Journal of Oceanology and Limnology   2023, Vol. 41 issue(2): 804-815     PDF       
http://dx.doi.org/10.1007/s00343-022-2105-2
Institute of Oceanology, Chinese Academy of Sciences
0

Article Information

LI Xiaochen, LUO Xianhu, DENG Ming, QIU Ning, SUN Zhen, CHEN Kai
Low-noise, low-power-consumption seafloor vector magnetometer
Journal of Oceanology and Limnology, 41(2): 804-815
http://dx.doi.org/10.1007/s00343-022-2105-2

Article History

Received May 8, 2022
accepted in principle Jun. 2, 2022
accepted for publication Jul. 15, 2022
Low-noise, low-power-consumption seafloor vector magnetometer
Xiaochen LI1, Xianhu LUO2, Ming DENG1, Ning QIU3,4, Zhen SUN3,4, Kai CHEN1     
1 China University of Geosciences, Beijing 100083, China;
2 Guangzhou Marine Geological Survey, China Geological Survey, Ministry of Natural Resources, Guangzhou 511458, China;
3 SCSIO Key Laboratory of Ocean and Marginal Sea Geology, South China Sea Institute of Oceanology, Chinese Academy of Sciences, Guangzhou 510301, China;
4 Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), Guangzhou 511458, China
Abstract: The seafloor vector magnetometer is an effective tool for marine geomagnetic surveys and seafloor magnetotelluric (MT) detection. However, the noise, power consumption, cost, and volume characteristics of existing seafloor vector magnetometers are insufficient for practical use. Therefore, a low-noise, low-power-consumption seafloor vector magnetometer that can be used for data acquisition of deep-ocean geomagnetic vector components is developed and presented. A seafloor vector magnetometer mainly consists of a fluxgate sensor, data acquisition module, acoustic release module, glass sphere, frame, burn-wire release, and anchor. A new low-noise data acquisition module and a fluxgate sensor greatly reduce power consumption. Furthermore, compact size is achieved by integrating an acoustic telemetry module and replacing the acoustic release with an external burn-wire release. The new design and magnetometer characteristics reduce the volume of the instrument and the cost of hardware considerably, thereby improving the integrity and deployment efficiency of the equipment. Theoretically, it can operate for 90 days underwater at a maximum depth of 6 000 m. The seafloor vector magnetometer was tested in the South China Sea and the Philippine Sea and obtained high-quality geomagnetic data. The deep-water environment facilitates magnetic field data measurements, and the magnetometer has an approximate noise level of 10 pT/rt (Hz)@1 Hz, a peak-to-peak value error of 0.2 nT, and approximate power consumption of 200 mW. The fluxgate sensor can measure the magnetic field in the lower frequency band and realize geomagnetic field measurements over prolonged periods.
Keywords: seafloor vector magnetometer    low noise    low power consumption    
1 INTRODUCTION

Seafloor magnetism is the foundation based on which we construct geomagnetic models, understand geodynamics (Jiang and Yu, 1998) and implement underwater engineering projects (Wu and Ryall, 1987). A geomagnetic model is a mathematical description of a geomagnetic field and constitutes an important research concept related to geomagnetism. At present, there are many geomagnetic models, and several countries have begun to develop local geomagnetic models for their regions (Oehler et al., 2018). Geomagnetic models mainly include global and regional magnetic field models. Since the geomagnetic signal is weak, the fluxgate sensor has a high sensitivity to environmental parameters, which can ensure the accuracy of magnetic field measurements (Topal et al., 2019). Geomagnetic surveys have important research value in geodynamics, upper-air physics, and earthquake prediction. They provide basic data for underwater navigation, marine disaster prediction, and detection of underwater magnetic minerals (Chen et al., 2020). Seafloor magnetotelluric (MT) is a geophysical method used to study the electrical structure of the Earth's interior using the natural electromagnetic field as a source. In essence, the MT impedance is calculated by measuring the orthogonal horizontal magnetic field and electric field on the seafloor (Constable et al., 1998). Seafloor MT data can be used to image the electrical anisotropy of the oceanic upper mantle (Johansen et al., 2019), thus providing an important research basis for observing the structure and dynamics of the mantle (Tada et al., 2014; Matsuno et al., 2020). Seafloor MT is also extensively used in the oil and gas fields (Constable and Srnka, 2007), imaging the electrical conductivity of the lithosphere-asthenosphere boundary (Key et al., 2013; Naif et al., 2013; Johansen et al., 2019), and natural gas hydrate detection (Weitemeyer et al., 2011).

