news & views high-intensity regions (Fig. 1a). However, due to the small size of the nanoparticle the gradient force remains weak and Brownian motion within the fluid is constantly nudging it in random directions, thus hindering its trapping. As the optical force is proportional to the irradiance of light (intensity per unit of surface area), it could be possible to increase the power of the laser to overcome Brownian motion. However, a high irradiance could damage or even destroy the nanoparticle. Hence, the first challenge is to generate an optical trap for a moderate irradiance. A second, more subtle issue is that to trap a single nanoparticle the size of the trap should be of the order of the nanoparticle size. Such a small optical trap would be so sensitive to its environment that the interaction with the single nanoparticle that should be trapped may perturb it enough to destroy its trapping ability. The technique presented by Quidant and co-workers overcomes these problems by using a modified near-field scanning optical microscope (NSOM) probe to create a subwavelength optical trap (Fig. 1b) and exploit, rather than fight, the trap’s high sensitivity to its environment. The idea of using a NSOM probe for optical manipulation is not new 5,6. The implementation of this idea, however, was hindered by the challenge of obtaining a strong enough trapping potential at a reasonable irradiance. The solution was to introduce a bowtie aperture at the end of the near-field probe (Fig. 2). Whereas, a NSOM probe usually consists of a chemically etched, metal-coated optical fibre with a round opening at the end, the authors’ probe has a bowtie opening at its apex. The bowtie acts as a plasmonic nanoantenna, confining light on a subwavelength scale. The associated large gradient of the electric field
Figure 2 | Schematic of the optical nanotweezers. A nanoparticle is trapped near a bowtie plasmonic aperture at the end of a near-field optical probe. Image courtesy of Johann Berthelot.
ensures that, when light is funnelled down the probe, an efficient optical trap is created at the apex. To make sure that the sensitivity of such a trap could be used as an advantage, Quidant and co-workers modified the bowtie nanoantenna such that its resonance is slightly blue-shifted compared with the trapping wavelength. This way only when a nanoparticle is present near the bowtie opening is the optical trap activated. The result is an efficient optical trap for a relatively weak irradiance. Because the trap is formed at the tip of a NSOM probe, the trapped nanoparticle can be moved at will, and released by turning the illumination off. The next challenge is to selectively trap a single nanoparticle. In the current scheme a particle has to wander near the bowtie to
activate the optical trap, many applications would require an ability to locate a specific particle and capture it. If the nanoparticles are on a substrate, a bowtie nanoantenna probe can be a very effective, polarization-sensitive imaging device7. Local spectroscopy via the probe could provide a way to discriminate between different types of nanoparticle. This would be particularly interesting with metallic nanoparticles. The nanoantenna at the end of the NSOM probe could be tailored to a particular plasmon resonance for materialselective, or even size-selective trapping. We can also envisage an array of near-field probes8 to capture and manipulate particles attached to different antigen–antibody complexes or DNA strands. As a result of the nanoantennas being at the apices of the probes, a nanoscale immunoassay or DNA biochip based on the selective capture and the parallel detection of biomolecules could be possible. ❐ Patrick C. Chaumet is at Aix Marseille Université, CNRS, Centrale Marseille, Institut Fresnel, UMR 7249, 13013 Marseille, France. Adel Rahmani is at the School of Mathematical Sciences, University of Technology Sydney, Broadway, New South Wales 2007, Australia. e-mail: [email protected]; [email protected] References 1. Ashkin, A. Phys. Rev. Lett. 24, 156–159 (1970). 2. Ashkin, A., Dziedzic, J. M. & Yamane, T. Nature 330, 769–771 (1987). 3. Ashkin, A. & Dziedzic, J. M. Science 235, 1517–1520 (1987). 4. Berthelot, J. et al. Nature Nanotech. 9, 295–299 (2014). 5. Novotny, L., Bian, R. X. & Xie, X. S. Phys. Rev. Lett. 79, 645–648 (1997). 6. Chaumet, P. C., Rahmani, A. & Nieto-Vesperinas, M. Phys. Rev. Lett. 88, 123601 (2002). 7. Vo, T.‑P. et al. Opt. Express 20, 4124–4135 (2012). 8. Chovin, A., Garrigue, P., Manek-Honninger, I. & Sojic, N. Nano Lett. 4, 1965–1968 (2004).
