Materials for Quantum

Technology

PAPER • OPEN ACCESS You may also like

- Characterisation of CVD diamond with
Optimizing ion implantation to create shallow NV high concentrations of nitrogen for
magnetic-field sensing applications
centre ensembles in high-quality CVD diamond Andrew M Edmonds, Connor A Hart,
Matthew J Turner et al.

- Fiber-coupled quantum light sources

To cite this article: Midrel Wilfried Ngandeu Ngambou et al 2022 Mater. Quantum. Technol. 2 045001 based on solid-state quantum emitters
Lucas Bremer, Sven Rodt and Stephan
Reitzenstein

- Anti-Stokes excitation of optically active

point defects in semiconductor materials
View the article online for updates and enhancements. Wu-Xi Lin, Jun-Feng Wang, Qiang Li et al.

This content was downloaded from IP address 189.216.232.142 on 12/01/2025 at 00:37

 Mater. Quantum Technol. 2 (2022) 045001 https://doi.org/10.1088/2633-4356/ac9948

PAPER

Optimizing ion implantation to create shallow NV centre

O P E N AC C E S S
ensembles in high-quality CVD diamond
R E C E IVE D
3 August 2022
Midrel Wilfried Ngandeu Ngambou1,∗ , Pauline Perrin2 , Ionut Balasa2 ,
R E VISE D
5 October 2022
Ovidiu Brinza1 , Audrey Valentin1 , Vianney Mille1 , Fabien Bénédic1 ,
Philippe Goldner2 , Alexandre Tallaire1,2 and Jocelyn Achard1,∗
AC C E PTE D FOR PUBL IC ATION
1
11 October 2022 LSPM, CNRS, Université Sorbonne Paris Nord, 99 Avenue JB Clément 93460, Villetaneuse, France
2
Chimie ParisTech, PSL Research University, CNRS, Institut de Recherche de Chimie Paris, 75005 Paris, France
PUBL ISHE D ∗
14 November 2022 Authors to whom any correspondence should be addressed.
E-mail: [email protected] and [email protected]
Original content from Keywords: nitrogen-vacancy centres, ion implantation, single crystal diamond, shallow NV layer, CVD diamond film
this work may be used
under the terms of the
Creative Commons
Attribution 4.0 licence. Abstract
Any further distribution The negatively charged nitrogen-vacancy centre (so-called NV-centre) in diamond is one of the
of this work must
maintain attribution to most promising systems for applications in quantum technologies because of the possibility to
the author(s) and the
title of the work, journal optically manipulate and read out the spin state of this defect, even at room temperature.
citation and DOI. Nevertheless, obtaining high NV densities (>500 ppb) close to the surface (5–20 nm) while
maintaining good spin properties remain challenging. In this work we rely on a versatile ion
implantation system allowing both implanting nitrogen using N2 + and creating vacancies with
He+ ion bombardment at variable energies and fluence to create shallow NV ensembles. By
optimizing the ion irradiation conditions as well as the surface preparation prior to treatment we
successfully increase the amount of created colour centres while demonstrating narrow magnetic
resonance linewidths.

1. Introduction

The nitrogen-vacancy (NV) centre in diamond is a widely studied defect that possesses unique spin and optical
properties. It is indeed possible to optically initialize and readout the electronic spin state with long coher-
ence times up to several milliseconds even at room temperature as demonstrated in [1]. These specific spin
properties have paved the way to ultra-sensitive, high-performance and innovative quantum sensors (magne-
tometers, gyroscopes, spectrum analysers, etc) which could open up perspectives in investigating properties
that remain inaccessible by conventional devices. Several proposals for integrating NV centres in sensing and
quantum applications rely on their tailored fabrication in ultra-pure host material and have been used for sens-
ing magnetic fields, spin-based qubits for quantum information processing [2–4]. To develop such devices,
the diamond films must possess very high crystalline quality and the density, environment, orientation of the
introduced colour centres have to be perfectly controlled. Microwave plasma-assisted chemical vapour deposi-
tion (MPACVD) has been identified as a key technology to produce such engineered ‘quantum grade’ diamond
films [5]. Nevertheless, some of these applications require a spatial accuracy of a nanometre for positioning NV
centres close to the diamond surface, which can hardly be achieved by direct CVD growth. Another impor-
tant challenge is to create a high density of shallow NV centres that preserve long NV spin coherence times
[6]. Several methods have been studied to overcome this difficulty such as tapered nanopillars fabrication,
delta doping or diamond with an integrated RF micro-antenna for the soin manipulation of NV centre [7–9].
A well-established method to generate these centres consists in nitrogen ion implantation of a high purity
CVD film followed by thermal annealing [10]. This technique has been the subject of several studies [11, 12]
and a better understanding of the accumulation and generation of defects during ion implantation has been
established [13].
In this paper, we report an efficient process for creating shallow NV centres in diamond using an implanta-
tion system based on the use of a single cavity electron cyclotron resonance (ECR) source module for producing
relatively high energy ion beams (5–50 keV) using different gas sources. By combining nitrogen and helium ion

