A modular optically pumped magnetometer system

Abstract

To address the demands in healthcare and industrial settings for spatially resolved magnetic imaging, we present a modular optically pumped magnetometer (OPM) system comprising a multi-sensor array of highly sensitive quantum magnetometers. This system is designed and built to facilitate fast prototyping and testing of new measurement schemes by enabling quick reconfiguration of the self-contained laser and sensor modules as well as allowing for the construction of various array layouts with a shared light source. The modularity of this system facilitates the development of methods for managing high-density arrays for magnetic imaging. The magnetometer sensitivity and bandwidth are first characterised in both individual channel and differential gradiometer configurations before testing in a real-world magnetoencephalography environment by measuring alpha rhythms from the brain of a human participant. We demonstrate the OPM system in a first-order axial gradiometer configuration with a magnetic field gradient sensitivity of 10fT/cm/Hz at a baseline of 4.5 cm. Single-channel operation achieved a sensitivity of 65fT/Hz . Bandwidths exceeding 200Hz were achieved for two independent modules. The system’s increased temporal resolution allows for the measurement of spinal cord signals, which we demonstrate by using phantom signal trials and comparing with an existing commercial sensor.

Full text

Quantum Sci. Technol. 9(2024) 035045 https://doi.org/10.1088/2058-9565/ad4e60
OPEN ACCESS
RECEIVED
12 February 2024
REVISED
2 May 2024
ACC EPT ED FOR PUB LICATI ON
21 May 2024
PUBLISHED
19 June 2024
Original Content from
this work may be used
under the terms of the
Creative Commons
Attribution 4.0 licence.
Any further distribution
of this work must
maintain attribution to
the author(s) and the title
of the work, journal
citation and DOI.
PAPER
A modular optically pumped magnetometer system
T Coussens1,2,∗, A Gialopsou1,3, C Abel1, M G Bason1,4, T M James1, W Evans2,
M T M Woodley1,2, D Nightingale1, D Nicolau1, L Page1, F Orucˇevi´
c1and P Krüger1,2
1Department of Physics and Astronomy, University of Sussex, Brighton BN1 9QH, United Kingdom
2Physikalisch-Technische Bundesanstalt (PTB), Abbestr. 2-12, 10587 Berlin, Germany
3Clinical Imaging Sciences Centre, Brighton and Sussex Medical School, University of Sussex, Brighton BN1 9RR, United Kingdom
4RAL Space, UKRI-STFC Rutherford Appleton Laboratory, Didcot OX11 0QX, United Kingdom
∗Author to whom any correspondence should be addressed.
E-mail: t.coussens@sussex.ac.uk
Keywords: modular, optically, pumped, magnetometer, array, OPM, MEG
Abstract
To address the demands in healthcare and industrial settings for spatially resolved magnetic
imaging, we present a modular optically pumped magnetometer (OPM) system comprising a
multi-sensor array of highly sensitive quantum magnetometers. This system is designed and built
to facilitate fast prototyping and testing of new measurement schemes by enabling quick
reconfiguration of the self-contained laser and sensor modules as well as allowing for the
construction of various array layouts with a shared light source. The modularity of this system
facilitates the development of methods for managing high-density arrays for magnetic imaging.
The magnetometer sensitivity and bandwidth are first characterised in both individual channel and
differential gradiometer configurations before testing in a real-world magnetoencephalography
environment by measuring alpha rhythms from the brain of a human participant. We demonstrate
the OPM system in a first-order axial gradiometer configuration with a magnetic field gradient
sensitivity of 10 fT /cm/√Hz at a baseline of 4.5 cm. Single-channel operation achieved a
sensitivity of 65 fT /√Hz. Bandwidths exceeding 200 Hz were achieved for two independent
modules. The system’s increased temporal resolution allows for the measurement of spinal cord
signals, which we demonstrate by using phantom signal trials and comparing with an existing
commercial sensor.
1. Introduction
Since the 1960s, superconducting quantum interference devices (SQUIDs) have been the benchmark for the
measurement of ultra-low magnetic fields [1,2]. However, developments in the spin-exchange
relaxation-free (SERF) regime over the past two decades have allowed optically pumped magnetometers
(OPMs) to achieve similar sensitivities to SQUIDs [3], with a number of distinct advantages. SQUIDs
require cryogenic cooling which increases costs, limits the portability of the device, and increases the
separation between the device and the magnetic source of interest. OPMs offer reduced separation distances
over SQUIDs, which is of particular interest for magnetoencephalography (MEG) [4–7], where OPM-MEG
is able to offer an improved spatial [8,9], and spatio-temporal [10] resolution.
