University of Siegen

MANUAL
Master Lab Course

Optically detected magnetic resonance

(ODMR)

Authors
Philipp Reuschel
Assegid M. Flatae
Florian Sledz
Mario Agio

Department of Physics
Laboratory of Nano-Optics
Contents
1 Abstract 2

2 Theoretical background 3
2.1 Zeeman effect . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
2.2 Rabi model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
2.2.1 Co-rotating frame and time evolution . . . . . . . . . . . . . . . . . 4
2.2.2 Rabi oscillations . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

3 Physical system 8
3.1 Diamond and color centers . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Nitrogen-vacancy center . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9
3.2.1 Atomic structure and creation . . . . . . . . . . . . . . . . . . . . . 9
3.2.2 Fluorescence spectrum . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.3 Optically detected magnetic resonance (ODMR) . . . . . . . . . . . . . . . 13
3.3.1 Energy diagram . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
3.3.2 Manipulation of the spin-states and its application . . . . . . . . . 13

4 Experimental setup and procedure 17

4.1 Setup . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17
4.2 Experimental tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20
4.3 Analysis tasks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.4 Output requirements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

I
List of Figures
1 Excitation probability of a driven two-level system. . . . . . . . . . . . . . 7
2 Hybridization of carbon and the diamond lattice. . . . . . . . . . . . . . . 8
3 Atomic structure of the NV center. . . . . . . . . . . . . . . . . . . . . . . 10
4 Fluorescence spectrum of the NV center. . . . . . . . . . . . . . . . . . . . 12
5 Simplified energy diagram of the NV center. . . . . . . . . . . . . . . . . . 14
6 ODMR spectrum of an NV center ensemble. . . . . . . . . . . . . . . . . . 16
7 Microwave circuit and planar antenna. . . . . . . . . . . . . . . . . . . . . 18
8 Experimental setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19
9 Screenshot of Labview prorgam. . . . . . . . . . . . . . . . . . . . . . . . . 21

List of Tables
1 Calibration data for microwave source. . . . . . . . . . . . . . . . . . . . . 27
1 Calibration data for microwave source (continued). . . . . . . . . . . . . . 28

II
Acknowledgements
We would like to thank Navid Soltani and Jan Krause for technical support during the
mechanical construction of the setup. We are very grateful for the support provided
by the electronical workshop, especially Jens Winter, in designing the microwave circuit
and the fabrication of the planar antenna. We would like to thank Professor Christof
Wunderlich and Michael Johanning for providing the equipment used to calibrate the
microwave circuit. Lastly, we would like to thank Arka Gosh for assistance during the
test of the optical setup and his simulation of the microwave antenna.

1
1 Abstract
In this document, we present a technique called optically detected magnetic resonance
(ODMR) by which the electron spin state of a crystal defect in diamond is optically
pumped for spin initialisation and readout. We will utilize the nitrogen-vacancy (NV)
color center in diamond, to measure the modulus and orientation of an external magnetic
field using a confocal microscopy setup. ODMR of NV has applications in magnetometry,
sensing, biomedical imaging, quantum information and for the fundamental study of
physics.
In the theoretical background, we will introduce the interaction between a two-level
system and an external magnetic field. Firstly, we will revise the Zeeman effect, which
lays the foundation for magnetic field sensing with the splitting of spectral lines. Secondly,
we will discuss one of the key concepts of quantum optics, namely the Rabi model, which
describes the effect of an oscillating external magnetic field on a two-level system.
The nitrogen-vacancy center is the physical system used in this experiment. Its creation,
optical properties and the working principle of ODMR for measuring external magnetic
fields is explained.
The experimental setup, a confocal system combined with a microwave source, is shown
and students will have a hands-on experience during this experiment. The basic idea for
this student’s experiment, particularly the design of the microwave circuit, is taken from
[14].
Lastly, students will have to perform data analysis with a program of their choice (e.g.,
Matlab, Python). This includes plotting and function fitting. A guide through these
steps is provided at the end of the manual.

2
2 Theoretical background
2.1 Zeeman effect
The Zeeman effect (named after Pieter Zeeman) describes the interaction between a spin
J (such as an electron) and an external static magnetic field B. It can be observed by
the splitting of spectral lines into multiple components when an external magnetic field
is applied to the physical system. Since the energy split is a function of the field strength,
the Zeeman effect can be used to measure magnetic fields.

Classically, one can calculate the magnetic moment pm of an electron by

e
pm = − ·l, (1)
2me
where me is the electron’s mass and l its angular momentum. When such a magnetic
moment is brought into an external magnetic field B, its potential energy Epot is given
by

Epot = −pm · B . (2)

When the external magnetic field B is aligned along the z-axis, one finds that

e~
Epot = mz Bz , (3)
2me
since lz = m~. The angular momentum l classically precesses around the z-axis since
the torque D = −(e/2me ) l × B is applied [4].