A seafloor vector magnetometer is an effective tool for marine geomagnetic surveys and MT detection. Given that the seafloor MT field signal is weak, and it is difficult to obtain high-precision MT data, the magnetometer should possess low noise properties. The magnetometer requires long-term continuous measurement observation on the seafloor and has high demands in terms of power consumption and time synchronization compared with onshore experiments; offshore operations are difficult and costly. Additionally, space on research vessels is limited. Given that seafloor geomagnetic observation necessitates array measurements, it is necessary to design the instrument to be compact.

Constable et al. (1998) designed a seafloor equipment system equipped with ac-coupling sensors, coils, and electric field amplifiers. The sampling rate of the logger was 25 Hz and the power consumption was less than 500 mW. The coil was used to collect magnetic field signals over a 1 000-s period, increasing the bandwidth. There is an effective MT response in water at depths equal to 1 km, and the resolution is improved in shallow seafloor. Constable (2013) later developed an improved receiver with a sampling rate of 125 Hz, power consumption of 370 mW, more acquisition channels, smaller instrument volume, improved offshore operation, and more flexible offshore data acquisition. Kasaya and Goto (2009) developed an ocean bottom electromagnetometer (OBEM) and ocean bottom electrometer (OBE) system with the advantages of low-power consumption, compact size, easy offshore operation, and low cost. Additionally, this system supports MT measurements and its measurement accuracy is 16 bits but can be improved further. Takumi et al. (2014) developed a marine magnetotelluric measurement system in a shallow water environment, which reduced the height and the electrode arm size and effectively prevented the motion noise caused by waves. This system's compact structure allows it to be used in small vessel trials, mainly in the contexts of shallow water and coastal electromagnetic exploration. Ogawa et al. (2018) designed a miniaturized magnetometer system with a fluxgate. The power consumption was only 36 mW, and the underwater operation time could be as long as 2 years. However, the system cannot complete independently marine geomagnetic measurements. It needs to be equipped with an ocean bottom seismometer platform, and the fluxgate was built in; therefore, it causes some electromagnetic interference. Quasar Geophysical Technologies has developed a small underwater electromagnetic receiver, QMax3 (QuasarGeo, 2020), which is currently the most advanced underwater electromagnetic measurement instrument with a considerably reduced size and a compact single-sphere structure. Chen et al.(2015a, b) developed OBEM-III for controlled-source electromagnetic detection; this system has a folding arm used to improve offshore efficiency as well as obtain better electromagnetic data and lower noise levels. However, this instrument has high power consumption (up to 1 600 mW) and a large size. Chen et al. (2017) developed a micro-ocean-bottom E-field receiver OBE. These authors of this study changed the previous structure, which comprised four glass spheres, into a single-sphere structure. This model has the advantages of compact size, low cost, and high operation efficiency. It can complete MT and CSEM acquisition tasks, but it cannot measure magnetic field data. Dong et al. (2021) developed a micro-ocean-bottom electromagnetic (MicrOBEM) system for CSEM data acquisition. The system supports the testing of three-component electric and two-component magnetic fields based on two induction coils. However, the system is insufficient for low-frequency noise, and its power consumption is expected to be reduced further.

In summary, the seafloor vector magnetometer at China University of Geosciences has the problems of large volume, high power consumption, and insufficient accuracy at low frequencies. Based on the existing research foundation (Chen et al., 2017; Dong et al., 2021), we developed a low-noise, lowpower-consumption seafloor vector magnetometer to meet the requirements of geomagnetic vector measurements. The design is equipped with a fluxgate sensor; it can improve measurement accuracy and support continuous seafloor operation for 90 days.