The advantages of OPMs over SQUIDs also benefit many other applications such as the diagnostic
measurement of batteries through non-invasive current mapping [11–13], magnetic nanoparticle
detection [14,15], defect analysis in metals [16] and for fundamental physics research [17,18]. However, a
number of challenges remain, especially around the construction of dense arrays of OPMs. Commercially
available OPMs currently focus on single-sensor designs; when multiple units are operated in close proximity,
issues arise from radio frequency (RF) cross-talk [19], cross-axis projection errors [20], as well as from high
surface temperatures of the sensors. To achieve dense OPM array systems, these problems must be addressed.
In addition, in medical applications there is increasing interest in understanding high-frequency
biomagnetic phenomena in MEG [21], magnetospinography [22] and magnetomyography [23]. Existing
© 2024 The Author(s). Published by IOP Publishing Ltd

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
OPM-MEG is limited to frequencies up to approximately 100 Hz whereas SQUIDs are able to achieve far
larger bandwidths [24]. Further developments in OPMs are therefore needed in order to apply their
improved spatial resolution to the higher temporal frequency domain.
Here, we present a modular magnetometer system designed from the ground up for inherent scalability
while maintaining high performance, allowing for easy adaptability of sensor locations and of sensing
regimes. The layout of the magnetometer system is determined by assembling individual modules each with
specific functionality. Additionally, optical elements inside these modules may be replaced or reconfigured to
allow implementation of different sensing schemes.
To illustrate the adaptability of the magnetometer system, two configurations of the instrument, each
operating in the ultra-sensitive SERF regime, are presented. A single-sensor configuration was used to
capture alpha rhythms from the brain of a human participant, demonstrating suitability for high sensitivity
measurements. Additionally, a two-sensor gradiometer, with superior sensitivity performance through
common-mode noise rejection, was characterised. Furthermore, we demonstrate that the bandwidth of our
system gives it an improved ability to measure higher frequency signals generated by a spinal cord phantom
compared with a commercially available OPM.
2. Method
Our system utilises 45 mm ×45 mm ×40mm cuboidal modules, which interconnect to produce a desired
OPM system layout. An example of a 2×2 sensor array is shown in figure 1. The enclosure of each module is
3D-printed from polycarbonate, allowing for design flexibility whilst also providing excellent mechanical
rigidity with a 139 ◦C glass transition temperature [25]. The system is assembled from three types of
modules: light source, sensor, and beam distribution modules.
The light source modules introduce and condition laser light for use in the magnetometer system. Light
can be internally generated by a vertical-cavity surface-emitting laser, or routed from external sources via
optical fibres. The sensor modules contain optical elements, vapour cells, magnetic field coils and
photodiodes required for magnetic field measurements. Finally, beam distribution modules are used for light
delivery, allowing for greater flexibility in positioning the sensors. Due to the adaptability of its optics
packages, measurement schemes and array configurations, the magnetometer can be tailored to a large
variety of applications.
The sensor module design presented in this paper operates in the high-sensitivity SERF regime, which
requires large atomic densities in a low magnetic field environment [26]. A single-axis gradiometer (figure 2)
is constructed using one light source module and two sensor modules in a similar configuration to that in
[27], and can be extended to a multi-axis gradiometer instrument as shown in figure 1. The SERF sensor
modules utilise the Hanle resonance (zero-field resonance) as a measurement scheme [28], relying on the
monitoring of the transmitted power of a resonant circularly polarised laser beam probing a vapour cell.