© 2022 The Author(s). Published by IOP Publishing Ltd

Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

implantation, NV centre ensembles have been produced at a depth of 30 nm. Their creation has been optimized
by choosing appropriate energies and fluence and their spin properties have been assessed by optically detected
magnetic resonance (ODMR). In addition, we explored different surface treatments of the diamond prior to
implantation such as surface terminations, polishing and subsequent reactive ion etching with inductively
coupled plasma (RIE-ICP) etching.

2. Experimental details

A schematic diagram of our experimental system is shown in figure 1(a). It consists of a secondary vacuum
system, a gas inlet, an implantation chamber and a Faraday cup, the latter allowing measuring ion current and
thus assessing fluence. Ion beam is generated in the chamber using an ECR plasma system. The gas flow injected
into the ECR source is controlled with a throttle valve and a pressure of about 10−5 mbar is maintained. The ion
fluence is adjusted by setting the microwave power of the ECR source between 0.3 and 4 W. We aim at fluence
below 2 × 1013 ions cm−2 , to avoid diamond graphitization reached at 1022 vac cm−3 [14, 15]. By applying
an acceleration voltage, the ion energy can be set from 5 to 50 keV while the beam focus can be adjusted with
another electrode that has to be biased to roughly 90% or less of the accelerating voltage. The beam size in this
experiment is about 0.8 cm2 , i.e. larger than the size of the single crystal diamonds that are used.
Prior to implantation, 20 μm-thick high purity CVD diamond layers have been grown on type Ib high
pressure high temperature (100)-oriented diamond substrates by MPACVD [16]. Following growth, the sur-
face of the samples has been conditioned following two different treatments: (i) a hydrogen plasma performed
in the CVD chamber to allow for H-termination or (ii) an ozone treatment using a deep-UV light to achieve
O-termination except for sample 5 that was submitted to a low-power microwave plasma oxygen treatment
(see table 1). To further assess the effect of surface states on ion implantation, another set of diamond films
have been polished using a standard ‘Scaife’ system to a roughness of the order of Ra = 0.1 μm over a sur-
face of 100 × 100 μm2 as measured by optical profilometry. It has been followed by a tri-acid cleaning of
H2 SO4 :HClO4 :HNO3 (2:3:3). On one sample, an RIE-ICP etching with a gas mixture of Ar:O2 (1:1) at a tem-
perature of 20 ◦ C and a pressure of 4 mbar has been performed during 7 min in order to remove around
2 μm of material off the surface. The objective of this treatment is to estimate both the impact of polish-
ing on NV centres formation and the possibility to limit this impact by removing the hardened zone created
by polishing before implantation. After polishing or ion etching, all samples were hydrogen terminated and
implanted. Table 1 summarizes the surface treatments carried out on the CVD diamond layers before and
after implantation. After implantation all samples were oxygen terminated in order to stabilize the shallow NV
centres.
The samples have then been exposed successively to nitrogen and helium ion beams in our implantation
system. The low energy of the ECR microwave source does not allow for a complete dissociation of the stable N2
molecules and therefore mostly N2 + ions are created in the plasma and accelerated. These ions have been used
to introduce nitrogen inside the diamond, but they also generate vacancies. The second ion beam treatment
with He+ ions only produces vacancies and has been used to optimize substitutional nitrogen (NS ) to NV
conversion. Based on SRIM simulations [17], the penetration depth of those ions has been plotted in figure 1(b)
as a function of acceleration voltage. In order to implant N and He at a similar depth of 30 nm, we have selected
a voltage of 40 keV for N2 + (i.e. 20 keV for individual N atoms) and 5 keV for He+ ions respectively.
The effect of helium has been investigated on sample No. 5 starting from hydrogen terminated ultrapure
CVD diamond film firstly implanted using nitrogen ions for 10 s over the entire surface. Then, helium has been
implanted in three different areas of the same sample during respectively 10, 20 and 30 s. To screen untreated
areas from the beam, we used a physical shadow mask placed onto the sample. The implantation conditions
are given in table 1 and a scheme of the implantation distribution for this sample is shown in figure 2.
Finally, the samples have been annealed under high vacuum (10−6 mbar) at 800 ◦ C for 2 h in order to
repair the damage induced by implantation and allow vacancies migrating to pair with nitrogen [17]. Figure 3
summarises the experimental process used to create NV centres.
Raman and photoluminescence (PL) measurements under optical excitation at 473 nm and 532 nm have
been performed in a non-confocal mode using a Horiba-HR800 spectrometer. A DiamondViewTM instru-
ment has been used to observe the fluorescence and phosphorescence. Confocal laser scanning microscope
has been used to perform images. ODMR measurements have been performed by a confocal microscopy set-
up in order to access the spin properties and measure coherence times of different implanted samples. A high
numerical aperture objective (NA = 0.95) has been used to focus the excitation green laser (532 nm) on the
sample and collect the PL from the NV− centres. The PL signal was spatially filtered by a 50 μm-pinhole
and finally recorded by a single photon counter detector (Laser Components COUNT-10C). Additionally, a
microwave field with a tuneable frequency was generated by a 100 μm diameter wire placed near the sample