A similar result can be obtained using quantum mechanics. The Hamiltonian Ĥ is defined
by

Ĥ = γ Jˆ · B , (4)
where γ is the gyro-magnetic moment and γ/2π ≈ 28.025 MHz mT−1 . Jˆ = Ĵx ex + Ĵy ey +
Ĵz ez is the total electronic angular momentum operator. For simplicity, we again assume
that the external magnetic field is aligned along the z-axis. This leads to

Ĥ = γBz Ĵz . (5)

To find the energy shift ∆E caused by the coupling between static external magnetic
field and a spin, we need to find eigenvalues for the equation

Ĥ |j, mj i = E |j, mj i , (6)

3
where j is the angular momentum number of the particle and mj its magnetic spin
number. We can solve this problem for the case of spin-1/2- and spin-1-particles using
matrix representation, defined by
 
  1 0 0
~ 1 0
Ĵz = and Ĵz = ~ 0 0 0  , (7)
2 0 −1
0 0 −1
respectively. For spin-1/2 we find mj = ±1/2 and for spin-1 we find mj ∈ {−1, 0, 1} with

Ĵz |j, mj i = ~mj |j, mj i . (8)

Therefore, the energy splitting ∆E of the Zeeman effect can be calculated simply by

∆E = ∆mj ~γBz . (9)

2.2 Rabi model

The Rabi model (named after Isidor Isaac Rabi) describes the interaction between a two-
level quantum system and a classical oscillating magnetic field. It is very famous since
it delivers an intuitive way of dealing with a two-level system. In this chapter, we will
briefly go over the basic idea of the model and coarsely do some derivation which leads
us to the results. A more detailed (and alternative) calculation can be found in [7] or on
quantum optics text books.

2.2.1 Co-rotating frame and time evolution

To start, we take a static external magnetic Bk aligned along the z-axis so we can
(according to the Zeeman model) define the ground state as |0i := |−zi and excited state
as |1i := |+zi. Without an oscillating field, we can write

~ω0 ~ω0
Ĥk = γBk Ĵz = (|ei he| − |gi hg|) = σz , (10)
2 2
where ω0 = γBk . Now we can add the oscillating field B(t) = B⊥ cos(ωt), which lies
within the xy-plane with an azimutal angle φ.

~ω0 ~
ĤS = σz − γB⊥ cos(ωt)[cos(φ)σx + sin(φ)σy ] (11)
2 2
~ω0 ~Ω
= σz − cos(ωt)[cos(φ)σx + sin(φ)σy ] (12)
2 2
~ω0 ~Ω
= σz − [exp(−i(ωt + φ))σ+ + exp(+i(ωt + φ))σ− ] , (13)
2 2

4
where Ω = γB⊥ is the Rabi frequency.

We now use the interaction picture to transform this Hamiltonian into a co-rotating
frame which oscillates at the same frequency ω as the varying magnetic field B⊥ . To do
this, we define

~ω0
Ĥ0 = Ĥk = σz (14)
2
~Ω
Ĥint = − [exp(−i(ωt + φ))σ+ + exp(+i(ωt + φ))σ− ] (15)
2
and use the unitary
ω
ÛRF = exp(−i t σz ) (16)
2
to transform the states |Ψi from the Schrödinger (S) picture into the rotating frame
(RF), such that

|ΨRF (t)i = Û†RF |ΨS (t)i . (17)

After some calculation, we arrive at the Hamiltonian in the rotating frame

~∆ ~Ω
ĤRF = − σz + [exp(−iφ)σ+ + exp(iφ)σ− ] (18)
2 2
~∆ ~Ω
=− σz + [cos(φ)σx + sin(φ)σy ] , (19)
2 2
with the detuning ∆ = ω − ω0 and the Rabi frequency Ω.

One finds the time evolution of an initial state |ΨRF (0)i

 
i
|ΨRF (t)i = ÛRabi (t) |ΨRF (0)i = exp − ĤRF t |ΨRF (0)i (20)
~
is simply a rotation on the Bloch sphere with

 
Ωtot
ÛRabi (t) = exp −i · σt (21)
2
   
Ωtot Ω Ωtot
= cos t 1−i sin t [cos(φ)σx + sin(φ)σy ] , (22)
2 Ωtot 2
√
where Ωtot = −∆ez + Ωe⊥ and Ωtot = ∆2 + Ω2 is the generalized Rabi frequency.

5
2.2.2 Rabi oscillations
We will now consider a two-level system which is initially in the ground state |−zi and
is driven by an external oscillating magnetic field Bx , which is aligned along the x-axis.
We are interested in the probability Pe to find the system in the excited state |+zi after
time T has passed:

Pe (T ) = | h+z| ÛRabi |−zi |2 (23)

      2
Ωtot Ω Ωtot
= h+z| cos T 1−i sin T (σ+ + σ− ) |−zi (24)
2 Ωtot 2
√ !
1 ∆2 + Ω 2
= sin 2
T . (25)
1 + (∆/Ω)2 2

This is a periodic sin-function with a Lorentian shaped amplitude which depends on the
detuning ∆ of the oscillating magnetic field relative to its field strength Ω.