2 MATERIAL AND METHOD

Owing to the target requirements, the following specifications were set and achieved. (1) Low noise. Select fluxgates are more suitable for low-frequency magnetic field signal acquisition than inductive coils. Furthermore, a new low-noise circuit is developed to implement precise data acquisition. Signal conditioning circuitry was added to filter high-frequency noise and eliminate common-mode interference. An analog-to-digital converter (ADC) with a high-resolution, high signal-to-noise ratio (SNR), and low drift characteristics was selected for the acquisition circuit. To balance the efficiency and noise, the power supply design used the cascade scheme of the switching power supply and low dropout regulator. (2) Low-power consumption. The miniature seafloor magnetometer has a new lowpower acquisition module with a low-power fluxgate sensor and advanced power management technology to reduce the power consumption of the entire unit to 200 mW compared with the rated consumption of 1 600 mW of the existing ocean bottom electromagnetic receiver (OBEM-III). In the acquisition circuit design, multichannel, a lowpower ADC is adopted, and the power consumption of each channel is only 2 mW. In terms of control, an ultralow-power microcontroller unit (MCU) module was designed. Peripherals can be turned off to make the entire machine enter sleep mode. (3) Low cost. The traditional acoustic release device is expensive and heavy. The release and recovery mechanism was designed by integrating an underwater acoustic telemetry modem and by adding an external burn-wire release mechanism. The lownoise, low-power-consumption seafloor vector magnetometer used a 43.18-cm glass sphere and a smaller fluxgate sensor to achieve a compact design. Compared with the micrOBE (2017) and micrOBEM (2021), the fluxgate geomagnetic vector acquisition was designed to ensure that the instrument remained compact; this reduced further the power consumption and hardware cost of the instrument, thereby improving the degree of integration and operation efficiency of the equipment.

Figure 1a shows the structure of the seafloor vector magnetometer, which mainly consists of a data acquisition module, fluxgate sensor, acoustic transducer, acoustic telemetry modem (ATM), glass sphere, frame, burn-wire release, and an anchor. Its operation is classified into two parts: the data acquisition module and acoustic release module. The 43.18-cm glass sphere was used to provide buoyancy equal to 26 kg. The titanium alloy cabin had an inner diameter of 70 mm, outer diameter of 80 mm, inner length of 500 mm, outer length of 585 mm, and pressure resistance of 60 MPa. The data acquisition module realized eleven-channel (11CH) electromagnetic signal acquisition and storage. The circuit was placed in the cabin with an ATM. The fluxgate sensor was built in the cabin to measure the three-axis orthogonal magnetic field signal. The acoustic release unit was composed of an ATM, a transducer, and a burn-wire release.

Fig.1 Deployment of the seafloor vector magnetometer instrument (a); photograph of the data acquisition module in the titanium alloy cabin (b)
2.1 Top design

The data acquisition module is one of the key technologies in equipment design and is necessary for stable operation. The data acquisition module was equipped with an acquisition circuit, an amplifier, and a lithium battery pack. To reduce the structural complexity, only the entire acquisition cabin side hole leads to the necessary instrument interface, including fluxgate sensor access (electrode and induction-coil interface standby), communication and charging interface, and a transducer interface. Figure 2 shows the internal structure of the acquisition capsule. The battery pack is on the left, and the unilateral interface between the acquisition capsule and the outside is on the right. The acquisition module includes (from left to right) the acquisition (ACQ) circuits, amplifier, and ATM (Fig. 1b). The battery pack supports thirty-six 18650 lithium batteries, and the power of each battery is 12.58 Wh. The magnetometer, which has a power consumption of 200 mW, can continuously work on the seafloor for 90 days.

Fig.2 Functional block diagram of the seafloor vector magnetometer The dashed box indicates the magnetometer system that is installed in the pressure cabin.

The circuit in the data acquisition module mainly includes the ACQ module and the ATM. The deck unit provides global positioning system (GPS) timing, direct current (DC) charging power supply, and a universal serial bus (USB) communication for the in-cabin circuits. The battery pack is located on the nonperforated side of the rear end of the acquisition module and consists of two groups. To ensure that the release function of the instrument will not be affected after it stops its acquisition due to insufficient power, the battery pack consists of a separate power supply. One group of batteries supplies power to the acquisition circuit, and the other group of batteries is connected to the ATM module.