NANOSCALE MRI
Dark spins in the spotlight
A single nitrogen–vacancy centre can be used to probe the location of electron spins with subnanometre precision.
Lloyd Hollenberg
T
he material and quantum properties of the negatively charged nitrogen– vacancy (NV) centre in diamond provide a range of quantum technology possibilities, including nanoscale sensing and imaging1,2. Remarkably, the electronic spin states of NV centres have relatively long quantum coherence at room temperature,
making them a sensitive probe for a variety of external perturbations. The NV centre can be tracked and read-out optically; its position, local magnetic and electric fields, and temperature can all be monitored through quantum control via the application of microwaves. Furthermore, these properties lend
NATURE NANOTECHNOLOGY | VOL 9 | APRIL 2014 | www.nature.com/naturenanotechnology
© 2014 Macmillan Publishers Limited. All rights reserved
themselves to biosensing applications as diamond is relatively bio-friendly 3–7. A number of papers have reported the use of NVs as sensors of electronic and nuclear spins in nanoscale volumes8–13, including in biological contexts14–16. Writing in Nature Nanotechnology, Mike Grinolds and co-workers from Harvard University, 253
news & views NV electronic and spin levels, optical readout/polarization Magnetic field dependence of spin level transitions
V
N
System crossing
FL
532 nm
+1
±1 → 0
637 nm
2.87 GHz
Low FL
−1 MW
±1
High FL
0
0
B-field
Spin states Magnetic field detection using the NV centre Microwave control
Magnetic field target, B NV
Laser
FL
Diamond host Ψ
Protocol cycle
Initialization
NV spin read-out via FL
Ψ' Exposure/control
B-field signal
Readout
Figure 1 | Magnetometry using the nitrogen–vacancy (NV) centre in diamond. The NV centre’s ground state is a spin-1 triplet of states (top left), which under excitation with green light can be distinguished via the fluorescence (FL) and be initialized into the |0〉 state. In a magnetic field the |±1〉 spin-projection states split and can be accessed via microwaves (MW) at the corresponding resonant transitions (top right). A typical measurement of magnetic field involves a cycle of initialization into a particular quantum state ψ, control and exposure to the field, and finally readout of the resulting state providing the B-field signal (lower schematic).
RF control of dark spins (ωDEER)
Scanning magnetic tip
Resonant nt ce field slice
Resonant dark spins Non-resonant dark spins
Dipolar coupling NV centre
Readout off NV centre NV centre entre tre
MW control of NV centre ceentre (ω (ωNV)
Figure 2 | Nano-MRI set-up17. A scanning magnetic tip on an atomic force microscope produces a spatial region of target spins resonant with an applied radiofrequency (RF) driving field (at frequency ωDEER). The signal corresponding to the volume of excited dark spins is detected by the NV magnetometer probe (controlled by microwaves at frequency ωNV) in the diamond substrate below.
