arXiv:quant-ph/0608233v2 26 Oct 2006
Room-temperature manipulation and decoherence of a single spin in diamond
R. Hanson, O. Gywat and D. D. Awschalom
Center for Spintronics and Quantum Computation,
University of California, Santa Barbara, California 93106, USA
(Dated: February 1, 2008)
We report on room-temperature coherent manipulation of the spin of a single nitrogen-vacancy
center in diamond and a study of its coherence as a function of magnetic field. We use magnetic
resonance to induce Rabi nutations, and apply a Hahn spin echo to remove the effect of low-frequency
dephasing. A sharp rise in the decoherence rate is observed at magnetic fields where the nitrogen-
vacancy center spin couples resonantly to substitutional nitrogen spins via the magnetic dipolar
coupling. Finally, we find evidence that away from these energy resonances spin flips of nitrogen
electrons are the main source of decoherence.
PACS numbers: 76.30.Mi,03.67.Lx,03.65.Yz
The study of single quantum systems is interesting
both for testing fundamental laws of physics as well as
for practical purposes, as computing with quantum sys-
tems promises an enormous increase in computing power
and quantum communication allows secure information
exchange1. In the solid state, coherent control of single
quantum systems has been achieved in a number of sys-
tems, e.g. superconducting Cooper pair boxes2and elec-
tron spins in quantum dots3. Among these, the nitrogen-
vacancy (N-V) center in diamond4is unique, because its
spin exhibits a long coherence time that persists up to
room-temperature5, whereas most other systems only al-
low coherent control at cryogenic temperatures.
Coherent manipulation of N-V centers on large ensem-
bles was first achieved many years ago6,7. Recently, how-
ever, coherent rotations and spin echoes of a single N-V
center spin were reported by Jelezko et al.8. This land-
mark experiment, that has not been reproduced by a
different group thus far, demonstrates that the N-V cen-
ter provides a testbed for quantum manipulation in the
solid state at room temperature9,10. On the other hand,
single-center spectroscopy allows the study of the local
environment of the N-V center11and has already un-
veiled anisotropic spin interactions and magnetic dipolar
coupling to spins of other defects in diamond12. Recent
results of these studies include the observation of strong
coupling between a single N-V center and the spin of
a single substitutional nitrogen atom13,14and the mea-
surement of the spin relaxation time of a single nitrogen
electron spin14. By combining single-center spectroscopy
with coherent control, the coherent interaction of the N-
V center spin with its environment can be probed, which
might ultimately lead to coherent quantum circuits9.
Here, we report coherent control of the electron spin
state of a single N-V center at room temperature. We
demonstrate that we can coherently drive single-spin ro-
tations (Rabi nutations) and undo low-frequency dephas-
ing by application of a Hahn spin echo sequence. We use
this capability to study the coherence of the N-V center
as a function of applied magnetic field. By comparing the
magnetic-field dependences of the decoherence rate and
the photoluminescence, we find that the coherence of the
N-V center spin is strongly affected by the resonant spin
exchange with the surrounding nitrogen spins at specific
magnetic fields. At other fields, we find evidence that the
dephasing caused by spin flips of the nitrogen electrons
is the main contribution to the observed decoherence.
The N-V center consists of a substitutional nitrogen
atom next to a vacancy in the diamond lattice. We study
the negatively charged state of the N-V center (N-V−),
which has a level structure as depicted in Fig. 1(a). It
has a spin triplet (3A) ground state, with a zero-field
splitting D= 2.88 GHz between the sublevels with spin
z-component mS=0 and mS=±115. The spin is quan-
tized along the N-V symmetry axis, a <111> crystal
axis16. There is a strong optical transition to an excited
(3E) triplet state that conserves spin. Linearly polarized
optical excitation preferentially pumps the spin system
into the ground state mS= 0 level17, enabling efficient
optical initialization of the spin state. Also, the average
photon emission rate is substantially smaller for transi-
tions involving the mS= ±1 levels than for the mS= 0
level18, which allows readout of the spin state by the
photoluminescence intensity IPL. These two effects have
been attributed to spin-dependent intersystem crossing
from the3E state to a nearby singlet (1A) state16,19,20.