2
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

Figure 1. (a) Schematic of the implantation system. (b) SRIM simulations giving the penetration depth of nitrogen and helium
ion in diamond as a function of ion the beam energy. The dotted lines given the implantation energies of nitrogen and Helium
corresponding to the same penetration depth used.

Table 1. Surface treatments before implantation and irradiation conditions.

Samples Surface treatment before Irradiation Fluence (ions cm−2 ) Surface treatment
implantation conditions after implantation

1 As-grown/hydrogen N2 + 10 s 2 × 1013 Oxygen termination

termination by ozone
2 As-grown/oxygen N2 + 10 s 2 × 1013 Oxygen termination
termination by ozone
3 Polishing/acid cleaning/ N2 + 10 s/He+ 10 s 2 × 1013 /4.1 × 1012 Oxygen termination
hydrogen termination by ozone
4 Polishing/acid cleaning/ N2 + 10 s/He+ 10 s 2 × 1013 /4.1 × 1012 Oxygen termination
RIE-ICP etching/hydrogen by ozone
termination
5 As-grown/hydrogen N2 + 10 s/He+ 0 2 × 1013 /0–4.1 × 1012 – Oxygen termination
termination s–10 s–20 s–30 s 8.2 × 1012 –1.2 × 1013 by O2 plasma

3
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

Figure 2. Nitrogen–helium co-implantation for NV centres formation performed on sample No. 5.

Figure 3. Schematic of NV centres fabrication process in diamond. (a) As grown CVD diamond layer. (b) Hydrogen surface
termination. (c) Nitrogen and (d) helium ion implantation. (e) Annealing in high vacuum at 800 ◦ C for 2 h. (f) CVD diamond
containing shallow NV centres layer. (g) Oxygen termination after implantation.

Figure 4. PL spectra acquired with a 532 nm laser and normalized to the diamond Raman peak for the diamond samples
(1 and 2) that were oxygen terminated (red curve) and hydrogen terminated (blue curve) before nitrogen implantation. All
samples were equally O-terminated after the irradiation process.

4
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

Figure 5. (a) Confocal laser scanning microscopic images and (b) PL spectra performed with a 473 nm laser and normalized to
the diamond Raman peak after ion implantation of the as grown, polished and polished and etched samples (sample 1, 3 and 4,
respectively).

surface and a permanent magnet has been used to lift the degeneracy of the ms = ±1 spin sublevels and com-
pletely separate lines pertaining to different NV orientations (by Zeeman effect). Thus, ODMR spectra were
acquired by collecting the PL while sweeping the microwave frequencies: a PL intensity dip corresponds to a
resonance between the microwave field frequency and the transition between ms = 0 and one of the ms = ±1
states (i.e., electron spin resonance ESR).