University of Basel, the HarvardSmithsonian Center for Astrophysics and the Vienna Center for Quantum Science and Technology report the demonstration of three-dimensional magnetic resonance 254
imaging of individual electron spins near the surface of diamond with subnanometre resolution17. This work extends the imaging capability of the NV centre towards ambient nanoscale nuclear magnetic resonance. It
also shines light on one of the main factors limiting the efficacy of NV sensing, that is, spins associated with the surface of diamond that are a ubiquitous source of decoherence of the NV spin states. The detection of magnetic fields using NV centres (Fig. 1) is based on the fact that the ground state of the NV is a spin-1 triplet. The state with zero spin projection, |0〉, has a zero-field splitting of 2.87 GHz from the states with non-zero spin projection, |±1〉, which means that transitions between the states can be induced and controlled by applying resonant microwaves. Under excitation with green light, the spin states can be distinguished by the variation in fluorescence from the brighter |0〉 state compared with the darker |±1〉 states. Following this readout procedure the NV ends up in the |0〉 state. This microwave control and spin-dependent difference in fluorescence is the basis of optically detected magnetic resonance (ODMR)18, and magnetic-field detection schemes. In its simplest terms, detection of magnetic fields occurs through a process of initialization, exposure and readout. For the detection of a static magnetic field, in the first step the NV centre is illuminated with green light, which initializes the system into the |0〉 ground state. For example, a π/2 microwave pulse places the system into a superposition of |0〉 and |+1〉 states. The system is then left to evolve in the magnetic field, so that the |+1〉 component acquires a phase, whereas the |0〉 does not. After an exposure time τ (the so-called free induction time in this example), a second π/2 pulse converts the magnetic-field-dependent phase accumulated in the |+1〉 state into a population difference, which can be readout through fluorescence measurements. By adding a π pulse in the midpoint of the protocol, oscillatory magnetic fields can be detected with even higher sensitivity. Critically, the decoherence time of the centre controls the exposure time and hence the sensitivity of detection. In the case of the π/2, τ, π/2 sequence, known as a Ramsey sequence, the free induction decay time T2* limits the sensitivity to static (d.c.) fields, whilst the sensitivity of the π/2, τ/2, π, τ/2, π/2 sequence known as a Hahn spin-echo sequence, is limited by the T2 decoherence timescale (usually much longer than T2*), and is hence more sensitive. More complicated control schemes can reach higher sensitivities by extending the quantum coherence of the NV, and the decoherence profile itself can be used to detect fluctuating fields with zero average. A single NV centre is not the most sensitive magnetometer, but it has the advantage of being an atomic size probe,
NATURE NANOTECHNOLOGY | VOL 9 | APRIL 2014 | www.nature.com/naturenanotechnology
© 2014 Macmillan Publishers Limited. All rights reserved
news & views and can be brought close, and with enough sensitivity, to nanoscale volumes, and operates at ambient temperatures. For some applications, sensitivity can be enhanced through a multiplicity of NV sensors. As the decoherence time of the NV centre determines the overall sensitivity to magnetic fields, it is a key factor governing its utility. This is a crucial aspect of the work by Grinolds and co-workers. The effects of inherent sources of decoherence such as carbon-13 nuclear spins and nitrogen donor electron spins can be reduced in highpurity diamond. However, as the NV sensor needs to be very close to the target, this typically means it is close to the diamond surface where maverick electron spins (dark spins) abound and are a major source of decoherence limiting the sensitivity of the NV. Although these spins have been probed in various detection experiments8–10, Grinolds et al. have actually imaged the location of these culprit dark spins with high resolution. They were also able to image the location of an isolated spin near a NV centre. In the set-up used by Grinolds and coworkers a magnetic atomic force microscope tip is brought into close proximity to a single NV centre that was implanted about 10 nm below the surface of a diamond substrate (Fig. 2). The magnetic field from the scanning tip splits the Zeeman levels of the target electrons. Because the magnetic
field felt by each electron is slightly different, however, the splitting varies according to their position, in analogy to conventional macroscopic magnetic resonance imaging. It is therefore possible to excite a very small volume of electron spins by applying a radiofrequency field with a frequency matching only the Zeeman splitting in that volume. Using a protocol known as double electron–electron resonance (DEER) the spin in the NV centre acquires a phase corresponding to the driven target spins in the resonant volume. Grinolds and coworkers were able to use this information to extract the spatial distribution of the spins in the very small target volume. The result is a beautiful image of the location of the dark spins, forming a layer located about 10 nm above the NV sensor, consistent with the nominal NV depth, which shows that these spins are associated with the diamond surface. The data is of sufficient quality to extract the effective density, of the surface spins (~0.5 spins nm–2) — an important piece of information required for the detailed understanding of the dynamics of this spin-bath10. A further step of the process allowed imaging of the location of a single dark spin coupled to the NV sensor to within subnanometre lateral resolution. The results demonstrate threedimensional electron-spin imaging to subnanometre resolution, which in itself could have a range of applications.