We study five different single N-V centers in a single-
crystal high-temperature high-pressure (type Ib) dia-
3A ( =1)
3E ( =1)
1A ( =0)
FIG. 1: (a) Electronic level structure of the N-V center. The
excited3E state is 1.945 eV above the3A ground state. The
spin sublevels of the electronic ground state are manipulated
in this work. (b) Electron spin resonance (ESR) between the
mS=0 and mS=−1 spin levels of NV 14 at B = 100 G.
mond, commercially available from Sumitomo Electric
Industries. This diamond contains nitrogen (N) impuri-
ties with a density of 1019−1020cm−3. These impurities
(also known as P1 centers) contain one unpaired elec-
tron carrying a spin of1
2. The N-V centers are typically
spaced by a few µm, much larger than the spatial resolu-
tion of our setup (≈ 0.3 µm), allowing the study of single
N-V centers. We detect the phonon-broadened3E-3A
transition of the N-V center using non-resonant photolu-
minescence microscopy along the  crystal axis (for
details on the setup see Ref. 12). We identify single N-V
centers through photon antibunching measurements and
Electron Spin Resonance (ESR) measurements. All ex-
periments are performed at room temperature.
We precisely align the external magnetic field B with
the  crystal axis, which is the symmetry axis for all
centers studied in this work. We will refer to this axis
as the z-axis. The spin Hamiltonian of the N-V center
HNV can then be written as
HNV = DS2
where µB is the Bohr magneton,?S is the N-V center
electron spin operator,¯A is the hyperfine tensor7,16and
?I is the spin operator of the N-V center nitrogen nucleus.
Virtually all nitrogen atoms in our diamond are14N with
total nuclear spin I=1. The electron g-factor is 2.00.
We use magnetic resonance to manipulate the N-V cen-
ter electron spin state. An AC magnetic field, perpen-
dicular to the static magnetic field, is generated by send-
ing a radio-frequency (RF) alternating current through a
gold wire of 25 µm diameter that runs close to the N-V
center. This flexible approach allows us to study many
different centers in a single sample. For a wire-to-N-V
center distance of 20 µm, the AC magnetic field at the
site of the N-V center is about 1 G for 16 dBm output
power at the RF source. Figure 1(b) shows the typical
continuous wave (CW) ESR signal for a single center (in
this case NV14). Away from the resonance, the photolu-
minescence is high because the spin is strongly polarized
into the mS= 0 state by the continuous optical excita-
tion. When the applied frequency matches the energy
splitting between the mS=0 and mS=−1 levels, the AC
magnetic field induces rotations between the spin levels.
This reduces the spin polarization and therefore the mea-
sured photoluminescence intensity IPLdrops.
For coherent control of the spin state, we separate
the polarization, the spin manipulation and the readout
in time by applying the sequence depicted in Fig. 2(a).
First, a strong laser pulse initializes the spin in the mS=0
state. Then the laser is turned off, and the AC magnetic
field is pulsed to coherently manipulate the spin state.
Finally, the spin state is read out by a second laser pulse.
This way, we avoid the decoherence that can be induced
by the optical excitation8.
Upon application of the AC magnetic field, the spin is
coherently driven between the two spin sublevels. The
probability PmS=0 to find the N-V center in the mS=
0 0.5 1.0
FIG. 2: Coherent manipulation of the electron spin of a single
N-V center. (a) Three-step sequence applied: the first laser
pulse (typically 5 µs long) initializes the spin, then the spin
manipulation is carried out in the dark, and finally the spin
state is read out by a second laser pulse (typically 2 µs long)21.
The sequence is typically repeated 1000 times to increase the
signal. Laser power is 500 µW. (b) Driven coherent oscil-
lations (Rabi nutations) of NV14 at B=850 G. The spin is
rotated between the mS=0 and the mS=−1 level, leading to
alternating high and low IPL. The Rabi nutation frequency
is proportional to the square-root of the applied RF power,
as expected.Curves are offset for clarity.
echo pulse sequence. (d) Hahn spin echo signal of NV14 at
B=850 G as a function of the total delay τ1+ τ2.