We can learn two things from this: Firstly, the maximum probability to excite the two-
level system depends only on the detuning ∆. It reaches its maximum for a resonant
driving field (∆ = ω − ω0 = 0) and is basically zero for large detunings (∆  Ω).
Secondly, the system will oscillate between its ground state |−zi and |+zi in time. These
oscillations are called Rabi oscillations. For zero detuning (∆ = 0) the system will be in
the excited state for Ωtot T = ΩT = π (π-pulse) and will be in a superposition between
ground and excited state when ΩT = π/2 (π/2-pulse). This behaviour is shown in fig. 1.

6
Figure 1: Excitation probability of a driven two-level system initially in the ground state.
a) Maximum excitation probability Pemax = [1 + (∆/Ω)2 ]−1 as a function of the
relative detuning ∆/Ω. b) Excitation probability Pe as a function of pulse du-
ration time T for different detunings ∆. The system performs Rabi oscillations.

7
3 Physical system
3.1 Diamond and color centers
Diamonds were first mentioned by Platon in Greece at around 4th century BC and were
long used as gemstones for their shininess and rarity. Nowadays, its main application is
industrial grinding and cutting where synthetically grown diamonds are used due their
exceptional hardness.

Diamond can be created by High Temperature High Pressure (HPHT), which mimics its
natural growth conditions but adds molten metal as a solvent. To grow diamond of high
purity, Chemical Vapour Deposition (CVD) is mostly used, where carbon is deposited
from the gaseous phase (e.g. methane) onto a substrate. Diamond nano-crystals can
be created by the explosion of a carbon source under oxygen deficiency (detonation dia-
mond ).

Its only constituents, carbon atoms, form a regular lattice, the diamond lattice. To
achieve those very strong covalent bonds, the energy levels of carbon hybridize. While
an isolated carbon atom has two electrons on the the 2s and 2p levels, they are hybridized
to the sp3-level to form a tetrahedron with a binding angle of about 109.5° (see fig. 2).
Pure diamond has a bandgap of 5.5 eV and is optically transparent. Some impurities form
atomic-sized defects in the lattice which can emit visible light. These are called color
centers and diamond can host a multitude of those. One famous example is the Hope
diamond, which has boron impurities giving it its dark blue color. Recent studies show
color centers are used for a variety of applications, such as in quantum nanophotonics,
bio labelling, light emitting diodes, and spintronics. Some of these defects, namely the
nitrogen vacancy (NV), the silicon vacancy (SiV) have been subject to rigorous research
owing to their relatively known crystallographic and electronic structures.

Figure 2: Hybridization of carbon and the diamond lattice. a) Energy levels of an isolated
carbon atom. b) sp3-hybdridization of carbon which forms the diamond lattice.
c) Atomic structure of the diamond lattice.

8
3.2 Nitrogen-vacancy center
3.2.1 Atomic structure and creation
The nitrogen-vacancy (NV) center is a color center in diamond, where one carbon atom
is removed (vacant lattice site) and another neighbouring carbon atom is substituted
with a nitrogen atom. It exists in two charge states, the negatively charged NV- and
the neutral NV0 . Since most applications rely on the additional electron, we will use NV
center synonymous for the negatively charged state.

As nitrogen is the most abundant impurity in diamond, the NV center was first discov-
ered in the 1970s and has since been extensively studied. It shows spin-optical properties
at room temperature which, together with diamond’s exceptional stability, enables its
use in a variety of applications without the need for additional cooling or vacuum.

The NV center’s symmetry axis is placed along the one of the four diamond lattice ori-
entations: [111], [1̄1̄1], [1̄11̄], [11̄1̄]. Its (electrical) dipole, responsible for fluorescence
emission, is orthogonal to this axis. Therefore, the main emission of the NV center’s
farfield is also along its specific lattice orientation [10]. The NV center’s lattice structure
is shown in fig. 3.

To create a NV center, one has two general possibilities. Due to naturally occurring
contamination of diamond with nitrogen, one can use this already present nitrogen and
additionally create crystal damage (vacancies) by particle irradiation (e.g. ions, elec-
trons). Another approach is to use an ultra-pure diamond sample and implant nitrogen
into it, which creates vacancies at the same time. The diamond will be heated (annealed )
to about 800 °C to help the nitrogen and vacancy diffuse together to form an NV center.
Recently, it has been demonstrated that it is also possible to use highly energetic laser
pulses on a nitrogen-rich sample to create NV centers. Firstly, the laser pulses create
vacancies and, as a second step, it anneal the diamond locally until color centers are
formed [2].

9
Figure 3: Atomic structure of the NV center. The arrow indicates its symmetry axis along
one of the four diamond lattice orientations (111), (1̄1̄1), (1̄11̄) and (11̄1̄). Since
the (electrical) dipole is orthogonal to the symmetry axis, the main farfield
emission will be in the direction of the symmetry axis. Figure taken from [10].