Here, the Hanle resonance can be understood by considering the pumping laser to be aligned parallel to
the y-axis, as per the sensor’s coordinate system in figure 2, and all three components of the magnetic field to
be nulled to zero (Bx=By=Bz=0) within the vapour cell. Under these conditions, atoms are pumped into
a ‘dark state’. When an orthogonal component to the pumping beam, such as Bx, is swept away from zero
field, the atoms are no longer pumped into a ‘dark state’ and the magnitude of the transmitted light follows a
Lorentzian distribution with peak transmission at Bx=0. By applying a modulating field Bx,RF cos (ωt)
across the zero-field resonance peak, and passing the photodiode response through a lock-in amplifier (LIA),
a dispersion signal is produced, of the form [29,30]
V(Bx) = AγBxτ
1+ (γBxτ)2,(1)
where γis the gyromagnetic ratio for the active atomic species, τis the relaxation time of the spin-polarised
atomic population, and Ais a proportionality constant (with units V/pT) determined by calibrating the
sensor output. The LIA also reduces 1/f noise, improving the signal-to-noise ratio of the resonance
signal [31]. In the following measurements, the sensitive component is Bx, however a similar measurement
could be made for the other orthogonal component Bz.
Each sensor module contains three orthogonal coil pairs positioned around a vapour cell, which are used
to apply offset and modulation fields. Each coil pair is formed from two hand-wound 10mm ×10 mm
square coils offset from the centre of the vapour cell by 5mm. The field Bxis modulated at frequency
ω=2π×926Hz with a peak-to-peak of 60 nT at the centre of the vapour cell. To achieve sufficient atomic
density for the SERF regime, the vapour cells are heated using alternating currents at 81kHz and 200 kHz,
driven through twisted manganin heating wires wrapped around the vapour cells. These AC frequencies are
2

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
Figure 1. An example of a multi-axis magnetic gradiometer connecting 4 sensor modules in a 2 ×2 array configuration with light
source and beam distribution modules. The laser beam is conditioned with a lens, a linear polariser (LP) and a quarter-waveplate
(λ/4). After passing through vapour cells, the beam power is monitored with photodiodes (PD). A non-polarising 50:50 beam
splitter (50:50 BS) and a mirror ensure that the probe light from a single beam is delivered to all sensors.
Figure 2. A schematic (top) and a photograph (bottom) of the single-axis gradiometer modular OPM consisting of a light source
module and two sensor modules. Modules can be snapped together to produce different array configurations.
chosen to be orders of magnitude higher than the rubidium atomic response bandwidth to avoid broadening
of the measured resonance. Two heater frequencies were implemented to avoid cross-talk. The AC currents
driving the heaters were spaced sufficiently far apart so that the difference between the two frequencies was
well beyond the bandwidth of the sensors. The vapour cell housing is thermally insulated to minimise the
required heating power and the surface temperature of the sensor modules. In so doing, a stand-off distance
of 8.5mm from the end of the sensor module housing to the cell centre was achieved.
Light from an external-cavity diode laser, tuned to the 87Rb D1 transition (795nm), is coupled into a
polarisation-maintaining optical fibre and input to the light source module of the sensor (figure 2). The
linearly polarised beam is then collimated to a 1/e2radius of 1.5mm before being circularly polarised and
split with a 50:50 non-polarising beamsplitter cube. The reflected beam is directed through a
5 mm ×5mm ×5 mm cubic glass 87Rb cell in sensor module 1, containing 50Torr of N2and 650 Torr of Ne
as buffer gases [32]. In sensor module 2, an identical vapour cell is probed with the beam coming from the
transmitted arm of the beamsplitter. Photodiodes are placed behind the vapour cells to monitor the probing
beam power. Their signals are passed through external transimpedance amplifiers and fed to a pair of
synchronised LIAs before being recorded with a digital acquisition card (DAQ) [33]. Cell temperatures were
recorded prior to taking magnetic field measurements using a thermistor placed at the base of the vapour
cell. After the temperature was recorded, the thermistor was disconnected. Once inter-module fixings were
applied, the alignment reproducibility was sufficient for swapping out sensor modules without a noticeable
3

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
loss of performance. Approximately 360 µW of light power was used per channel. The losses in the optical
system were approximately 40% from the fibre output to the photodiode with an unheated vapour cell.
These losses may be improved by using vapour cells with anti-reflective coatings and mitigating Rb build-up
on the optical windows.
Sensitivity measurements were taken inside a four-layer mu-metal cylindrical shield [34] with an internal
magnetic field magnitude below 10 nT and magnetic noise of 10 fT/√Hz. The offset fields were applied
using the sensor module’s internal coils, and modulation fields were generated using field coils built into the
shield. The LIA output voltages were calibrated by applying well-defined Bxfields.