(c) Hahn spin
0 state thus varies periodically, as described by Rabi’s
where f1is the Rabi nutation frequency, ∆f the detuning
from the resonance frequency and t the time from the
start of the AC magnetic field pulse. The Rabi frequency
depends on the amplitude B1of the AC magnetic field:
2gµBB1/h, where h is Planck’s constant and the
2is due to the rotating wave approximation.
In Fig. 2(b) we plot IPLas a function of pulse width
for three different RF output powers. We observe that
IPL, which is proportional to PmS=0, oscillates with a
frequency that depends on the applied power. The blue
curves are fits to the data using a cosine multiplied by
an exponential with a single decay constant T′
that this time is different from the dephasing time T∗
These measurements demonstrate our ability to coher-
ently manipulate the spin state of a single N-V center.
That each Rabi oscillation curve is well fit by a single-
frequency cosine is surprising at first sight, since the hy-
perfine coupling to the nitrogen nucleus of the N-V cen-
ter is expected to split the resonance condition into three
frequencies, spaced by about 2 MHz16(see Eq. 1). From
Eq. 2 we see that, for a fixed value of B1, the nutation fre-
quency depends on the spin state of the nitrogen nucleus.
As we averageoverdifferent nuclear spin configurations, a
beating pattern is expected in Fig. 2(b). The absence of
this beating could be explained by a hyperfine-induced
polarization of the nitrogen nuclear spin under strong
optical excitation. Such a polarization was observed re-
cently around 500 G13, but is possible at other fields as
well. Indeed, we observe a single Rabi frequency at all
magnetic fields probed (50-1020 G). Note also that “for-
bidden” transitions, in which the nuclear spin is simulta-
neously changed with the electron spin, are not observed
For long pulses, the oscillations damp out due to inter-
actions with the environment. We note that the decay
2increases with increasing Rabi frequency, which
can be ascribed to the refocussing effect of the Rabi nu-
tations: continuous driving can be viewed as a series of
concatenated π-pulses23. The data in Fig. 2(b) therefore
suggests that some of the observed damping is due to
dephasing and thus can be eliminated. To test this hy-
pothesis, we apply a Hahn spin echo to the N-V center.
The Hahn pulse sequence is depicted in Fig. 2(c). A π/2
pulse creates a coherent superposition of the mS=0 and
mS= −1 states, which is allowed to dephase during a
time τ1. After a π-pulse, the dynamics are reversed and
a spin echo occurs after a time τ2equal to τ1. We apply
a final π/2 pulse to map the echo signal to the readout
basis. By fixing τ1and measuring the spin echo signal as
a function of τ2, we have confirmed that the echo signal
is maximum for τ1= τ2(data not shown).
Figure 2(d) shows the echo signal as a function of 2τ
(here τ1= τ2= τ). The echo signal decays exponentially
on a timescale of T2=6 µs, about three times longer than
the decay time of the Rabi nutations. This demonstrates
that for this N-V center the fluctuations in the environ-
ment that dominate the decay are slow on the microsec-
ond timescale. The influence of higher-frequency dephas-
ing sources can in principle be eliminated by applying
multiple spin echoes or more elaborate pulse sequences.
We investigate the sources of decoherence in more de-
490 500510520 530
FIG. 3: Dependence of the coherence on the applied magnetic
field B. (a) Decoherence rate 1/T′
B. (b) Energy levels of the N-V center and of N electron
spin as a function of B along the N-V symmetry axis. An
energy resonance occurs around B=514 G. (c) IPL of NV14
as a function of B. The dips reflect the reduced spin po-
larization of the N-V center due to resonant spin exchange
with the N electron spin through magnetic dipolar interac-
showing stronger decoherence at the fields where IPL dips.
The dashed line depicts the average off-resonance value of
at B=850 G versus normalized amplitude of the dip in IPL
for five different N-V centers.