NATURE NANOTECHNOLOGY | VOL 9 | APRIL 2014 | www.nature.com/naturenanotechnology
© 2014 Macmillan Publishers Limited. All rights reserved
Furthermore, they provide quantitative information on dark spins, which are a key limiter to the ultimate aspirations of threedimensional nuclear spin imaging. These results are therefore a major step towards practical implementations of nanoscale MRI using NV centres. ❐ Lloyd Hollenberg is in the Centre for Quantum Computation & Communication Technology, School of Physics University of Melbourne, Victoria 3010, Australia. e-mail: [email protected] References 1. Doherty, M. W. et al. Phys. Rep. 528, 1–45 (2013). 2. Mater. Res. Soc. Bull. 38 (Nitrogen-Vacancy Centers: Physics and Applications special issue), 127–167 (2013). 3. Neugart, F. et al. Nano Lett. 7, 3588–3591 (2007). 4. Mohan, N., Chen, C.-S., Hsieh, H.-H., Wu, Y.-C. & Chang, H.-H. Nano Lett. 10, 3692–3699 (2010). 5. McGuinness, L. P. et al. Nature Nanotech. 6, 358–363 (2011). 6. Le Sage, D. et al. Nature 496, 486–489 (2013). 7. Kucsko, G. et al. Nature 500, 54–58 (2013). 8. Grotz, B. et al. New J. Phys. 13, 055004 (2011). 9. Mamin, H. J., Sherwood, M. H. & Rugar, D. Phys. Rev. B86, 195422 (2012). 10. McGuinness, L. P. et al. New J. Phys. 15, 073042 (2013). 11. Tetienne, J.-P. et al. Phys. Rev. B87, 235436 (2013). 12. Staudacher, T. Science 339, 561–563 (2013). 13. Mamin, H. J. et al. Science 339, 557–560 (2013). 14. Kaufmann, S. et al. Proc. Natl Acad. Sci. USA 110, 10894–10898 (2013). 15. Ermakova, A. et al. Nano Lett. 13, 3305–3309 (2013). 16. Steinert, S. et al. Nature Commun. 4, 1607 (2013). 17. Grinolds, M. S. et al. Nature Nanotech. 9, 279–284 (2014). 18. Gruber, A. et al. Science 276, 2012–2014 (1997).
Published online: 23 March 2014
255
Figure 2 | Schematic of the optical nanotweezers. A nanoparticle is trapped near a bowtie plasmonic aperture at the end of a near-field optical probe. Image courtesy of Johann Berthelot.