3. Results and discussion

3.1. Effect of surface termination on NV creation

PL analysis under optical laser excitation at 532 nm has been performed after nitrogen implantation and
annealing on hydrogen and oxygen terminated CVD diamond samples (samples 1 and 2, respectively). We
underline here again that after implantation all samples were equally oxygen terminated and therefore the
effect of surface termination was only evaluated during the implantation process. Figure 4 shows the emission
resulting from NV0 and NV− centres at 575 nm and 637 nm respectively. Both spectra have been normalized
to the diamond Raman peak. Although NV centres have been successfully created whatever the surface termi-
nation before implantation, it can be noticed that for the hydrogen terminated sample the NV centres creation
is more efficient. This is probably due to the fact that hydrogen-terminated diamond surfaces induces p-type
surface transfer doping leading to a hole accumulation layer which could change the charge state of implanta-
tion defects and their recombination behaviour. In particular, the formation of thermally stable defects such
as di-vacancy (V2 ) or NVH could be strongly reduced as described by Fávaro de Oliveira et al [18].

3.2. Effect of polishing and RIE-ICP etching on NV creation

A critical issue for the formation of shallow NV centres in diamond is the surface roughness which has to be
optimized for an implantation to a depth of a few nm. Surface roughness of the diamond films following CVD

5
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

Figure 6. (a) Optical fluorescence images performed on sample 5 that has been submitted to different He+ irradiation doses.
(b) PL spectra normalized to the diamond Raman peak for different implanted areas.

growth can indeed reach several tens of nm depending on the thickness. For therefore, it may be necessary
to perform surface polishing. Figure 5 shows a comparison between the microscopic images and PL spectra
acquired with a 473 nm laser excitation of the as-grown (sample 1), as-polished and polished followed by
ICP etching samples (samples 3 and 4, respectively). The measurement by optical profilometry on an area of
100 × 100 μm2 showed that the roughness has been reduced by the polishing from 0.8 μm for the as grown
sample to 0.1 μm for the polished sample. It can be noticed that polishing has strongly reduced the NV cen-
tres creation efficiency. This is probably due to the fact that polishing induces sets of parallel polishing lines
(figure 5(a)) produced by the abrasive particles moving across the surface even though it may appear smooth
to the naked eye [19]. Moreover, it is well known that subsurface damage can be produced by the mechani-
cal polishing [20]. In order to suppress these polishing effects, an additional surface treatment consisting in
RIE-ICP etching has been performed. And as illustrated in figure 5(b), the PL spectrum of the polished and
etched sample indicates higher creation efficiency for NV centres. This is believed to be due to the removal of
subsurface damage induced by polishing and located in the first few micrometres at the surface that are etched
away by the plasma.

3.3. Effect of helium co-implantation

PL imaging has been performed with the DiamondViewTM for sample 5 that has received different helium
irradiation durations in different regions after nitrogen-helium co-implantation followed by thermal anneal-
ing. The result is presented in figure 6(a). Red luminescent regions are visible corresponding to the emission
of NV centres created by the implantation process. We note here that the overall brightness is slightly reduced
as compared to the previously implanted sample as reported in figure 4. This is likely due to the oxygen plasma
termination performed at the end of the process that possibly etched away part of the surface NVs. Further
optimization of this cleaning procedure should be performed. Nevertheless, the brightness difference between
different regions already indicates a higher creation efficiency in the co-implanted areas with 10 and 20 s of
treatment. Non-uniform NV creation is also visible within individual regions due to the ion distribution in
the beam or to inaccurate positioning of the physical mask during irradiation.
To confirm this tendency, PL spectra performed at 532 nm laser excitation is given in figure 6(b). The
presence of NV centres is clearly detected on the three parts of the diamond that have been implanted in a
second step using He+ beam and it can be confirmed that an optimum is found for an implantation duration
of 20 s. The region only implanted with nitrogen shows weak NV fluorescence. For too high an implantation
dose (30 s), the NV fluorescence reduces again, probably due to graphitization of the diamond film or to the
formation of defects that quench NV emission.
Figure 7(a) shows four typical ODMR spectra from the four different areas of the co-implanted sample.
For the area that has not been implanted with helium, no electron spin resonance (ESR) contrast could be
measured due to a too weak signal. For the other regions, the hyperfine interaction with 14N that has a 1/2 spin