10
3.2.2 Fluorescence spectrum
The NV center can be excited with an energy larger than the energy difference between
its ground and excited state, 1.945 eV. This energy can be delivered by a laser with a
wavelength λ ≤ 637 nm according to Planck’s law
c
E = hν = h , (26)
λ
where h is Planck’s constant, ν the frequency and c the speed of light.

Using non-resonant excitation, the system is excited into a higher vibronic band of the
excited state |ei and decays to the vibronic ground state of |ei within picoseconds. Emis-
sion into the zero-phonon line (ZPL) happens, when the system decays from the vibronic
ground state of |ei into the vibronic ground state of its ground state |gi (lifetime of about
10 ns). Not only absorption but also emission can leave the system in a higher vibronic
state. Consequently, the system can gain or loose energy by absorption or excitation of
a vibronic state of the surrounding diamond lattice. Quantized excitation of a lattice are
called phonons. The excitation of phonons is the dominating process for the NV cen-
ter, which explains a relatively small emission into the ZPL at 637 nm but a prominent
phonon sideband (PSB) shifted to longer wavelengths with a width of about 100 nm (see
fig. 4).
The fluorescence spectrum can be explained with an energy diagram. This diagram will
be discussed in the next section and will lead us directly to optically detected magnetic
resonance (ODMR).

11
Figure 4: Fluorescence spectrum of the NV center at room temperature under excitation
with a green (532 nm) laser. The spectrum shows small emission into the ZPL
at 637 nm and the characteristic PSB at higher wavelengths with a width of
approx. 100 nm.

12
3.3 Optically detected magnetic resonance (ODMR)
3.3.1 Energy diagram
The energy diagram of the NV center can be seen in fig. 5. This diagram visualizes the
structure of the main energy levels and their spin-subsystems. In general, the NV center
has three main energy levels, the ground state |gi, the excited state |ei and a singlet
state |si. Additionally, the ground and excited states are split into a spin-1 subsystem
with magnetic spin numbers ms = 0, ±1.

The ms = ±1 states are nearly degenerate with no external magnetic field and are sepa-
rated from the ms = 0 state by the zero-field splitting Dg = 2.87 GHz. When an external
magnetic field is applied, the ms = ±1 states will split by an energy difference of 2~γBk
according to the Zeeman effect.

Upon optical pumping, e.g. with a green (532 nm) laser, something remarkable happens:
The |g,0i state will be excited to vibronics levels of |e,0i and mostly decay back to the
|g,0i ground state under emission of a photon. The |g, ±1i state on the other hand will
be excited to vibronic levels of |e, ±1i and predominantly decay non-radiatively through
the singlet state |si to the |g,0i ground state. There are two important consequences:
Firstly, the |±1i-sublevel shows less fluorescence and, secondly, continuous optical pump-
ing will polarize the system to the |0i-sublevel. This is called spin polarization.

We conclude that we can initialize the spin-sublevel to the |0i state by extended laser
excitation and read out the current spin-state by probing the fluorescence signal with
a short laser pulse. High fluorescence indicates the |0i sub-state and low fluorescence
indicates the |±1i sub-state.

3.3.2 Manipulation of the spin-states and its application

The magic of the NV center relies on the combination of the optical readout of the spin
state (|0i or |±1i) in combination with the a zero-field splitting Dg in the gigahertz-range.

This means, that the transition between this two-level system can be driven with an
oscillating magnetic field with frequency ω. Since Dg is in the gigahertz-range, we will
use a microwave (MW) source for this. The system will perform Rabi oscillations with
Rabi frequency Ω ∝ B⊥ , which depend on the amplitude of the applied MW field. We
call a plot of the fluorescence intensity I over the applied MW frequency ω the ODMR
(optically detected magnetic resonance) spectrum. There are different methods to aquire
this spectrum. We will focus on cw -ODMR, in which the laser and MW are continuously
applied and the MW frequency is swept. There are pulsed measurement schemes, which
can be understood by applying the Rabi model, but those require more complex equip-

13
Figure 5: Simplified energy diagram of the NV center. The frequency axis (ν) is not
to scale. Radiative transitions are marked as solid while non-radiative decays
are marked as dotted arrows. There exists a ground state |gi, an excited
state |ei and a singlet state |si. Their additional vibronic states are shaded in
grey. Ground and excited state feature a spin-1 sublevel with magnetic spin
numbers ms = 0, ±1. Those are split by the zero-field splitting Dg = 2.87 GHz.
The transition between them can be driven with Rabi frequency Ω. Upon
application of an external magnetic field, the ms = ±1 sublevel split by γBk /π,
where γ/2π ≈ 28.025 MHz mT−1 . |e,0i will radiatively decay to |g,0i. |e, ±1i
will predominantly decay non-radiatively through |si to |g,0i.

14
ment and are not part of this lab course.