2of NV14 as a function of
2of NV14 as a function of magnetic field,
2. The red lines in (c) and (d) are Lorentzian fits. (e) T′
tail by measuring Rabi nutations over a wide range of
magnetic field. In Fig. 3(a) we plot 1/T′
B. We observe that 1/T′
2is more or less constant over the
whole range, except around B=514 G. As was found in
previous work12,14,24, this magnetic field marks an energy
2as a function of
resonance between the mS= 0 and mS= −1 levels of the
N-V center and the mS= +1/2 and mS= −1/2 electron
spin levels of substitutional N impurities, see Fig. 3(b).
The magnetic dipolar coupling leads to resonant spin ex-
change at this field, and this could be the cause of the
observed reduction of coherence.
In order to find evidence for this hypothesis, we com-
pare the magnetic field dependences of 1/T′
At the resonance condition, the magnetic dipolar cou-
pling leads to spin exchange between the N-V center and
the N spins, and the resulting reduction in spin polar-
ization of the N-V center translates into a dip in IPL
(see also Refs. 12,14). As can be seen in Fig. 3(c), the
resonance has sidepeaks due to the strong hyperfine in-
teraction of the N impurity12,25. Fig. 3(d) shows 1/T′
over the same magnetic field range. Peaks in 1/T′
observed at exactly the same magnetic fields where IPL
dips, which demonstrates that the magnetic dipolar cou-
pling is indeed the cause of the reduced coherence.
Both the central peak in 1/T′
IPL are well fit to a Lorentzian (red lines in Figs. 3(c)-
(d)). The width of the dip in IPL is a bit larger than
the width of the peak in 1/T′
broadening induced by the continuous laser excitation.
Note that this broadening is absent in the measurement
2, since the laser is turned off during the spin ma-
Away from the energy resonances, the magnetic dipo-
lar coupling becomes inefficient in exchanging spins and
2goes down. However, the N spins can still contribute
to decay of the Rabi nutations via the z-component of the
magnetic dipolar field that they create at the site of the
N-V center. If an N electron flips its spin, this magnetic
field is reversed. Such events shift the total magnetic field
felt by the N-V center and lead to dephasing. The more
2and the central dip in
2, which is likely due to
N spins are nearby, the more fluctuations in the magnetic
field and the stronger the decay. Therefore we expect a
relation between T′
2and the strength of the dipolar fields
if the N spins are also the dominant source of decoherence
away from the resonances. We use the amplitude of the
dips in IPLat B=514 G as a measure for the strength of
the dipolar fields at a given N-V center. In Fig. 3(e) we
2versus the dip amplitude normalized to the off-
resonance IPL. The five centers investigated follow the
expected trend, suggesting that dephasing due to N elec-
tron spin flips is the main decoherence mechanism away
from the energy resonances.
Our results on single centers confirm and extend the
findings from previous measurements on large ensembles
of N-V centers. Hiromitsu and coworkers observed a rise
in the Hahn echo decay rate around 514 G26. Kennedy
and coworkers found that the Hahn echo decay time was
shorter for samples with a higher density of N impuri-
ties5. Our results show that the local density of N spins
determines the coherence time of single centers in our
sample, with corresponding fluctuations observed from
center to center.
In summary, we have demonstrated the ability to co-
herently manipulate the spin state of a single N-V center
in diamond and probe its coherence. Using this tool,
we have investigated the dependence of the coherence
on magnetic field, showing that the magnetic dipolar
coupling to substitutional N impurities is the dominant
source of decoherence in this diamond. As a next step,
coherent two-spin operations can be envisioned for N-V
centers that are strongly coupled to a single N spin13,14.
We thank J.M. Elzerman, R.J. Epstein and F.M.
Mendoza for discussions. This work was supported by
AFOSR, DARPA/MARCO, DARPA/CNID and ARO.