To demonstrate the suitability of the modular OPM system for MEG applications, a single-sensor
magnetometer configuration was operated to capture human brain alpha rhythms [35] which occur in the
8Hz–12 Hz band. The experiment was conducted in accordance with the Declaration of Helsinki Ethical
Principles, and was approved by the Brighton and Sussex Medical School Research Governance and Ethics
Committee (ER/BSMS3100/1). The participant gave written informed consent to take part after explanation
of the procedure and purpose of the experiment.
The MEG measurements were taken inside a three-layer mu-metal cylinder, with an internal diameter of
50 cm and axial length of 100 cm. The end caps on one side of the shield were removed for the duration of
the measurement to allow access for the participant. The sensor setup was similar to the three-module
gradiometer configuration shown in figure 2, with the difference that only one sensor module was present.
The single-sensor-module OPM was placed centrally at the back of the mu-metal cylinder (away from the
open end), supported by a wooden arm, against which the participant was able to rest their head, as shown in
figure 5. A prerecorded voice prompted the participant to open and close their eyes every 12 s to stimulate an
alpha-band response when their eyes were shut (photic blocking) [36]. The optimum placement of the
sensor over the primary visual cortex (Oz position) was measured according to the standard 10–10
system [37]. The sensor was operated sensitive to the Bxaxis (as shown in figure 2) which measured the
radial magnetic field component arising from the participant’s brain. The analogue signal from the
OPM-MEG system was sampled at 200Hz using the same DAQ setup as previously described.
Finally, the system’s temporal resolution enables the capture of higher frequency biomagnetic
phenomena. To demonstrate this, a phantom spinal cord signal was generated to form a comparison between
the modular OPM system and a commercial OPM. The phantom signal was created using
WebPlotDigitizer [38] based on signals recorded in [39] using a SQUID array with a bandwidth of
12kHz [40]. For the extracted signal, frequencies up to 650Hz contained approximately 90% of the phantom
signal’s power. The comparison was undertaken inside the four-layer mu-metal cylindrical shield with the
phantom signal created using the internal coil of the shield. Magnetic field measurements were acquired
sequentially from a single-channel modular OPM system and a commercial OPM [41]. The phantom spinal
cord signal was scaled to ensure the signal was well above the noise floor for both OPM systems, resulting in a
peak-to-peak of approximately 4nT. The amplitude of the phantom signal was chosen to be well within the
dynamic range of both sensors to enable a fair comparison. The two systems were placed in approximately
the same position, within the homogeneous region of the coils. Both systems were operated with a single
sensitive axis (single-axis mode for the commercial sensor) and were aligned parallel to the internal coil
generating the phantom signal. The analogue outputs of the commercial sensor and modular OPM LIA were
recorded using the aforementioned DAQ with a sample frequency of 40kHz. The recorded data were filtered
with the same notch filters to encompass the first three harmonics of the modulation frequency for the
modular system (926Hz) and the commercial system (923Hz).
3. Results and discussion
The sensitivity measurements were taken over a period of 60s during which the magnetometer signals from
sensor modules 1 and 2 were recorded. Data were analysed for each sensor separately, and as a software
gradiometer (sensor module 1 minus sensor module 2). The frequency spectra in figure 3show the noise
floors of sensor modules 1 and 2 to be 65fT /√Hz and 83 fT/√Hz, respectively, in the 5Hz–45Hz band.
When the difference is taken between the two channels to form a gradiometer, the sensitivity improves to
47fT /√Hz in the 5 Hz–45Hz band, corresponding to magnetic field gradient sensitivity of 10 fT /cm/√Hz
with a baseline of 4.5 cm. This improvement is attributed to the common mode reduction in light intensity
noise. As the response of the 87Rb atoms and the LIA low-pass filter both limit bandwidth, the linear density
spectrum has been divided by the transfer function of each sensor module [5]. By applying an oscillating
signal of constant amplitude using the internal coils (approx 200 pT) with varying frequency, the transfer
function of each sensor module was measured up to 320Hz. Each transfer function follows a first-order
low-pass filter and the bandwidths of the sensor modules were found to be (213 ±5)Hz for module 1, and
4

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
Figure 3. The 1 Hz–200Hz frequency-compensated linear spectral density of the demodulated signals from sensor module 1
(blue), sensor module 2 (red) and from the balanced photodiodes gradiometer configuration (yellow). The gradiometer
configuration removes common-mode noise between the two sensors, which can be seen by the noise reduction at 50 Hz, and in
the broader peak in the 55Hz–75 Hz band. Sensor module 1 has a noise-floor of 65 fT/√Hz, sensor module 2, 83 fT/√Hz, and
the gradiometer configuration 47 fT /√Hz in the 5Hz–45 Hz band.