The ODMR spectrum will not only allow the calculation of Dg but, since the magnetic
sub-levels are split by γBk /π, also the strength of a static external magnetic field Bk . In
principle, one resonance for each crystallographic axis will be visible. Depending on the
diamond orientation relative to the magnetic field, e.g., parallel to the [111]-direction,
some of them might merge to an indistinguishable resonance dip with a higher contrast
(see fig. 6). In general, the contrast of the resonances depends on the orientation of
the electrical and magnetic dipole in regards to the experimental setup. The laser po-
larization defines the electric field at the focus. The collection efficiency depends on the
electrical dipole orientation and the NA of the objective. The transition frequency of the
spin-subsystem depends on the MW polarization.

It has been shown that the splitting between the ms sub-levels also depends on tem-
perature, external electric fields and lattice strain which further enables the NV center’s
use as a single atom sensor. These effects, however, are not easily distinguishable which
complicates the application as a multi-parameter sensor. Nonetheless, possible applica-
tions of the NV center and ODMR in particular range from quantum cryptography [1]
and quantum memory [6] to magnetic [3] and electric [5] field and temperature sensing
[13] to the violation of Bell-inequalities [8].

15
Figure 6: ODMR spectrum of an NV center ensemble with an external magnetic field
aligned at 45° to the [111] crystal axis. Eight-fold splitting of the resonances is
visible.

16
4 Experimental setup and procedure
4.1 Setup
As discussed in the theory, measuring ODMR requires three main ingredients: Firstly,
we need a strong light field (e.g. a laser) to excite the NV centers. Secondly, we need
to collect the fluorescence signal and, thirdly, we need to apply a magnetic field with
microwave frequency to drive the spin-sublevels.

In this experiment, we excite the NV centers using a 5 mW, 532 nm diode laser (Thor-
labs CPS532 ) in a home-built confocal microscopy setup. The laser intensity is con-
trolled using a variable neutral density filter. The sample is excited through a 50x 0.95
NA objective (Zeiss C Epiplan-Apochromat 50x/0.95 DIC M27 ) which is mounted on a
transnational z-stage for efficient excitation and fluorescence signal collection. The sam-
ple, single-crystalline NV rich diamond sample (2.6 mm x 2.6 mm x 1 mm), is placed on
a planar antenna which provides a strong, homogeneous, broadband magnetic field [11]
(see fig. 7b). The antenna along with the sample is placed on x-y-stage to align their
position relative to the microscope objective.

To generate the microwave signal, which drives the Rabi oscillations, we use a voltage-
controlled oscillator (VCO, Mini-Circuits TB-ROS-3050C+). The VCO’s input voltage
is fed by a Labview-controlled function generator (Keysight 33210A), whose output sig-
nal is amplified by an operational amplifier circuit (amplification A = 1.5) to match the
VCO’s input range of 1.5 V to 15 V. The microwave signal is amplified by an amplifier
(Mini-Circuits ZRL-3500+) to a power of more than 20 dBm at a spectral bandwidth
of 0.1 MHz (see fig. 7a). The signal is transmitted to the antenna by SMA cables. The
calibration data of this circuit (tuning voltage vs. microwave frequency) can be found in
the appendix.

The static magnetic field is created with a solenoid and a current source. You can assume
the following parameters for the solenoid: Number of windings N = 302, average radius
R = 28.9 mm and length l = 18 mm. These values have errors smaller than 5 %. The
current will be varied from 0 A to 1.6 A. The solenoid axis is angled at about 30° with
respect to the z-axis.

From the objective the signal is sent through a 550 nm dichroic mirror (Semrock Di03-
R532 ) which transmits light that has a wavelength larger than 550 nm (fluorescence) and
reflects light below that wavelength (laser). To ensure that no residual laser light goes
to the detectors, an additional 550 nm longpass filter (Thorlabs FELH0550 ) is placed
along the detection path, which reflects all light with wavelengths below 550 nm. Before
the detection, a pinhole (50 µm diameter) can be placed at the common focus point of

17
 (a) (b)

Figure 7: (a) Microwave circuit. 1. Tuning voltage input. 2. Operational amplifier (am-
plification A = 1.5). 3. Voltage-controlled oscillator (VCO). 4. Microwave
amplifier. 5. DC power supplies. (b) Planar microwave antenna with SMA
connector. The red circle indicates the proper diamond sample position.

two lenses (confocal), which can be placed in the beam path and limits the collection
volume to the confocal volume. This helps to suppress unwanted background (such as a
sample holder). By assuming a focused Gaussian beam, one can show that the confocal
dimensions (Airy disc ∆r and depth of field ∆z) are

λ λ
∆r ≈ 0.61M and ∆z ≈ 2M 2 , (27)
NA NA2
where M is the magnification of the objective, NA is its numerical aperture and λ the
wavelength of the fluorescence signal [9].

There are different types of detectors that are used in the setup. By flipping a mirror in
the beam path, the light gets reflected to the corresponding detector. We use a CMOS
(complementary metal–oxide–semiconductor) camera (Thorlabs DCC1645C ) to align the
laser beam and to check the alignment of the confocal setup. In addition, it helps to
image the fluorescence signal using proper filters. A spectrometer (Oceanoptics QEPro)
is used to characterize the fluorescence signal and to make sure that NV centers are
excited. By integrating over the NV spectrum and varying the applied MW frequency,
we can use the spectrometer to measure ODMR.