2.2 Fluxgate sensor

The focus of marine geomagnetic instrumentation has been on drift-free measurements of magnetic fields at DC frequencies using fluxgate magnetometers (Constable et al., 1998). Compared with the proton precession magnetometer, optical pump, induction coil magnetic sensor, and other electromagnetic sensors, the fluxgate sensor has numerous advantages pertaining to its small size, low-power consumption, and low cost. The seafloor vector magnetometer is equipped with a unipolar differential fluxgate sensor. The Bartington Mag648 low-power, threeaxis fluxgate sensor is used for weak magnetic field signal acquisitions. Compared with the inductioncoil magnetic sensor, its size is smaller and the power consumption is lower. Owing to the lowfrequency and low-noise measurement requirements of long-period MT, the fluxgate sensor measurement scheme has better advantages. Figure 3 shows the comparison of noise levels between the fluxgate and induction-coil sensor, it is more suitable for testing low-frequency magnetic field signals. These threeaxis fluxgate sensors offer low-noise magnetic field measurements from DC up to 30 Hz. The power consumption of Mag648 is less than 15 mW, the noise is less than 10 pT/rt(Hz)@1 Hz, and the measurement range is ±60 μT. Their exceptionally low-power consumption and compact size make them ideal for battery-powered applications. The volume of the cabin is D55 mm×L260 mm, air weight is 2.4 kg, and water weight is 1.8 kg. Figure 4a shows the internal circuit of the fluxgate sensor, and Fig. 4b shows the photograph of the cabin.

Fig.3 Comparison of noise levels between fluxgate and induction-coil sensor
Fig.4 Photograph of fluxgate sensor (MAG-648 from Bartington) (a); photograph of fluxgate sensor made of titanium alloy with a pressure cable (b), capable of withstanding 60 MPa, and using an MCBH8M watertight connector
2.3 Data acquisition module

The data acquisition module realizes signal measurements and storage and includes an MCU, ADC, TCVCXO oscillators, DC/DC, and a TF card. The personal computer interacts with the data acquisition module via a USB interface. According to the schematic of the acquisition module shown in Fig. 5, the data acquisition module supports 11 channels of synchronous acquisition, three of which are externally connected with fluxgate sensors, and another spare channel can be externally connected with electrodes and coils. The MCU directly manages multiple ADCs and writes data to TF cards. The temporal information comes from the GPS module outside the deck box, and the time data are read through the UART port reserved for the module by the MCU. The sampling timing of the magnetic field is important for observations. The internal clock precision of MCU is only 1 s, and passive crystal is unstable, long time work accumulate large errors. Therefore, using the relative time recording method, high-precision clock source frequency division is employed to obtain the local clock, and the calculation of the current time only needs the local clock and synchronous GPS times. To make the acquisition circuit and the host computer directly communicate, a USB hub module is designed to realize the conversion between the USB and the UART port. At the same time, the PC can also access the TF card through the USB hub and the TF reader, thus allowing it to download and upload information without the MCU, thus improving the data transmission rate. The average measured data download speed is more than 18 MBps. The battery pack supports thirty-six 18650 lithium batteries, which is 25.2 V at full charge, and supplies power to each module via DC/DC links. The current monitor is added to detect the current in real time, and the threshold can be controlled by the program. When the current exceeds the set threshold, the MCU shuts off the peripheral power supply to achieve the function of overload protection.

Fig.5 Block diagram of the acquisition circuit, configured to collect three-component fluxgate data, supporting both external induction coils and electrodes
2.4 Acoustic release module

The ATM, external transducer, and burn-wire release cooperate to achieve acoustic ranging, burnwire switching, and acquisition subject to the control of the constant current source Is of the underwater acoustic module output. The static power consumption of the acoustic telemetry circuit is approximately 165 μA. The traditional acoustic release device is bulky and costly. The weight of the Oceano 2 500 Universal from iXblue is 30 kg in air and 22 kg in water, and the size is D143 mm× L849 mm. The primary goal of miniaturization is to solve the bulkiness problem of the acoustic release device. Accordingly, the customized underwater acoustic telemetry board, transducer, and burn-wire mechanism were used. The underwater acoustic telemetry board had a diameter of 68 mm and a height of 50 mm. It can coordinate the underwater acoustic deck unit to achieve ranging and burn-wire switch functions. After the underwater acoustic telemetry board receives the burn-wire open command, the constant current source is enabled, the current is approximately 0.7 A, and the burnwire mechanism is fused after approximately 30 s. Based on the underwater acoustic telemetry control software, the serial number of the instrument can be selected, the gain, signal transmission amplitude, and response amplitude can be changed, and the switching state of the burn-wire can be controlled. The query status returns to the device status screen to view parameters such as voltage, temperature, distance, noise energy, signal energy, and burn-wire status.