ensures that, when light is funnelled down the probe, an efficient optical trap is created at the apex. To make sure that the sensitivity of such a trap could be used as an advantage, Quidant and co-workers modified the bowtie nanoantenna such that its resonance is slightly blue-shifted compared with the trapping wavelength. This way only when a nanoparticle is present near the bowtie opening is the optical trap activated. The result is an efficient optical trap for a relatively weak irradiance. Because the trap is formed at the tip of a NSOM probe, the trapped nanoparticle can be moved at will, and released by turning the illumination off. The next challenge is to selectively trap a single nanoparticle. In the current scheme a particle has to wander near the bowtie to
activate the optical trap, many applications would require an ability to locate a specific particle and capture it. If the nanoparticles are on a substrate, a bowtie nanoantenna probe can be a very effective, polarization-sensitive imaging device7. Local spectroscopy via the probe could provide a way to discriminate between different types of nanoparticle. This would be particularly interesting with metallic nanoparticles. The nanoantenna at the end of the NSOM probe could be tailored to a particular plasmon resonance for materialselective, or even size-selective trapping. We can also envisage an array of near-field probes8 to capture and manipulate particles attached to different antigen–antibody complexes or DNA strands. As a result of the nanoantennas being at the apices of the probes, a nanoscale immunoassay or DNA biochip based on the selective capture and the parallel detection of biomolecules could be possible. ❐ Patrick C. Chaumet is at Aix Marseille Université, CNRS, Centrale Marseille, Institut Fresnel, UMR 7249, 13013 Marseille, France. Adel Rahmani is at the School of Mathematical Sciences, University of Technology Sydney, Broadway, New South Wales 2007, Australia. e-mail: [email protected]; [email protected] References 1. Ashkin, A. Phys. Rev. Lett. 24, 156–159 (1970). 2. Ashkin, A., Dziedzic, J. M. & Yamane, T. Nature 330, 769–771 (1987). 3. Ashkin, A. & Dziedzic, J. M. Science 235, 1517–1520 (1987). 4. Berthelot, J. et al. Nature Nanotech. 9, 295–299 (2014). 5. Novotny, L., Bian, R. X. & Xie, X. S. Phys. Rev. Lett. 79, 645–648 (1997). 6. Chaumet, P. C., Rahmani, A. & Nieto-Vesperinas, M. Phys. Rev. Lett. 88, 123601 (2002). 7. Vo, T.‑P. et al. Opt. Express 20, 4124–4135 (2012). 8. Chovin, A., Garrigue, P., Manek-Honninger, I. & Sojic, N. Nano Lett. 4, 1965–1968 (2004).
NANOSCALE MRI
Dark spins in the spotlight
A single nitrogen–vacancy centre can be used to probe the location of electron spins with subnanometre precision.
Lloyd Hollenberg
T
he material and quantum properties of the negatively charged nitrogen– vacancy (NV) centre in diamond provide a range of quantum technology possibilities, including nanoscale sensing and imaging1,2. Remarkably, the electronic spin states of NV centres have relatively long quantum coherence at room temperature,
making them a sensitive probe for a variety of external perturbations. The NV centre can be tracked and read-out optically; its position, local magnetic and electric fields, and temperature can all be monitored through quantum control via the application of microwaves. Furthermore, these properties lend
NATURE NANOTECHNOLOGY | VOL 9 | APRIL 2014 | www.nature.com/naturenanotechnology
© 2014 Macmillan Publishers Limited. All rights reserved
themselves to biosensing applications as diamond is relatively bio-friendly 3–7. A number of papers have reported the use of NVs as sensors of electronic and nuclear spins in nanoscale volumes8–13, including in biological contexts14–16. Writing in Nature Nanotechnology, Mike Grinolds and co-workers from Harvard University, 253
news & views NV electronic and spin levels, optical readout/polarization Magnetic field dependence of spin level transitions
V
N
System crossing
FL
532 nm
+1
±1 → 0
637 nm
2.87 GHz
Low FL
−1 MW
±1
High FL
0
0
B-field
Spin states Magnetic field detection using the NV centre Microwave control
Magnetic field target, B NV
Laser
FL
Diamond host Ψ
Protocol cycle
Initialization
NV spin read-out via FL
Ψ' Exposure/control
B-field signal
Readout
Figure 1 | Magnetometry using the nitrogen–vacancy (NV) centre in diamond. The NV centre’s ground state is a spin-1 triplet of states (top left), which under excitation with green light can be distinguished via the fluorescence (FL) and be initialized into the |0〉 state. In a magnetic field the |±1〉 spin-projection states split and can be accessed via microwaves (MW) at the corresponding resonant transitions (top right). A typical measurement of magnetic field involves a cycle of initialization into a particular quantum state ψ, control and exposure to the field, and finally readout of the resulting state providing the B-field signal (lower schematic).