18
 (a)

(b)

Figure 8: (a) Experimental setup sketch. (BS) Beamsplitter. (CMOS) Complementary

metal–oxide–semiconductor. (LP) Longpass filter. (VCO) Voltage-controlled
oscillator. (b) Image of the real setup. Red and green lines are added to
indicate the laser and fluorescence beam path.

19
4.2 Experimental tasks
1. Wear laser safety goggles.
2. Now, turn on the laser and wait for five to ten minutes until it stabilizes its output
power.
3. Open the Oceanview software, select Quickview and change the integration time
to 10 ms.
4. Focus the laser beam on the diamond surface by turning the micrometer screw on
the z-stage. Find an area on the sample, where you can see a clear NV ZPL on the
spectrum. You can move the diamond sample relative to the objective by turning
the micrometer screws on the x- and y-stage. Make sure to limit unwanted signal
from the substrate. Wait for a stable ZPL and a good signal-to-noise-ratio. When
you found a satisfying signal, take a spectrum and save it to a .txt-file.
5. Close the Oceanview software, open the Labview ODMR software. In the ODMR
software, set the exposure time to a suitable value (e.g., 20 ms), select a voltage
from 1 V to 10 V with a step size of 10 mV steps. The spectrum is automatically
integrated from 600 nm to 800 nm to create the ODMR signal. The data is saved
into the sub-folder data within the ODMR folder. The voltage is saved in the first
column and the intensity is saved in the second column of the file. The columns
are separated by semi-colons. Select refresh in the drop-down menu and afterwards
connect to the signal generator by selecting it in the drop-down menu. Press connect
devices. You should hear a click from the relais inside the signal generator. Now,
click start measurement. If you want to repeat the measurement, click on the arrow
in the top left corner, click connect devices and then, again, start measurement. You
can use this first measurement to characterize the laser fluctuations.
A screenshot of this Labview program is shown in fig. 9.
6. You can now connect the MW antenna.
7. Unplug the cables from the solenoid. Switch on the lab power supply to check that
it is set to its minimum current value. You can switch between current control and
voltage control by turning the corresponding knob. Now, connect the solenoid to
the power supply. Remember, that you may not exceed the solenoid’s maximum
current rating of 1.6 A. Place the solenoid as close to the antenna as possible.
Measure/estimate the distance between the sample and the solenoid’s center. You
can read the exact current from the multi-meter, which is connected in line with
the solenoid and the power supply.
8. Select a high current (max. I = 1.6 A) to generate a strong external magnetic
field such that eight distinct resonances are visible, when you measure the ODMR

20
Figure 9: Screenshot of Labview prorgam. The operational steps are the following: 0. Use
the dropdown menu. Click refresh. Select the function generator. 1. Specify
the integration time of the spectrometer. A time of 0.01 − 0.02 s should be
sufficient. 2. Define the tuning voltage interval (U = 1 − 10 V) and the step
size. For a fast sweep, you can use ∆U = 50 mV. It is recommended to not go
below ∆U = 5 mV because of a rising acquisition time. Usually. ∆U = 10 mV
is a good trade-off between acquisition time and resolution. 3. Connect the
devices. You should hear a click from the relais inside the signal generator.
4. Start the measurement. R. If you want to start a new measurement, click on
the arrow and repeat steps 1−4. S. You can observe the fluorescence spectrum,
and the ODMR spectrum (I vs. ν and I vs. U ) by changing the index tab.

21
 spectrum. If that is not the case, rotate the diamond sample and or vary the
current until they are visible. Take a picture of the position of the diamond sample
on the sample holder relative to the solenoid once it is well aligned.

9. Sweep the solenoid current from 0 A to 1.6 A in 200 mA steps. Measure the ODMR
spectrum each time (1 V to 10 V with step size of 10 mV).

10. Lastly, check that all your data is saved properly. You need a picture of the
alignment of the sample relative to the solenoid. Additionally, you need an ODMR
spectrum with a disconnected antenna and eleven ODMR spectra with a connected
antenna and a varied external magnetic field (I = 0 − 1.6 A).

22
4.3 Analysis tasks
1. Plot the fluorescence spectra (NV and its background) in one figure.

2. The calibration data for the microwave circuit (tuning voltage U ↔ microwave
frequency ν) can be found in the appendix. You can either use linear interpolation
or a first to third order polynomial to do a fit. Explain why you chose your method
(ease of use, uniform residuals...). Use this calibration for all further data analysis.

3. Plot the fluorescence fluctuations, which (mostly) correspond to the laser fluctu-
ations, measured without MW antenna over time. You get the time scale from
the integration time and the voltage steps. Characterize the fluctuations by their
signal to noise ratio SNR = amplitude/standard deviation.

4. Plot all measured ODMR spectra with axes intensity, ν and solenoid current I.
Sort them in terms of I. Suitable plot types for such three-dimensional datasets
are, e.g., pcolor in matplotlib or Matlab, waterfall in Matlab.