The underwater acoustic deck unit has also been miniaturized (Fig. 6) with dimensions of 230 mm× 150 mm×87 mm.

Fig.6 Schematic of the hardware of the deck unit used for operating the acoustic telemetry modem
3 RESULT 3.1 Indoor test

To test its function and circuit performance indices, the basic parameters of the acquisition module, such as the noise floor, dynamic range, time drift, and power consumption were tested indoors.

Accordingly, we short-circuited the input of each channel to the ground, set the sampling rate to 300 Hz, and set the power supply voltage to 25 V. After a period of testing, we checked the noise time series (Fig. 7a) and calculated the noise power spectral density (Fig. 7b) to represent its statistical characteristics in the frequency domain. The calculated noise level of the electric field channel (E) was better than 0.1 nV/m/rt (Hz)@1 Hz (the measured pole distance was 8 m), that of the fluxgate channel (B) was 0.2 μV/rt(Hz)@1 Hz, and that of the induction-coil channel (H) was 0.5 μV/rt (Hz)@1 Hz. The sensitivity of the fluxgate sensor and induction coil was 50 μV/nT and 100 mV/nT, respectively, and the equivalent noise was better than 10 pT/rt(Hz)@1 Hz and 0.1 pT/rt(Hz)@1 Hz. The peak-to-peak errors of E, B, and H were 2 nT/m, 0.2 nT, and 2 pT. Furthermore, the dynamic range of each channel was approximately 118 dB, 118 dB, and 123.4 dB, respectively. The design adopted a crystal oscillator with a precision of 50 parts per billion and a time drift of less than 5 ms/d.

Fig.7 Noise time series test results (a); power spectral density diagram of channel (b)

The voltage source was used to simulate the 25-V power supply (6-series) lithium battery. The power consumption test results of the acquisition board in different states are listed in Table 1. The test results show that with the use of the 25-V power supply, the standby power consumption is less than 25 mW, and the acquisition power consumption is approximately 175 mW; this operational setting realizes low-power data acquisition. The power consumption of the acquisition circuit and fluxgate sensor is 175 and 25 mW, respectively. The internal battery pack of the instrument contains thirty-six 18650 batteries, which can work for approximately 90 days at most with the magnetic field signal collected by the fluxgate sensor (approximately 200 mW).

Table 1 Power consumption test results of the acquisition board in different states
3.2 Marine experiment

We conducted sea trials in two different places to evaluate the field performance of the low-noise, lowpower-consumption seafloor vector magnetometer, and assess and verify its advantages of compact size and low-power consumption in deep water.

The first field operation was conducted in the southern part of the South China Sea. We used the R/V Shiyanerhao (Experiment in Chinese) of the South China Sea Institute of Oceanology, Chinese Academy of Sciences, to conduct deep-water MT survey experiments. The water depth of the experiment area (Fig. 8) was between 1 400 m and 2 000 m. Five station-testing tasks were completed. The placement coordinates of the seafloor MT station are listed in Table 2. The measurement data of station S5 were invalid owing to cable failure. Thus, these measurements were performed again. Deep-water operations not only tested the pressure performance of the equipment but also prolonged the floating time of the equipment. However, the low-electromagnetic interference in deep-water conditions helps improve the MT SNR. In the deep-water area, the flow rate is low, and there is no obvious electromagnetic interference from moving seawater, which is relatively quiet most of the time.