1M. Nielsen and I. Chuang, “Quantum Computation
and Quantum Information” (Cambridge University Press,
2Y. Nakamura, Yu. A. Pashkin and J. S. Tsai, Nature 398,
3J. R. Petta, A. C. Johnson, J. M. Taylor, E. A. Laird, A.
Yacoby, M. D. Lukin, C. M. Marcus, M. P. Hanson and A.
C. Gossard, Science 309, 2180 (2005).
4G. Davies and M. F. Hamer, Proc. R. Soc. A 348, 285
5T. A. Kennedy, J. S. Colton, J. E. Butler, R. C. Linares
and P. J. Doering, Appl. Phys. Lett. 83, 4190 (2003).
6E. van Oort, N. B. Manson and M. Glasbeek, J. Phys. C
21, 4385 (1988).
7F. T. Charnock and T. A. Kennedy, Phys Rev. B 64,
8F. Jelezko, T. Gaebel, I. Popa, A. Gruber and J.
Wrachtrup, Phys. Rev. Lett. 92, 076401 (2004).
9F. Jelezko, T. Gaebel, I. Popa, M. Domhan, A. Gruber
and J. Wrachtrup, Phys. Rev. Lett. 93, 130501 (2004).
10I. Popa, T. Gaebel, M. Domhan, C. Wittmann, F. Jelezko
and J. Wrachtrup, Phys. Rev. B. 70, 201203(R) (2004).
11A. Gruber, A. Draebenstedt, C. Tietz, L. Fleury, J.
Wrachtrup and C. von Borczyskowski, Science 276, 2012
12R. J. Epstein, F. M. Mendoza, Y. K. Kato and D. D.
Awschalom, Nature Phys. 1, 94 (2005).
13T. Gaebel, M. Domhan, I. Popa, C. Wittmann, P. Neu-
mann, F. Jelezko, J.R. Rabeau, N. Stavrias, A. D. Green-
tree, S. Prawer, J. Meijer, J. Twamley, P. R. Hemmer and
J. Wrachtrup , Nature Phys. 2, 408 (2006).
14R. Hanson, F. M. Mendoza, R. J. Epstein and D. D.
Awschalom, Phys. Rev. Lett. 97, 087601 (2006).
15N. R. S. Reddy, N. B. Manson and E. R. Krausz, J. Lu-
min. 38, 46 (1987); E. van Oort, N. B. Manson, and M.
Glasbeek, J. Phys. C 21, 4385 (1988); D. A. Redman, S.
Brown, R. H. Sands, S. C. Rand, Phys. Rev. Lett. 67, 3420
16J. H. N. Loubser and J. A. vanWyk, Rep. Prog. Phys. 41,
17See e.g. J. Harrison, M. J. Sellers and N. B. Manson, J.
Lumin. 107, 245 (2004).
18F. Jelezko, I. Popa, A. Gruber, C. Tietz, and J. Wrachtrup,
A. Nizovtsev and S. Kilin, Appl. Phys. Lett. 81, 2160
19N. B. Manson, J. P. Harrison, and M. J. Sellars, Phys. Rev.
B 74, 104303 (2006).
20A. P. Nizovtsev, S. Ya. Kilin, F. Jelezko, I. Popa, A. Gruber
and J. Wrachtrup , Physica B 340-342, 106 (2003).
21Multiple synchronized outputs of a pattern generator are
used to control the laser via an Acoustic-Optical modula-
tor, to gate the photon counter and to trigger an arbitrary
waveform generator that in turn gates the high-frequency
source leading to the desired sequence of RF bursts.
22J. J. Sakurai, “Modern Quantum Mechanics” (Addisson-
23L. M. K. Vandersypen and I. L. Chuang, Rev. Mod. Phys.
76, 1037 (2004).
24E. van Oort, P. Stroomer and M. Glasbeek, Phys. Rev. B
42, 8605 (1990).
25W. V. Smith, P. P. Sorokin, I. L. Gelles, and G. J. Lasher,
Phys. Rev. 115, 1546 (1959).
26I. Hiromitsu, J. Westra, and M. Glasbeek, Phys. Rev. B
46, 10600 (1992)