5. Determine the zero-field splitting D ± E (for I = 0 A) and its error by fitting two
Lorentzians plus linear background f (ν)

A1 A2
f (ν) = 2
+ + Cν + F (28)
1 + [(ν − D − E)/Γ1 ] 1 + [(ν − D + E)/Γ2 ]2

to the zero-field ODMR spectrum. Ai are the amplitudes and Γi their full width at
half maximum (FWHM). C and F are the fit parameters of the linear background.
It will be helpful to provide starting values and upper and lower bounds to the fit
parameters. Specify the error of D and E. Plot the data set and your fit function
together in one figure.

6. Fit the ODMR spectrum with the largest external magnetic field with Lorentzian
functions. Eight dips should be visible. Instead of fitting the sum of eight Lorentzians,
it might be sufficient to fit each dip separately with a single Lorentzian and a linear
background of the form gi (ν)

Gi
gi (ν) = + Hi ν + K i , (29)
1 + [(ν − ν0,i )/Γi ]2

where Gi is the amplitude of the dip, ν0,i the position of its center, Γi its FWHM
and i ∈ [1; 8]. Hi and Ki are the fit parameters of the linear background. Specify
ν0 and Γ and their errors. It is highly recommended to provide starting values and
upper and lower bounds to the fit parameters. Plot the data set and the fitted
functions in their range together in one figure.

23
 7. You can assume a symmetric splitting of all dips around D due to their magnetic
spin number ms = ±1. Their energy (frequency) difference is

h ∆ν0 = ∆E = 2 ~γBi , (30)

where Bi is the magnetic field parallel to their crystal axis and γ/2π ≈ 28.025 MHz mT−1 .
Therefore, you can determine the parallel external magnetic field Bi and its error
for each of the four crystal axes.

8. We will now determine the magnitude1 of the external magnetic field.

• Assume the crystal axes are

e1 = ez , (31)
e2 = sin(α) cos(β)ex + sin(α) sin(β)ey + cos(α)ez , (32)
e3 = sin(α) cos(β)ex − sin(α) sin(β)ey + cos(α)ez , (33)
e4 = sin(α)ex + cos(α)ez , (34)

where α = 109.5° and β = 120°.

• Calculate the unit vectors in x-, y-, and z-direction in terms of crystal axis
vectors. Calculate Bx , By and Bz in terms of the Bi accordingly. For example,
the unit vector of the x-axis can be calculated by

ex = [e4 − cos(α)e1 ]/ sin(α) (35)

and, consequently, the magnetic field in x-direction Bx is

Bx = [B4 − cos(α)B1 ]/ sin(α) . (36)

• You can determine the absolute value |B| by

q
|B| = Bx2 + By2 + Bz2 . (37)

1
One can actually also determine the orientation of the magnetic field using NV centers. To do so, one
has to know the crystal orientation corresponding to the ODMR resonances. This can be done with
an static external magnetic field aligned along one of the crystal axes [12] or with polarimetry.

24
 • You now know the amplitude of the external magnetic field B. Compare this
to the analytical solution of the on-axis magnetic field of a solenoid Bs

!
µ0 IN l/2 − z l/2 + z
Bs (ρ = 0, z) = p + p , (38)
2 l R2 + (l/2 − z)2 l R2 + (l/2 + z)2

where N is the number of turns, I is the solenoid current, l is the length

of the solenoid and R its radius. ρ is the radial and z is the axial distance
from the center (!) of the solenoid. The solenoid axis is angled at about 30°
with respect to the z-axis. Estimate the errors for ρ and z from the setup.
Accompany this with the picture you took from the sample and the solenoid.
Compare this analytical solution to the B-field determined by ODMR.

9. Compare your results of the external magnetic field |B| with the ODMR spectra
plotted in the diagram. Give an estimate of the minimum external magnetic field
needed to resolve all eight resonances. The parameter Γ from the Lorentzian fits
might be useful for this.

10. To summarize, we would like you to: Plot the three fluorescence spectra. Calibrate
the microwave circuit and plot the ODMR spectra for multiple solenoid currents.
Determine the zero-field splitting. At a large external magnetic field, determine
the B-field amplitude and compare it to the analytical model of the solenoid used
to generate the field.

4.4 Output requirements

A good lab report does more than presenting data; it demonstrates the writer’s compre-
hension of the concepts behind the data. Merely recording the expected and observed
results is not sufficient; you should also identify how and why differences occurred, ex-
plain how they affected your experiment, and show your understanding of the principles
the experiment was designed to examine. Bear in mind that a format, however helpful,
cannot replace clear thinking and organized writing. You still need to organize your ideas
carefully and express them coherently.

Typical components to be fulfilled are: Title Page, Abstract, Introduction, Methods and
Materials (or Equipment), Experimental Procedure, Results, Discussion, Conclusion,
References. Figures should have figure captions with explanations. All graphs should
have labeled axes and visible data points and curves. Curve fits and fitted parameters
should be included. All quantities should be followed by the calculated error.