6
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

Figure 7. (a) ODMR spectra recorded by monitoring the NV defect PL intensity for sample 5 and (b) ODMR linewidths as a
function of the counts rate for different areas of the implanted sample.

leads to the splitting of the ESR line into three components visible on the spectra. These measurements have
been made repeatedly on several positions at each of the four different regions of the diamond surface. The
ESR lines have been fitted with a Lorentzian and the extracted linewidth is plotted in figure 7(b). The extracted
linewidth is equal or better than values already reported in the litterature [21–23]. An almost linear relationship
between PL intensity and linewidth is observed for count rates larger than 1.1 MHz with a similar trend for
the regions implanted with 10 and 20 s of He+ . Indeed, NV creation is not fully uniform over the implanted
area and can vary by up to a factor of 7. For those areas with a moderate helium implantation time, NV density
is likely to be the main limiting factor and the reason for the enlargement of the ESR lines. On the contrary,
areas implanted for a longer time (30 s) always showed much weaker PL intensities but broad linewidth on
the order of 1.7 MHz. In this case, it is believed that the presence of other defects or strain induced by the
implantation is limiting ESR linewidth. Higher He+ implantation doses are therefore not suitable to obtain
bright intensity with narrow ESR linewidths. According to [24], the corresponding values of the coherence
time (T 2 ∗ ) are calculated using the approximation
 
Γ = 1/ πT ∗2

7
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

Table 2. Full width at half maximum (FWHM) and coherence

time (T 2 ∗ ) for the different implanted areas of sample 5.

Implanted area FWHM (MHz) T 2 ∗ (ns)

N2 + only Not enough signal Not enough signal

N2 + /10 s He+ 1.05 301

(three measurements) 1.17 271
1.51 210

N2 + /20 s He+ 1.06 300

(three measurements) 1.47 217
1.53 208

N2 + /30 s He+ 1.6 199

(two measurements) 1.9 173

where Γ represents the full width at half maximum. These values are from 173 to 301 ns as given in table 2
and. In the regions implanted with 10 and 20 s of He, we have almost the same T 2 ∗ values. While in the one
implanted at 30 s of He we have smaller T 2 ∗ values confirming the previously observed tendency.

4. Conclusion

Achieving shallow NVs with high concentrations and good spin properties in diamond is a real challenge, but
it is a pre-requisite to the development of sensitive quantum sensors. Low energy ion implantation in high-
quality CVD diamond films provides a way to achieve such near-surface NV ensembles but the optimization of
this process is crucial. In this work we investigated different surface treatments prior to ion implantation and
employed co-implantation with N2 + and He+ with carefully chosen energies to reach similar range in matter.
He+ ions are used to add vacancies, and to improve substitutional nitrogen to NV conversion. We successfully
created bright NV centre ensembles at about 30 nm from the CVD diamond surface and observed that this
process is more efficient when moderate He+ ion doses are used (4–8 × 1012 ions cm−2 ). ODMR linewidths
measured in continuous mode of the order of 1.5 MHz were achieved. In addition, we found that the initial
surface state of the diamond samples prior to implantation strongly influences the creation efficiency of NVs.
Hydrogen-terminated diamonds as well as surfaces without polishing-induced surface damage lead to a better
NV creation yield. This approach will guide efforts to the development of diamond NV sensors with better
sensitivity.

Acknowledgments

This work has received funding from the Diamond-NMR Project No. ANR-19-CE29-0017-04, the Ile-de-
France region in the framework of DIM SIRTEQ, and from the European Research Council (ERC) under
the European Union’s Horizon 2020 research and innovation programme (Grant Agreement No. 101019234,
RareDiamond). ANR (Agence Nationale de la Recherche) and CGI (Commissariat à l’Investissement d’Avenir)
are also gratefully acknowledged for their financial support through Labex SEAM (Science and Engineering
for Advanced Materials and devices): ANR-10-LABX-096 and ANR-18-IDEX-0001. Finally, we acknowledge
the C(PN)2 of University Sorbonne Paris Nord for giving access to RIE-ICP facility.