RF control of dark spins (ωDEER)
Scanning magnetic tip
Resonant nt ce field slice
Resonant dark spins Non-resonant dark spins
Dipolar coupling NV centre
Readout off NV centre NV centre entre tre
MW control of NV centre ceentre (ω (ωNV)
Figure 2 | Nano-MRI set-up17. A scanning magnetic tip on an atomic force microscope produces a spatial region of target spins resonant with an applied radiofrequency (RF) driving field (at frequency ωDEER). The signal corresponding to the volume of excited dark spins is detected by the NV magnetometer probe (controlled by microwaves at frequency ωNV) in the diamond substrate below.
University of Basel, the HarvardSmithsonian Center for Astrophysics and the Vienna Center for Quantum Science and Technology report the demonstration of three-dimensional magnetic resonance 254
imaging of individual electron spins near the surface of diamond with subnanometre resolution17. This work extends the imaging capability of the NV centre towards ambient nanoscale nuclear magnetic resonance. It
also shines light on one of the main factors limiting the efficacy of NV sensing, that is, spins associated with the surface of diamond that are a ubiquitous source of decoherence of the NV spin states. The detection of magnetic fields using NV centres (Fig. 1) is based on the fact that the ground state of the NV is a spin-1 triplet. The state with zero spin projection, |0〉, has a zero-field splitting of 2.87 GHz from the states with non-zero spin projection, |±1〉, which means that transitions between the states can be induced and controlled by applying resonant microwaves. Under excitation with green light, the spin states can be distinguished by the variation in fluorescence from the brighter |0〉 state compared with the darker |±1〉 states. Following this readout procedure the NV ends up in the |0〉 state. This microwave control and spin-dependent difference in fluorescence is the basis of optically detected magnetic resonance (ODMR)18, and magnetic-field detection schemes. In its simplest terms, detection of magnetic fields occurs through a process of initialization, exposure and readout. For the detection of a static magnetic field, in the first step the NV centre is illuminated with green light, which initializes the system into the |0〉 ground state. For example, a π/2 microwave pulse places the system into a superposition of |0〉 and |+1〉 states. The system is then left to evolve in the magnetic field, so that the |+1〉 component acquires a phase, whereas the |0〉 does not. After an exposure time τ (the so-called free induction time in this example), a second π/2 pulse converts the magnetic-field-dependent phase accumulated in the |+1〉 state into a population difference, which can be readout through fluorescence measurements. By adding a π pulse in the midpoint of the protocol, oscillatory magnetic fields can be detected with even higher sensitivity. Critically, the decoherence time of the centre controls the exposure time and hence the sensitivity of detection. In the case of the π/2, τ, π/2 sequence, known as a Ramsey sequence, the free induction decay time T2* limits the sensitivity to static (d.c.) fields, whilst the sensitivity of the π/2, τ/2, π, τ/2, π/2 sequence known as a Hahn spin-echo sequence, is limited by the T2 decoherence timescale (usually much longer than T2*), and is hence more sensitive. More complicated control schemes can reach higher sensitivities by extending the quantum coherence of the NV, and the decoherence profile itself can be used to detect fluctuating fields with zero average. A single NV centre is not the most sensitive magnetometer, but it has the advantage of being an atomic size probe,
NATURE NANOTECHNOLOGY | VOL 9 | APRIL 2014 | www.nature.com/naturenanotechnology
© 2014 Macmillan Publishers Limited. All rights reserved
