2.1

Fluxgate magnetometers

The fluxgate magnetometer is the most common type of magnetic sensor used in space. It relies upon gating the ambient magnetic field by saturating a high permeability core.³ Through Lenz’s law, an output coil picks up this changing flux, and an output voltage is produced. Gating a core at a frequency f, produces an AC voltage at 2f. The signal amplitude provides information on the magnitude and direction of the ambient magnetic field. Many fluxgates have been successfully used in space thanks to their vector nature, low cost, and low power requirements.

2.2

Anisotropic Magnetoresistive (AMR) sensors

AMR sensors are based on elements whose resistance decreases when a magnetic field is applied.⁴ The sensor response is dependent on the direction of the magnetic field lines due to their anisotropic response. These sensors commonly use permalloy, which is magnetised in a certain direction, known as the easy direction. Upon application of a magnetic field, this magnetisation rotates towards the direction of the magnetic field with an angle-dependent on the external field strength. The changes in resistance are converted to a voltage using a two-legged current path which feeds two terminals on an operational amplifier. Operating in a closed-loop, fields of around 0.1 nT can be detected. These sensors are light, small, require between 0.1 and 0.5mW of power, and can be operated at temperatures between –55 °C and 200 °C.

2.3

Proton precession/Overhauser magnetometers

The proton precession magnetometer is a scalar sensor that uses nuclear magnetic resonance to detect magnetic fields. The underlying principle is that many atoms possess a net magnetic moment and thus spin, or precess, in magnetic fields. The starting point of such a sensor is to align the atoms such that the magnetic moments point in the same direction, i.e. to polarise a sample. After such an alignment, the protons then relax in the ambient magnetic field. This relaxation takes the form of a change in the orientation of the magnetic moment, which rotates about the magnetic field. As it does so, an AC voltage is generated in a coil whose frequency is proportional to the magnitude of the field.⁵

The basic proton precession magnetometer uses a large volume, liquid sample rich in protons. The polarising fields are also high, 100 G or more. This, combined with the fact that the detectable ambient fields should be greater than 20 μT, has limited the utility of such sensors in space.

The Overhauser magnetometer improves on the proton precession magnetometer by polarising the sample more efficiently. A source of free radical electrons are added to the liquid sample and then pumped with radiofrequency radiation. Electron-proton coupling then acts to polarise the protons. This method is around ten times more efficient than the DC method and yields 100 times larger signals.⁶

2.4

Helium magnetometers

These magnetometers use a glass cell filled with He-4 – an inert, noble gas that is extremely stable. This gas is manipulated into a metastable state by an RF discharge. From this state, it is optically pumped using light at 1083 nm. Field-dependent magnetic resonances are then optically detected using an infrared photodetector. Scalar read-out can be achieved using magnetically-driven spin precession that causes transitions between three internal states of the atom.⁷ The magnetometer can also be configured as a vector sensor using either extra magnetic ‘bias’ fields to null the read-out or using modulation/demodulation techniques. Two types of optical pumping techniques are used for He magnetometers. Early versions used RF discharge lamps at 1083 nm, while later versions replaced this with pumping using semiconductor and fibre lasers. The latter, in particular, gives a high sensitivity below .

Advantages of He-4 magnetometers include small heading errors and no ‘dead zones’ (angles at which the sensitivity of measuring the ambient magnetic field are drastically reduced). The drawback of such technology is in SWaP. A small He-4 magnetometer that uses less than 1 W of power is challenging because of the RF discharge and the lack of low-power (<10mW) excitation lasers.

2.5

Optically-pumped magnetometers (OPMs) using alkali atoms

Optically-pumped magnetometers use light beams to redistribute atoms into magnetically sensitive states. Once in these states, atoms then precess around the ambient magnetic field. This precession is read-out by modulating another light field’s amplitude or plane of polarisation. Photodiodes are used to convert this probing light into an electrical signal. As the precession frequency, via the gyromagnetic ratio, is proportional to fundamental physical constants, these are absolute magnetometers. In contrast to proton precession magnetometers, OPMs exploit atoms (typically caesium, rubidium, or potassium) in the vapour phase. To increase the number of atoms that contribute to the magnetometer signal, the vapour is held in a glass cell and heated to 50 °C to 180 °C.