25
If at any point during your data processing or report writing you have any doubts, you
are encouraged to contact your teaching assistants. We are more than happy to help
you finish your task and learn as much as possible. We wish you good luck with your
experiments!

26
Appendix

Table 1: Calibration data for microwave source (tuning voltage U vs. microwave fre-
quency ν and power P .)
U (V) ν (GHz) P (dBm)
1.0 2.602 20.0
1.2 2.606
1.4 2.610
1.6 2.626
1.8 2.640
2.0 2.652 22.0
2.2 2.664
2.4 2.674
2.6 2.684
2.8 2.696
3.0 2.706 22.1
3.2 2.718
3.4 2.730
3.6 2.740
3.8 2.752
4.0 2.764 22.2
4.2 2.776
4.4 2.788
4.6 2.796
4.8 2.814
5.0 2.828 22.2
5.2 2.842
5.4 2.854
5.6 2.870
5.8 2.882
6.0 2.896 22.2
6.2 2.910
6.4 2.926
6.6 2.932
6.8 2.954
7.0 2.968 22.2
7.2 2.982
7.4 2.996
7.6 3.010
continued on next page

27
Table 1: Calibration data for microwave source (continued).
U (V) ν (GHz) P (dBm)
7.8 3.024
8.0 3.036 22.4
8.2 3.056
8.4 3.062
8.6 3.074
8.8 3.084
9.0 3.094 22.5
9.2 3.104
9.4 3.114
9.6 3.122
9.8 3.130
10.0 3.136 21.8

28
References
[1] Alexios Beveratos et al. „Single Photon Quantum Cryptography.“ In: PRL 89.18
(Oct. 2002), p. 187901. url: https://link.aps.org/doi/10.1103/PhysRevLett.
89.187901.
[2] Yu Chen Chen et al. „Laser writing of individual nitrogen-vacancy defects in dia-
mond with near-unity yield.“ In: Optica 6.5 (2019), p. 662. doi: 10.1364/optica.
6.000662.
[3] C. L. Degen. „Scanning magnetic field microscope with a diamond single-spin sen-
sor.“ In: Appl. Phys. Lett. 92.24 (June 2008), p. 243111. issn: 0003-6951. doi:
10.1063/1.2943282. url: https://doi.org/10.1063/1.2943282.
[4] Wolfgang Demtröder. Experimentalphysik 3. Jan. 2016. isbn: 978-3-662-49093-8.
doi: 10.1007/978-3-662-49094-5.
[5] F. Dolde et al. „Electric-field sensing using single diamond spins.“ In: Nature
Physics 7.6 (2011), pp. 459–463. issn: 1745-2481. url: https://doi.org/10.
1038/nphys1969.
[6] Torsten Gaebel et al. „Room-temperature coherent coupling of single spins in dia-
mond.“ In: Nature Physics 2.6 (2006), pp. 408–413. issn: 1745-2481. url: https:
//doi.org/10.1038/nphys318.
[7] Christopher Gerry and Peter Knight. Introductory Quantum Optics. Cambridge
University Press, 2004. doi: 10.1017/CBO9780511791239.
[8] B. Hensen et al. „Loophole-free Bell inequality violation using electron spins sepa-
rated by 1.3 kilometres.“ In: Nature 526.7575 (2015), pp. 682–686. issn: 14764687.
doi: 10.1038/nature15759.
[9] Lukas Novotny and Bert Hecht. Principles of Nano-Optics. Cambridge University
Press, 2006. doi: 10.1017/CBO9780511813535.
[10] Linh Pham et al. „Enhanced metrology using preferential orientation of nitrogen-
vacancy centers in diamond.“ In: Physical Review B 86 (July 2012). doi: 10.1103/
PhysRevB.86.121202.
[11] Kento Sasaki et al. „Broadband, large-area microwave antenna for optically de-
tected magnetic resonance of nitrogen-vacancy centers in diamond.“ In: Review of
Scientific Instruments 87.5 (2016), p. 053904. doi: 10.1063/1.4952418. eprint:
https://doi.org/10.1063/1.4952418. url: https://doi.org/10.1063/1.
4952418.
[12] S Steinert et al. „High Sensitivity Magnetic Imaging Using an Array of Spins in
Diamond.“ en. In: Rev. Sci. Instrum. (), p. 6.

29
[13] David M. Toyli et al. „Fluorescence thermometry enhanced by the quantum coher-
ence of single spins in diamond.“ In: Proc Natl Acad Sci USA 110.21 (May 2013),
p. 8417. url: http://www.pnas.org/content/110/21/8417.abstract.
[14] Haimei Zhang et al. „Little bits of diamond: Optically detected magnetic resonance
of nitrogen-vacancy centers.“ In: American Journal of Physics 86.3 (2018), pp. 225–
236. doi: 10.1119/1.5023389. eprint: https://doi.org/10.1119/1.5023389.
url: https://doi.org/10.1119/1.5023389.

30