news & views and can be brought close, and with enough sensitivity, to nanoscale volumes, and operates at ambient temperatures. For some applications, sensitivity can be enhanced through a multiplicity of NV sensors. As the decoherence time of the NV centre determines the overall sensitivity to magnetic fields, it is a key factor governing its utility. This is a crucial aspect of the work by Grinolds and co-workers. The effects of inherent sources of decoherence such as carbon-13 nuclear spins and nitrogen donor electron spins can be reduced in highpurity diamond. However, as the NV sensor needs to be very close to the target, this typically means it is close to the diamond surface where maverick electron spins (dark spins) abound and are a major source of decoherence limiting the sensitivity of the NV. Although these spins have been probed in various detection experiments8–10, Grinolds et al. have actually imaged the location of these culprit dark spins with high resolution. They were also able to image the location of an isolated spin near a NV centre. In the set-up used by Grinolds and coworkers a magnetic atomic force microscope tip is brought into close proximity to a single NV centre that was implanted about 10 nm below the surface of a diamond substrate (Fig. 2). The magnetic field from the scanning tip splits the Zeeman levels of the target electrons. Because the magnetic
field felt by each electron is slightly different, however, the splitting varies according to their position, in analogy to conventional macroscopic magnetic resonance imaging. It is therefore possible to excite a very small volume of electron spins by applying a radiofrequency field with a frequency matching only the Zeeman splitting in that volume. Using a protocol known as double electron–electron resonance (DEER) the spin in the NV centre acquires a phase corresponding to the driven target spins in the resonant volume. Grinolds and coworkers were able to use this information to extract the spatial distribution of the spins in the very small target volume. The result is a beautiful image of the location of the dark spins, forming a layer located about 10 nm above the NV sensor, consistent with the nominal NV depth, which shows that these spins are associated with the diamond surface. The data is of sufficient quality to extract the effective density, of the surface spins (~0.5 spins nm–2) — an important piece of information required for the detailed understanding of the dynamics of this spin-bath10. A further step of the process allowed imaging of the location of a single dark spin coupled to the NV sensor to within subnanometre lateral resolution. The results demonstrate threedimensional electron-spin imaging to subnanometre resolution, which in itself could have a range of applications.
NATURE NANOTECHNOLOGY | VOL 9 | APRIL 2014 | www.nature.com/naturenanotechnology
© 2014 Macmillan Publishers Limited. All rights reserved
Furthermore, they provide quantitative information on dark spins, which are a key limiter to the ultimate aspirations of threedimensional nuclear spin imaging. These results are therefore a major step towards practical implementations of nanoscale MRI using NV centres. ❐ Lloyd Hollenberg is in the Centre for Quantum Computation & Communication Technology, School of Physics University of Melbourne, Victoria 3010, Australia. e-mail: [email protected] References 1. Doherty, M. W. et al. Phys. Rep. 528, 1–45 (2013). 2. Mater. Res. Soc. Bull. 38 (Nitrogen-Vacancy Centers: Physics and Applications special issue), 127–167 (2013). 3. Neugart, F. et al. Nano Lett. 7, 3588–3591 (2007). 4. Mohan, N., Chen, C.-S., Hsieh, H.-H., Wu, Y.-C. & Chang, H.-H. Nano Lett. 10, 3692–3699 (2010). 5. McGuinness, L. P. et al. Nature Nanotech. 6, 358–363 (2011). 6. Le Sage, D. et al. Nature 496, 486–489 (2013). 7. Kucsko, G. et al. Nature 500, 54–58 (2013). 8. Grotz, B. et al. New J. Phys. 13, 055004 (2011). 9. Mamin, H. J., Sherwood, M. H. & Rugar, D. Phys. Rev. B86, 195422 (2012). 10. McGuinness, L. P. et al. New J. Phys. 15, 073042 (2013). 11. Tetienne, J.-P. et al. Phys. Rev. B87, 235436 (2013). 12. Staudacher, T. Science 339, 561–563 (2013). 13. Mamin, H. J. et al. Science 339, 557–560 (2013). 14. Kaufmann, S. et al. Proc. Natl Acad. Sci. USA 110, 10894–10898 (2013). 15. Ermakova, A. et al. Nano Lett. 13, 3305–3309 (2013). 16. Steinert, S. et al. Nature Commun. 4, 1607 (2013). 17. Grinolds, M. S. et al. Nature Nanotech. 9, 279–284 (2014). 18. Gruber, A. et al. Science 276, 2012–2014 (1997).
Published online: 23 March 2014
255