Many modern OPMs use miniature diode lasers, such as vertical-cavity surface-emitting lasers (VCSELs), to provide both the optical pumping and probing light. Such developments have significantly reduced the SWaP characteristics. The intended range of the magnetometer must be considered when designing the device. In sensors that use rubidium, the non-linear Zeeman effect presents problems with fields of over 10 μT. Schemes to overcome this difficulty using extra light beams have been demonstrated.⁸

2.6

Nitrogen-vacancy centres in diamond

Negatively-charged nitrogen-vacancy (NV) centres are point defects in a diamond lattice. The NV centre consists of a nitrogen atom adjacent to a carbon vacancy. The negatively-charged NV centre can be spin polarised using one wavelength of light, like in other optical magnetometers, and then read out at a different wavelength.⁹ The sensing element of NV centres are typically small and single defects can be engineered in nano-diamonds for high-resolution magnetometry.¹⁰ Researchers have been able to demonstrate field sensitivities of around . The sensor detects both DC and AC vector magnetic fields using different techniques - typically up to several GHz with bandwidths of up to roughly 100 kHz.

Table 3.

Magnetometer instruments on a range of missions. A ✓ indicates that the sensor was used on the spacecraft.

2.7

Historic magnetometer choices

In 3, we identify the magnetometer technologies used in space missions from the 1950s through to the present day. This table highlights the particularly wide-use of fluxgates, often in combination with Helium sensors. In addition, it indicates the range of technologies now available for space missions.

3.1

OPM space heritage

Laser-pumped alkali atom magnetometers are beginning to attract the attention of the space community again. One example is the coupled dark-state magnetometer (CDSM). This is a scalar magnetometer, designed for the China Seismo-Electromagnetic Satellite, which was launched in 2018. It was designed by Space Research Institute (IWF) of the Austrian Academy of Sciences. The CDSM uses CPT to reduce the sensitivity of the sensor to temperature. Unlike many other scalar OPMs, the detection scheme it uses is omnidirectional, i.e. has no dead zones. It is an all-optical sensor design without excitation coils or electromechanical parts. The instrument has a high accuracy of 0.19 nT, but a relatively large detection noise of at 1 s integration time. It has a mass of 1.7 kg and power consumption of 3.4W.¹²

Researchers at Johns Hopkins University and the National Institute of Standards and Technology (NIST) in the US have also developed a OPM for space applications.¹³ The focus of their work is not geared towards a particular space mission, rather in developing and qualifying the sensor. The collaboration used the rubidium isotope ⁸⁷Rb held in a vapour cell with a volume of only 1 mm³ to enable efficient heating using a resistive heater implemented in multiple metal layers of a transparent sapphire substrate. The prototype instrument has a total mass of less than 0.5 kg and uses less than 1 W of power while achieving a sensitivity of at 1 Hz, comparable to other OPMs. Their estimate of the combined variability due to solenoid field variations and instrument RMS noise amounted to about 0.1 nT. This is believed to be due to the power supply stability, which could easily be improved in future sensors.

3.2

All-optical sensors

In comparison to the electrical currents needed in many alternative sensing technologies, it is possible to produce OPMs that rely only upon the transmission and detection of light. One potential all-optical configuration is shown in Figure 1. In this approach the pumping and probing laser light is transferred to the cell and collected by optical fibres. Laser light, in combination with absorptive filters on the vapour cell faces, is also used to increase and stabilise the cell temperature.¹⁴

Figure 1.

Example configuration of an all-optical OPM sensor for deployment in space.

An all-optical approach has two distinct advantages. The first is that the OPM can be passive, i.e. not produce magnetic fields itself, that would otherwise perturb other sensors. Secondly, the optical components that are used to guide light are significantly lighter than copper cabling. Reducing mass is particularly attractive to missions that place sensors at the end of a long boom. For example, a 125 μm diameter optical fibre used to connect a sensor at the end of a 3 m boom weighs around 400 g. A 32 AWG, 125 μm copper wire has a roughly 8.5 × larger mass. Thus, moving to optical sensors could offer significant savings in boom design.

