),
and drive an electro optic phase modulator (PM) to gen-
erate side bands on the laser light at arbitrary frequencies.
The AWG marker channels drive a set of home-built electri-
cal switches which gate the drive signal of an acousto-optic
modulator (AOM) to create short pulses from the laser light.
This gated AOM signal is mixed with fast arbitrary voltage
pulses(AWG2,Tektronix AFG 3102), amplified (Mini-Circuits
ZHL-5W-1+), and sent to an acousto-optic modulator (AOM
Brimrose 400MHz) to synthesize controllable-amplitude laser
pulses with rise times as short as 2.5ns. We use a single light
source (Toptica DL Pro 795nm) for different tasks (optical
pumping, pulse generation, etc). The laser frequency is set
via a wavemeter (Bristol 871) and locked to a thermally and
acoustically isolated high finesse optical cavity (Stable Laser
Systems) via the Pound-Drever-Hall method and a fast feed-
back loop acting on the laser current and piezo voltage (Top-
tica PDH and FALC Modules). Transmitted signals from the
crystal are directed to a variable-bandwidth photo detector
(NewFocus 2051) and displayed on an oscilloscope (Lecroy
Waverunner 8100A) configured by the PC and synchronized
with the experimental sequence by a trigger signal.
the crystal structure leads to effective in-equivalence be-
tween the sites due to six different orientations that the
ion and its entire local environment can take within the
lattice [26]. However, for a few specific directions relative
to the crystalline axes, ions at in-equivalent sites can be
cast into classes that share the same projections of ap-
plied electromagnetic fields ( ~
E,~
B) onto their local site
axes.
Given the orientation of our magnetic field, parallel to
the crystalline ~
B|| <111>axis, we cast the ions into
two different classes as depicted in Fig. 2 a. The ions
in each separate class experience different magnetic field
projections onto the axes in their local frame. One class
3
features magnetic field projections along the ion’s local
X and Z axes (blue), and the other projections along
the local Y and Z axes (red). This difference becomes
evident from a simple spectral hole burning experiment,
which we use to introduce the remaining results of the
paper.
Frequency Oset (MHz)
Absorption (cm-1)
c.
Energy
H
H
B = 0
Magnetic Field (B)
+1/2
-1/2
 g
+1/2
-1/2
 e
F
τ1 =1ms
τ1 =64ms
ζ
1-ζ
a.
b.
<001>
<010>
<100>
FIG. 2. a. Depiction of the six ion substitution sites relative
to the cubic crystal cell. For a magnetic field along the crys-
talline <111>axis the sites are cast into two classes featuring
different field projection, shown in red and blue. Small black
arrows indicate the local site X,Y,Z, axes for the red site in
the <100> <010>plane. b. Energy level structure of the
Tm ion transition of interest with previously measured life-
times (τ), branching ratios (ζ), and ground and excited split-
tings (∆g, ∆e). c. A pair of hole-burning spectra with the
main spectral hole pictured in the center. Each hole, shown
in an associated color to part a., is burned with a polariza-
tion that selects one of the two classes of ions and produces
anti-holes with different splittings (indicated by the colored
arrows). Additional modulations of the optical depth out-
side the hole originate from a slight orientation offset during
crystal cutting and polishing.
In spectral hole burning, a long optical pump pulse
excites atomic population in a narrow spectral window
within an inhomogeneously broadened absorption line.
Excited ions subsequently decay — either back into the
original state, or into another energy level that often be-
longs to the electronic ground state manifold. Scanning
a weak laser beam over a spectral interval centered on
the frequency of the original pump pulse reveals sections
of decreased and increased absorption — so-called spec-
tral holes and anti-holes. Spectral holes occur at fre-
quencies of reduced ground-state population, i.e. with
offset ∆ = 0 (for the central hole), ∆ = De(for the
side hole), and anti-holes can be observed whenever the
ground-state population is increased, which happens at
∆ = Dgand ∆ = Dg±De. Here, Dgand Deare ground
and excited state splittings. Consult Ref.[13, 27, 28] for
more details on spectral hole burning.
In our case, the ground and excited state splittings de-
pend on the magnitude and direction of the applied mag-
netic field — which vary for each class of Tm ions [26].
At 7.5 mT we recorded a pair of hole burning spectra,
shown in Figure 2 c, for orthogonal pumping polariza-
tions. This allowed us to selectively address ions in either
of the two classes. We found two sets of anti-holes, each
of which split according to the different field projections
experienced by the two classes of Tm3+ ions. This ability
to select out a single class of ions becomes important in
section VI as spectral diffusion depends on the magnetic
projection on each specific class of ions.
IV. MODELING RESULTS USING
THREE-LEVEL RATE EQUATIONS
To further understand the effects of a given hole burn-
ing process on our REIC ensemble we turn to solutions
of the Maxwell-Bloch equations that describe the inter-
action of light with one or many atomic systems. These
differential equations can be quite difficult to solve, given
the complexity that there does not exist a single fixed
Rabi frequency to drive all ions [29]. In the limit of exci-
tation pulse lengths much shorter or much longer than T2
the rate equation approach has been shown to be an ef-
fective model [30]. Thus for the case of Tm:YGG, due to
the long T2and low optical depth we reduce the Maxwell-
Bloch equations to a set of rate equations that describe
the conserved total atomic population and how it flows
through the different available levels as a function of time
[14].
Note that in the case of narrow band excitation of an
inhomogeneously broadened transition where only a cer-
tain portion of atoms are driven, frequency dependence
must be added. Following [14] and [15, 31] we describe
the dynamics of our atomic ensemble with equations 1-3.
ng(t)
t =R(∆)(neng) + 1ζ
Te
ne+1
Tb
nb(1)
ne(t)
t =R(∆)(ngne)1
Te
ne(2)
nb(t)
t =ζ
Te
ne1
Tb
nb(3)
This system of coupled differential equations describes
the relative change of atomic population, n(g,e,b)(t, ∆) in
three ion levels as a function of time, and frequency offset,
∆. The branching ratio ζdetermines how much popula-
tion decays from the excited state |eithrough the bottle-
neck state |bibefore reaching the ground state |giwith
respective level lifetimes Teand Tb. To drive the system,
摘要:

QuadraticZeemanSpectralDi usionofThuliumIonPopulationinaYttriumGalliumGarnetCrystalJacobH.Davidson,1,AntarikshaDas,1,yNirAlfasi,1RufusL.Cone,2CharlesW.Thiel,2andWolfgangTittel1,3,41QuTechandKavliInstituteofNanoscience,DelftUniversityofTechnology,2628CJDelft,TheNetherlands2DepartmentofPhysics,Montan...

收起<<
Quadratic Zeeman Spectral Diusion of Thulium Ion Population in a Yttrium Gallium Garnet Crystal Jacob H. Davidson1Antariksha Das1yNir Alfasi1Rufus L. Cone2Charles W. Thiel2and Wolfgang Tittel1 3 4.pdf

共12页,预览3页

还剩页未读, 继续阅读

声明:本站为文档C2C交易模式,即用户上传的文档直接被用户下载,本站只是中间服务平台,本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间,仅对用户上传内容的表现方式做保护处理,对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私,请立即通知玖贝云文库,我们立即给予删除!

相关推荐

分类:图书资源 价格:10玖币 属性:12 页 大小:5.21MB 格式:PDF 时间:2025-05-02

开通VIP享超值会员特权

  • 多端同步记录
  • 高速下载文档
  • 免费文档工具
  • 分享文档赚钱
  • 每日登录抽奖
  • 优质衍生服务
/ 12
客服
关注