Figure 4. The magnetometer response amplitudes for sensor modules 1 (blue) and 2 (red) as a function of applied signal
frequency. The data points are fitted to the response of a single order low pass filter. Dashed lines indicate a −3dB bandwidth of
(213 ±5)Hz for module 1, and (219 ±4)Hz for module 2.
(219 ±4)Hz for module 2, as shown in figure 4. The response from module 1 was found to deviate by less
than 10% from linearity for offsets >−7.4nT and <6.1 nT.
For this measurement, the 87Rb vapour cells were operated at temperatures of 129◦C and 124◦C in
sensor modules 1 and 2, respectively. We note that the optimal sensitivity of 87 Rb vapour in the SERF regime
occurs at higher temperatures [42]. For example, an increase of the cell temperature from 124 ◦C to 150 ◦C
corresponds to a factor-4 increase in vapour density [43]. Implementing this would, therefore, require an
alternative to the printed polycarbonate material. Considering the chosen conditions for this experiment,
light intensity noise (greater than the shot noise limit) was established to be the main limitation in sensitivity.
By improving the intensity noise to the shot noise limit, the performance of single-channel operation would
improve to approximately 30fT /√Hz. Further gains could be made in the low-frequency domain by
implementing active feedback on vapour cell temperature and laser power.
Data from the alpha rhythm measurement with a single sensor module were analysed in the frequency
domain, with the spectrogram shown in figure 5. We see an increase in the activity in the 8 Hz–12Hz
alpha-band when the participant’s eyes are closed, with a strong correlation with the TTL cue. Although we
5

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
Figure 5. Spectrogram of the 8 Hz–12Hz alpha-band response to a participant opening/closing their eyes (a) and a schematic of
the participant lying in the shield with the back of their head pushed up against the modular OPM system (b). The overlaid white
trace denotes the audio cue trigger signal from the stimulus PC. High values of the trigger signal indicate the time when the
participant is instructed to close their eyes, and the low value is when the participant has their eyes open. The spectrogram colour
scheme details the high periods of activity in yellow, and the low periods in green. A peak in activity of 1 pT /√Hz is observed in
the 8Hz–12 Hz region during the third eyes-closed time period.
already show that the current modular sensor is suitable for MEG, future improvements of the
signal-to-noise ratio could be made by suppressing common-mode noise using a gradiometer.
High-frequency measurement capability is important for resolving rapidly varying biomagnetic
phenomena [21–23]. A larger bandwidth was observed for the modular OPM system (approximately
210Hz–220 Hz) compared with that of the commercial OPM (approximately 135Hz [44]). However,
increased bandwidth comes at a cost of reduced sensitivity. Fundamentally, considering the spin projection
noise of the atoms, improving the sensitivity, i.e. lowering the minimum detectable magnetic field signal by
increasing the relaxation time τ(δB∼τ−1/2[26]), entails a reduced sensor bandwidth (∼τ−1[45]).
However, in practice, other technical noise contributions may interfere with this relationship [46]. The
relaxation time can be modified by tuning multiple parameters, including the rates of optical pumping, spin
destruction collisions and collisions with the vapour cell wall, as well as the presence of magnetic field
gradients [47]. It is also possible to shift the sensitive bandwidth range to higher frequencies by applying DC
field offsets [48].
In this case we present the advantages of increased bandwidth on the measurement of spinal cord activity
with a phantom signal. The extracted spinal cord signal (as described in the Method section) is used to drive
the shield’s internal coil. The responses to this signal as recorded by the modular and commercial OPM
systems are presented in figure 6. Time delays were optimised using root mean square deviation (RMSD)
minimisation with respect to the raw phantom trace and signal amplitudes normalised to the peak at 20ms.