4.1

Scalar OPMs to complement vector fluxgate measurements

Scalar magnetometers measure the Zeeman resonance frequency proportional to the absolute value of the magnetic field. They can operate in a range of fields, including in the Earth’s magnetic field. These OPMs offer very high sensitivities of less than .¹⁵ Such performance is achieved by developing mitigations for detrimental atomic collisions, such as spin-exchange collisions, which are present in hot alkali-metal vapour magnetometers operating in a finite magnetic field. A collaboration between the Romalis group in Princeton and Twinleaf inc. demonstrated a finite field gradiometer using an intense pulsed laser to polarise a ⁸⁷Rb atomic ensemble and a compact VCSEL probe laser to detect paramagnetic Faraday rotation in a single multipass cell. They reported differential magnetic sensitivity of over a broad dynamic range including Earth’s field magnitude and common-mode rejection ratio higher than 10⁴.¹⁶

This application area is based on the fact that certain OPM schemes have the advantage of intrinsic calibration provided by fundamental physical constants. These constants govern the atomic vapour response to magnetic fields. As a result, they are able to provide exceptional accuracy and sensitivity with limited drift. The compact scalar sensors considered here are therefore well suited to act as an onboard calibration source for a more conventional vector fluxgate magnetometer. This hybrid sensor approach has been flown previously (as a fluxgate and helium magnetometer combination), for example on the Cassini mission, and has been shown to provide better accuracy that either instrument alone.

A simplified scalar magnetometer scheme has been developed with a specific focus on space applications. The prototype instrument achieved a sensitivity of in a 10 μT background field, with a total mass below 500 g and a power consumption below 1W.¹³ The above approach uses an RF coil to drive the atomic response, typically in the M_x configuration.¹⁷ This has two potential drawbacks for space applications. The first is the addition of an electrical connection to the sensor to provide the RF coil drive. The second is that the M_x approach has two ‘dead-zones’, one perpendicular to (polar) and one parallel to (equatorial) the direction of propagation of the laser beam. If the measurement field approaches these regions, the measurement sensitivity drops. An alternative approach is to instead drive the atomic response by either frequency-modulation (Bell-Bloom magnetometer, or FM NMOR: frequency-modulated nonlinear magneto-optical rotation) or amplitude modulation (AM NMOR: amplitude-modulated nonlinear magneto-optical rotation) of the laser beam. A comparison of this approach with that of the M_x scheme has shown that the two have comparable sensitivities but that the NMOR scheme has the added simplicity of only requiring optical connections to the sensor unit and the removal of the equatorial dead-zone.¹⁸

Magnetometers based on this approach have been shown to achieve sub-picotesla sensitivities in Earth scale background fields (10 μT). During operation, the modulation frequency tracks changes in the magnetic field. This tracking has been demonstrated across the 0 nT to 40 000 nT range. As a result, such schemes are ideal candidates for planetary missions – where large field ranges are encountered – in addition to being onboard calibration sources for alternative sensors.¹⁹ Finally, a simple extension to the sensor unit to include two non-overlapping orthogonal beams would allow continuous dead-zone-free measurements to be performed.

4.2

Vector read-out with low SWaP

Two vector components were measured by the Budker group in 2014.²⁰ Vector capability is achieved by effective modulation of the field along orthogonal axes and subsequent demodulation of the magnetic-resonance frequency. This modulation is provided by the AC Stark shift induced by circularly polarised laser beams. The sensor had a noise floor of and and was limited by power and frequency noise in the laser beams. Extending this to a full vector magnetometer – i.e. measuring all three components of the magnetic field was achieved by the neutron electric dipole moment (nEDM) collaboration.²¹ This magnetometer uses four laser beams and is operated in a pulsed mode, with each experimental cycle repeating every 40 ms. The sensor was designed for long-time stability and achieves a scalar resolution better than 300 fT for integration times ranging from 80 ms to 1000 s. The magnetic field direction was measured with a resolution better than 10 μrad for integration times from 10 s up to 2000 s. In contrast to other vector magnetometers the scalar resolution is not degraded by extracting vector information.

In addition to these full vector techniques, scalar magnetometers can be adapted to yield 2D vector information by measuring both the DC probe light transmission and the AC signal at the Larmor frequency. Combining two sensors gives full vector read-out. While the angular sensitivity is better than 0.02° with a measurement time of 100 ms – systematic errors of around 1 degree currently prohibit greater uptake of this method.²² Vector information can also be extracted using OPM schemes with elliptically polarised light. However, these have relied upon the ability to change the direction of the bias field and dead zones were present.^{23, 24} Typically, they would be of limited use in space missions.