Data availability statement

All data that support the findings of this study are included within the article (and any supplementary files).

ORCID iDs

Midrel Wilfried Ngandeu Ngambou https://orcid.org/0000-0002-0261-9072

Ionut Balasa https://orcid.org/0000-0002-0147-9776
Fabien Bénédic https://orcid.org/0000-0003-3350-4021
Philippe Goldner https://orcid.org/0000-0001-8517-0911
Jocelyn Achard https://orcid.org/0000-0001-7000-7230

8
Mater. Quantum Technol. 2 (2022) 045001 M W Ngandeu Ngambou et al

References

[1] Gruber A, Dräbenstedt A, Tietz C, Fleury L, Wrachtrup J and von Borczyskowski C 1997 Scanning confocal optical microscopy
and magnetic resonance on single defect centers Science 276 2012–4
[2] Bar-Gill N, Pham L M, Jarmola A, Budker D and Walsworth R L 2013 Nat. Commun. 4 1743
[3] Gaebel T et al 2006 Nat. Phys. 2 408–13
[4] Li L et al 2015 Nat. Commun. 6 6173
[5] Achard J, Jacques V and Tallaire A 2020 J. Phys. D: Appl. Phys. 53 313001
[6] Suter D and Jelezko F 2017 Prog. Nucl. Magn. Reson. Spectrosc. 98–99 50–62
[7] Jaffe T et al 2020 Nano Lett. 20 3192–8
[8] Zhou T X, Stöhr R J and Yacoby A 2017 Appl. Phys. Lett. 111 163106
[9] Fávaro de Oliveira F, Momenzadeh S Ali, Antonov D, Scharpf J, Osterkamp C, Naydenov B, Jelezko F, Denisenko A and
Wrachtrup J 2016 Nano Lett. 16 2228–33
[10] Meijer J, Burchard B, Domhan M, Wittmann C, Gaebel T, Popa I, Jelezko F and Wrachtrup J 2005 Appl. Phys. Lett. 87 261909
[11] Morhange J F, Beserman R and Bourgoin J C 1975 Japan. J. Appl. Phys. 14 544–8
[12] Naydenov B et al 2010 Appl. Phys. Lett. 96 163108
[13] Khmelnitsky R A, Dravin V A, Tal A A, Zavedeev E V, Khomich A A, Khomich A V, Alekseev A A and Terentiev S A 2015 J. Mater.
Res. 30 1583–92
[14] Uzan-Saguy C, Cytermann C, Brener R, Ritcher M, Shaanan M and Kalish R 1995 Appl. Phys. Lett. 67 1194–6
[15] Valentin A, De Feudis M, Brinza O, Tardieu A, William L, Tallaire A and Achard J 2018 Phys. Status Solidi a 215 1800264
[16] Achard J, Silva F, Tallaire A, Bonnin X, Lombardi G, Hassouni K and Gicquel A 2007 J. Phys. D: Appl. Phys. 40 6175–88
[17] Ziegler J F, Ziegler M D and Biersack J P 2010 Nucl. Instrum. Methods Phys. Res. B 268 1818–23
[18] Fávaro de Oliveira F 2017 Nat. Commun. 8 15409
[19] Chen Y and Zhang L 2013 Polishing of Diamond Materials (Engineering Materials and Processes) (London: Springer)
[20] Achard J, Tallaire A, Mille V, Naamoun M, Brinza O, Boussadi A, William L and Gicquel A 2014 Phys. Status Solidi a 211 2264–7
[21] Osterkamp C, Balasubramanian P, Wolff G, Teraji T, Nesladek M and Jelezko F 2020 Adv. Quantum Tech. 3 2000074
[22] Tetienne J et al 2018 Phys. Rev. B 97 085402
[23] van Dam S B et al 2019 Phys. Rev. B 99 161203
[24] Barry J F, Schloss J M, Bauch E, Turner M J, Hart C A, Pham L M and Walsworth R L 2020 Rev. Mod. Phys. 92 015004