Time delays of 0.7ms for the modular OPM and 3.7 ms for the commercial OPM were observed. We believe
this delay is characteristic of each system’s signal processing chain. The modular OPM system provides a
more accurate replication of the signal trace, with a RMSD of 0.05 compared to the commercial sensor’s
RMSD of 0.10. The features of the high-frequency phantom signal around 12ms are not well resolved by the
commercial OPM. We attribute this to the lower bandwidth of the commercial system rather than
6

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
Figure 6. A comparison in the responses of the modular (a) and commercial (b) OPM systems to the phantom spinal cord signal.
Time delays were optimised using RMSD minimisation with respect to the raw phantom trace. The uncompensated raw data are
shown as faded traces.
non-linearity in its dynamic range as even small-amplitude high-frequency oscillations in the phantom
signal are more accurately reproduced by the modular OPM than the commercial OPM (see for example the
signal from 7ms–10 ms in figure 6).
We have designed a modular OPM system and demonstrated its operation in single- and dual-channel
configurations. We have also utilised the system to record MEG signals from a human participant and
demonstrated an improved bandwidth over a commercial OPM system in resolving high-frequency features
from a phantom spinal cord signal. The ability of this highly customisable OPM system to implement new
measurement schemes, array configurations, dual-scheme OPM arrays, or multi-modal quantum sensor
systems, offers a unique combination of adaptability and capability for in-field use. This modular system is
an important development tool for the advancement of dense OPM arrays in medical and industrial fields,
providing a rapid testing platform to benchmark OPM system architectures.
Data availability statement
The data that support the findings of this study will be openly available following an embargo at the
following URL/DOI: https://doi.org/10.25377/sussex.c.7010940.
Acknowledgment
This work was supported by the UK Quantum Technologies Hub for Sensors and Timing (EPSRC Grant
EP/T001046/1), University of Sussex Strategic Development Fund and Innovate UK: Batteries—ISCF 42186
Quantum sensors for end-of-line battery testing.
CRediT author statement
T. Coussens: Conceptualisation, Methodology, Investigation, Visualisation, Writing—original draft,
Writing—review & editing. A. Gialopsou : Investigation, Visualisation, Writing—review & editing. C. Abel:
Methodology, Writing—review & editing. M. Bason: Methodology, Writing—review & editing. T. James:
Visualisation, Writing—review & editing. W. Evans: Writing—review & editing. M. Woodley:
Writing—review & editing. D. Nightingale: Writing—review & editing. D. Nicolau: Writing—review &
editing. L. Page: Writing—review & editing. F. Orucˇevi´
c: Writing—review & editing, Supervision, Funding
acquisition. P. Krüger: Writing—review & editing, Supervision, Funding acquisition.
7

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
ORCID iDs
T Coussens https://orcid.org/0000-0002-2106-9125
A Gialopsou https://orcid.org/0000-0002-6236-6038
C Abel https://orcid.org/0000-0002-6707-0586
M G Bason https://orcid.org/0000-0003-1921-524X
W Evans https://orcid.org/0000-0001-8323-0722
M T M Woodley https://orcid.org/0000-0003-4581-1643
D Nightingale https://orcid.org/0009-0000-9659-4728
D Nicolau https://orcid.org/0009-0008-5170-2332
L Page https://orcid.org/0009-0000-0113-1940
F Orucˇevi´
chttps://orcid.org/0000-0002-6114-4580
P Krüger https://orcid.org/0000-0001-9908-6454
References
[1] Ryhänen T, Seppä H, Ilmoniemi R and Knuutila J 1989 SQUID magnetometers for low-frequency applications J. Low Temp. Phys.