Making use of the two naturally occurring isotopes of Rb, researchers have implemented all-optical vector magnetometers with two orthogonal optical pumping beams. Amplitude modulations occurred at ⁸⁵Rb and ⁸⁷Rb Larmor frequencies, respectively. Simultaneously detection of the magnetic field in each direction was extracted using a single probe beam in the third direction. In a magnetic field ranging from 10 μT to 50 μT, a field angle sensitivity of better than above 10 Hz was shown.²⁵

4.3

OPMs for low-field environments

A significant subset of OPMs are those designed to work at very low fields - zero-field OPMs typically operate in fields around 10 nT. At such fields, the rate of detrimental collisions of atoms in the vapour exceeds the rate at which they are precessing. This effectively averages out a large source of sample decoherence which would otherwise have detrimental effects on the sensitivity of the magnetometer. This regime is known as the Spin-Exchange Relaxation Free (SERF) regime. To operate in the SERF regime, the atomic density needs to be higher than in other classes of OPM. Increasing the temperature of the alkali vapour is the main method employed to reach this density.

World-record magnetic field sensitivities have been achieved in the zero-field/SERF regime: with a measurement volume of 0.3 cm³. Despite this sensitivity, there is room for improvement, with a theoretical fundamental sensitivity limit of below .²⁶ Similar OPMs have also been operated as gradiometers. One such sensor from NIST has a baseline of 20 mm and is interrogated by the same laser beam resulting in a noise floor of above 20 Hz. The maximum rejection of magnetic field noise is 1000 at 10 Hz.²⁷

While the operational range of zero-filed OPMs is limited, there are a number of ways to extend this. Nulling coils can be used to compensate for external fields so that the field inside the OPM remains close to zero. This has been shown to work from 0 nT to 60 000 nT.²⁸ The range can also be extended by operating just outside the SERF regime, i.e. at lower vapour cell lower temperatures. This comes at the cost of reduced sensitivity via increased spin relaxation.

Full vector sensitivity has been demonstrated using a scheme in which an incident laser beam is reflected at 90° through a vapour cell and three orthogonal modulation fields are applied.²⁹ Such a magnetometer has magnetic-field sensitivities of along two directions and along the other axis.

An important consideration when using this OPM scheme is that, in comparison to other OPMs, they need careful calibration. In practice, this calibration can be done before the mission. Re-calibration is unlikely to be needed provided that there are not significant changes in sensor head temperature.

There are a range of environments in the solar system in which zero-field OPMs would be well suited (see Table 1). The magnetic fields of the moon are one such example. Miniature magnetospheres on the lunar surface are related to ‘lunar swirls’. These magnetospheres exhibit similar characteristics to normal planetary magnetospheres, namely, a collisionless shock. However, a crucial difference is that they are significantly smaller than those found on planets – on the order of several 100 km.³⁰ The Lunar Prospector recorded values in the tens of nanotesla.³¹ Interplanetary magnetic field measurements, such as the heliosphere, are also interesting candidates for measurement using zero-field OPMs. Some regions of space investigated by the current Solar Explorer mission are prime examples. Here, accuracies on the picotesla level are required - within the sensitivities of current OPMs.

Zero-field OPM development is a growth field, especially for studies of bio-magnetism. In this area, measurements of magnetic fields outside the body are used to provide information about the heart (magnetocardiography, MCG) and the brain (magnetoencephalography, MEG). As a result, one can expect future improvements both in SWaP and accuracy - driven by the commercial and scientific demands for improved sensors.

4.4

Gradiometer measurements

Conventional space magnetometers are typically employed in a gradiometer configuration. Two or more sensors are used, often deployed on a boom extending from the spacecraft body. In the basic configuration, the sensor closer to the spacecraft monitors any electromagnetic interference (EMF) from the spacecraft. This signal can then be subtracted from the measurements of the second sensor to give a more accurate measure of the ambient magnetic field. Numerous approaches to gradiometer configuration and cancellation algorithms exist to maximise the gradiometer performance. Post-processing of the data combined with downlink rates often limits the time-frame for which the most accurate data is available to several days.

OPMs can be configured as gradiometers in the same multi-sensor approach as any other magnetic field sensor. This technique has been used to demonstrate applications including MEG, MCG, and low-field nuclear magnetic resonance (NMR). Miniature OPM gradiometers are currently close to commercialisation.³²

OPMs can also be configured as intrinsic gradiometers.^{33, 34} When the probing laser beam passes back through the same cell (in a different location) – or through a second cell - the difference in the magnetic field between the two points is measured. This real-time subtraction removes the need for post-processing of the data, significantly increasing the speed at which it is available for investigation. The number of measurement components and electronics is also reduced in comparison to a multi-sensor setup.