Fig.8 The experimental area Left: the working area; right: the stations. Map review No. GS(2021)5442.
Table 2 Point coordinates of seafloor magnetotelluric (MT) station

Figure 9a shows recordings spanning 4 500 s of the entire time series of station S3. The correlation of electromagnetic components is significant, and there is no obvious interference. The amplitudes of the Ey and Bx components are stronger than those of the Ex and By components. The peak-to-peak value of the electric field is 400 nV/m, and the peak-to-peak value of the magnetic field is 10 pT. The time series of all stations were filtered, unqualified data were reselected and reprocessed, and the power spectra of the data files were edited with better quality. Data processing included gain correction and pulse interference suppression. The pulse was caused by electromagnetic interference owing to the instantaneous current change associated with the writing process on the TF card. Figure 9b shows the comparison before and after pulse interference was deselected. Figure 10 shows the data processing results of five stations. The Ey and Bx curves of station S4, and the Ex and By data of station S5 are poor, whereas the other MT curves show high-quality data. The frequency of most stations was lower than 1 000 s.

Fig.9 Station S3 sample time series collected at a water depth of 1 826 m in the southern part of the South China Sea (a); comparison of outcomes before and after pulse interference was deselected (b), in which the red circle is the processed pulse interference signal
Fig.10 Apparent resistivity and impedance phase calculated from MT-sounding data of five stations Red: Ex and By components; blue: Ey and Bx components.

The first offshore experiment was conducted at five stations. The recovery rate was 100%. Although station S5 was retested owing to channel faults, valuable MT-sounding data were obtained for the other stations. The magnetometer measurement scheme was verified, but the reliability of the measurement channel needs to be strengthened further. At a water depth of nearly 2 000 m, the sinking time is 60 min, the speed is 33 m/min, the floating time is 80 min, and the floating speed is approximately 25 m/min. To lift the floating speed of the instrument, the buoyancy ratio needs to be improved further.

The second offshore experiment was conducted by R/V Haiyangdizhiliuhao (Marina Geology #6 in Chinese) 202104 (voyage number). The sea trial operation area was located in the Philippine Sea. The Philippine Sea is 3 450 km long from north to south and 2 000 km wide from east to west. It covers an area of 5.8 million km2 and has an average depth of 6 km. The instrument deployment site was located at approximately 2 400 km from Guangzhou (Fig. 11), at the intersection of the Palau Ridge and the spreading center of the Central Basin. The water depth was approximately 5.5 km.

Fig.11 Site location map for the second seafloor vector magnetometer experiment in the Philippine Sea Map review No. GS(2021)5442.

Before the seafloor magnetometer was used in the sea, instrument pressure and underwater acoustic release module tests were conducted on August 28, 2021 at a water depth of ~5.3 km in the launching area. Figure 12 shows some of the experimental steps. As shown in Fig. 12b, the water depth of the floating upward receiver was 4.647 km and the signal energy and noise level were within the allowed range. Figure 12 shows the APP interface on the mobile phone to complete the instruction transmission and status display of ranging and the burn-wire switch. Owing to the weight of the hammer, the instrument was not fully placed on the sea floor to ensure the safety of the equipment (Fig. 12a). As shown in Fig. 12c, the app was returned to the control interface, and the burn-wire was opened, which proves that the improved miniaturized and low-cost acoustic telemetry system can achieve deep-water command transmission and control.

Fig.12 Instrument pressure test and underwater acoustic release system communication test (a); APP status query interface when the ship-water depth was 4.647 km (b); APP returns to the control interface (c), and the burn-wire was opened

After the completion of the pressure and acoustic release tests, the seafloor vector magnetometer was ready for launch on September 2, 2021. Firstly, the acquisition cable, fluxgate cable, and acoustic release device were connected to the deck, and the telemetry and burn-wire release functions were tested on the deck. Subsequently, the acquisition module and acoustic release unit were connected by a stainless steel cable, the user interface was opened, GPS time was synchronized, storage status was checked, and the sampling parameters were set. Following data collection for more than one week, the receiver was recovered from a depth of 5 425 m on September 11. After the confirmation of the successful release, the distance between the rising surface and the water surface was monitored every 20–30 min, and the device was ready for recovery.