76 287–386
[2] Pizzella V, Penna S D, Gratta C D and Romani G L 2001 SQUID systems for biomagnetic imaging Supercond. Sci. Technol. 14 R79
[3] Kominis I K, Kornack T W, Allred J C and Romalis M V 2003 A subfemtotesla multichannel atomic magnetometer Nature 422 596
[4] Sander T H, Preusser J, Mhaskar R, Kitching J, Trahms L and Knappe S 2012 Magnetoencephalography with a chip-scale atomic
magnetometer Biomed. Opt. Express 3981–90
[5] Colombo A P, Carter T R, Borna A, Jau Y-Y, Johnson C N, Dagel A L and Schwindt P D D 2016 Four-channel optically pumped
atomic magnetometer for magnetoencephalography Opt. Express 24 15403–16
[6] Sheng J, Wan S, Sun Y, Dou R, Guo Y, Wei K, He K, Qin J and Gao J-H 2017 Magnetoencephalography with a Cs-based
high-sensitivity compact atomic magnetometer Rev. Sci. Instrum. 88 094304
[7] Limes M, Foley E, Kornack T, Caliga S, McBride S, Braun A, Lee W, Lucivero V and Romalis M 2020 Portable magnetometry for
detection of biomagnetism in ambient environments Phys. Rev. Appl. 14 011002
[8] Tan S, Roth B J and Wikswo Jr J P 1990 The magnetic field of cortical current sources: the application of a spatial filtering model to
the forward and inverse problems Electroencephalogr. Clin. Neurophysiol. 76 73–85
[9] Boto E, Bowtell R, Krüger P, Fromhold T M, Morris P G, Meyer S S, Barnes G R and Brookes M J 2016 On the potential of a new
generation of magnetometers for MEG: a beamformer simulation study PLoS One 11 e0157655
[10] Gialopsou A et al 2021 Improved spatio-temporal measurements of visually evoked fields using optically-pumped magnetometers
Sci. Rep. 11 22412
[11] Bason M G, Coussens T, Withers M, Abel C, Kendall G and Krüger P 2022 Non-invasive current density imaging of lithium-ion
batteries J. Power Sources 533 231312
[12] Hu Y, Iwata G Z, Mohammadi M, Silletta E V, Wickenbrock A, Blanchard J W, Budker D and Jerschow A 2020 Sensitive
magnetometry reveals inhomogeneities in charge storage and weak transient internal currents in Li-Ion cells Proc. Natl Acad. Sci.
USA 117 10667–72
[13] Peng T, Gao W, Wu Y, Ma Y, Zhang S and Hu Y 2023 A real-time non-contact localization method for faulty electric energy storage
components using highly sensitive magnetometers (arXiv:2308.07708 [physics])
[14] Johnson C, Adolphi N L, Butler K L, Lovato D M, Larson R, Schwindt P D D and Flynn E R 2012 Magnetic relaxometry with an
atomic magnetometer and SQUID sensors on targeted cancer cells J. Magn. Magn. Mater. 324 2613–9
[15] Everaert K, Sander T, Körber R, Löwa N, Waeyenberge B V, Leliaert J and Wiekhorst F 2023 Monitoring magnetic nanoparticle
clustering and immobilization with thermal noise magnetometry using optically pumped magnetometers Nanoscale Adv. 52341–51
[16] Maddox B, Cohen Y and Renzoni F 2022 Through-skin pilot-hole detection and localization with a mechanically translatable
atomic magnetometer Appl. Phys. Lett. 120 014002
[17] Pustelny S et al 2013 The global network of optical magnetometers for exotic physics (GNOME): a novel scheme to search for
physics beyond the standard model Ann. Phys. 525 659–70
[18] Abel C et al 2019 PicoTesla absolute field readings with a hybrid 3He/87Rb magnetometer Eur. Phys. J. D73 150
[19] Nardelli N V, Krzyzewski S P and Knappe S A 2019 Reducing crosstalk in optically-pumped magnetometer arrays Phys. Med. Biol.
64 21NT03
[20] Borna A, Iivanainen J, Carter T R, McKay J, Taulu S, Stephen J and Schwindt P D D 2022 Cross-axis projection error in optically
pumped magnetometers and its implication for magnetoencephalography systems NeuroImage 247 118818
[21] Iivanainen J, Zetter R and Parkkonen L 2020 Potential of on-scalp MEG: robust detection of human visual gamma-band responses
Hum. Brain Mapp. 41 150–61
[22] Chander B S, Deliano M, Aza˜
nón E, Büntjen L and Stenner M-P 2022 Non-invasive recording of high-frequency signals from the
human spinal cord NeuroImage 253 119050
[23] Marquetand J et al 2021 Optically pumped magnetometers reveal fasciculations non-invasively Clin. Neurophysiol. 132 2681–4
[24] Drung D, Hinnrichs C and Barthelmess H-J 2006 Low-noise ultra-high-speed dc SQUID readout electronics Supercond. Sci.
Technol. 19 S235
[25] Shamim N, Koh Y P, Simon S L and McKenna G B 2014 Glass transition temperature of thin polycarbonate films measured by flash
differential scanning calorimetry J. Polym. Sci. B52 1462–8
[26] Allred J C, Lyman R N, Kornack T W and Romalis M V 2002 High-sensitivity atomic magnetometer unaffected by spin-exchange
relaxation Phys. Rev. Lett. 89 130801
[27] Nardelli N V, Perry A R, Krzyzewski S P and Knappe S A 2020 A conformal array of microfabricated optically-pumped first-order
gradiometers for magnetoencephalography EPJ Quantum Technol. 71–11
[28] Dupont-Roc J, Haroche S and Cohen-Tannoudji C 1969 Detection of very weak magnetic fields (10−9gauss) by 87Rb zero-field
level crossing resonances Phys. Lett. A28 638–9
8

Quantum Sci. Technol. 9(2024) 035045 T Coussens et al
[29] Cohen-Tannoudji C, Dupont-Roc J, Haroche S and Lalo¨
e F 1970 Diverses résonances de croisement de niveaux sur des atomes
pompés optiquement en champ nul. I. Théorie Rev. Phys. Appl. 595–101
[30] Tierney T M et al 2019 Optically pumped magnetometers: from quantum origins to multi-channel magnetoencephalography
NeuroImage 199 598–608
[31] Kastler A 1973 The Hanle effect and its use for the measurements of very small magnetic fields Nucl. Instrum. Methods 110 259–65
[32] 5 mm ×5 mm ×5 mm ×cubic glass 87 Rb cells containing 50 Torr of N2and 650 Torr of Ne were procured from Twinleaf LLC,
USA.
[33] LabJack T7 pro, LabJack Corporation, USA.
[34] MS-2 mu-metal shield, Twinleaf LLC, USA.
[35] Cohen D 1968 Magnetoencephalography: evidence of magnetic fields produced by alpha-rhythm currents Science 161 784–6
[36] Cohen D 1972 Magnetoencephalography: detection of the brain’s electrical activity with a superconducting magnetometer Science
175 664–6
[37] Koessler L, Maillard L, Benhadid A, Vignal J P, Felblinger J, Vespignani H and Braun M 2009 Automated cortical projection of EEG
sensors: anatomical correlation via the international 10–10 system Neuroimage 46 64–72
[38] Rohatgi A 2022 Webplotdigitizer: version 4.6 (available at: http://automeris.io/WebPlotDigitizer/)
[39] Miyano Y et al 2020 Visualization of electrical activity in the cervical spinal cord and nerve roots after ulnar nerve stimulation
using magnetospinography Clin. Neurophysiol. 131 2460–8
[40] Adachi Y, Kawai J, Haruta Y, Miyamoto M, Kawabata S, Sekihara K and Uehara G 2017 Recent advancements in the SQUID
magnetospinogram system Supercond. Sci. Technol. 30 063001
[41] QZFM Gen. 1, QuSpin Inc., USA.
[42] Shah V, Knappe S, Schwindt P D D and Kitching J 2007 Subpicotesla atomic magnetometry with a microfabricated vapour cell Nat.
Photon. 1649
[43] Seltzer S J 2008 Developments in alkali-metal atomic magnetometry PhD Thesis Princeton University
[44] Osborne J, Orton J, Alem O and Shah V 2018 Fully integrated standalone zero field optically pumped magnetometer for
biomagnetism Proc. SPIE 10548 105481G
[45] Shah V K and Wakai R T 2013 A compact, high performance atomic magnetometer for biomedical applications Phys. Med. Biol.
58 8153–61
[46] Fabricant A, Novikova I and Bison G 2023 How to build a magnetometer with thermal atomic vapor: a tutorial New J. Phys.
25 025001
[47] Shah V and Romalis M V 2009 Spin-exchange relaxation-free magnetometry using elliptically polarized light Phys. Rev. A
80 013416
[48] Savukov I, Kim Y J, Shah V and Boshier M G 2017 High-sensitivity operation of single-beam optically pumped magnetometer in a
kHz frequency range Meas. Sci. Technol. 28 035104
9