Seafloor vector magnetometer was successfully recovered from Sept. 2 to 11, and nearly 9 days of complete ultra-deep-water magnetic field data were obtained. These are the first magnetic data acquired from this depth in the Philippine Sea by China. The geomagnetic variation data at 60 s intervals are shown in Fig. 13, including the northern (x), eastern (y), and vertical (z) components of geomagnetic variation. Geomagnetic data from the Dalat geomagnetic station were also plotted for comparison. The observed data from seafloor magnetometer have similar characteristics to those from Dalat (Fig. 13). There are seven significant geomagnetic diurnal variations; the peak-to-peak values of the x, y, and z components are 40 nT, 60 nT, and 60 nT, respectively. There is a minor difference between the data from the seafloor magnetometer and those from Dalat, which is due to the low-pass difference of the seawater. The altitude and orientation of the instrument deviate from those of the three-axis component of the underwater geomagnetic coordinate system, which is not leveled, and the local and regional conductivity structures are different at each site. Therefore, we plotted the curve of the total geomagnetic field (Fig. 13) and found that the peakto-peak value was 80 nT. The data from the geomagnetic observation station were used as a remote reference to prove the validity of the test data. In the calculated period range, as shown in Fig. 14, the coherence between the geomagnetic data from Dalat and those from the seafloor magnetometer site in the geomagnetic total field is greater than 0.4. The offshore test was conducted in pressure conditions, and the recovery rate of one station was 100%. The reliability of the small-volume and low-cost burn-wire scheme was verified. The recovered and downloaded data were continuous and complete, and the measurement scheme of the entire machine was verified. After testing, according to a proper counterweight calculation, the sinking and rising speeds were the same at a water depth of 5.5 km, and the speed was approximately 1.5 m/s.

Fig.13 Time series plots of the three orthogonal components and total field of the geomagnetic field The black color denotes the seafloor vector magnetometer data, and the red color denotes the Dalat geomagnetic observatory data.
Fig.14 Coherence between the geomagnetic data from Dalat and those from the seafloor vector magnetometer site in terms of the total geomagnetic field

Compared with previous seabed electromagnetic receivers (Table 3), this design can execute fluxgate sensor tests and has achieved innovative measurements in terms of the full-band low-noise magnetic field vector, low-power consumption, and underwater acoustic release. The successful development of the instrument indicates that it has an ultra-deep-water magnetic field operation capability of 6 000 m, and provides strong technical support for deep-ocean exploration research.

Table 3 Comparison of main performance parameters
4 DISCUSSION

This study focused on the development of a miniaturized seafloor vector magnetometer with low-noise, low-power consumption, and low-cost characteristics. This magnetometer can record MT data and geomagnetic field with sampling frequency of 100 Hz as well as battery voltage, current, storage and hardware statuses. The systems were used for the two marine experiments and successfully acquired and recorded electromagnetic and geomagnetic field variations during the test period. The data indicated the expected performance of the seafloor vector magnetometer. The observed geomagnetic field changes are in good agreement with the observation data of Dalat ground geomagnetic observatory. The correlation of electromagnetic components of MT data is significant, and there is no obvious interference. We developed a low-power acquisition circuit, and the power consumption of the entire system was reduced from the original consumption of 1 600–200 mW. The seafloor magnetometer integrated an underwater acoustic telemetry solution that eliminated the need for traditional acoustic release devices. This resulted in a considerable reduction in size and cost. The depth of the geomagnetic field test at 6 000-m verified the reliability and stability of the instrument and obtained high-quality geomagnetic field data. The fluxgate sensor can measure the magnetic field in a lower frequency band and realize long-period geomagnetic field measurement. In future work, we will broaden the observation frequency band, while supporting electrodes, induction coils, and fluxgate measurements. The circuit has been completed, but it is still necessary to address the possible mutual interference between sensors and the slow floating caused by insufficient buoyancy.

5 DATA AVAILABILITY STATEMENT

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

6 ACKNOWLEDGMENT

We thank Zhigang ZHANG from Guangzhou Marine Geological Survey for his assistance with the module architecture. Zhongliang WU from Guangzhou Marine Geological Survey provided valuable advice on hardware design. We appreciate the extensive support offered by the captains, ship crew, and marine technicians of the R/Vs Haiyangdizhiliuhao (Marine Science #6) and Shiyanerhao (Experiment #2).

References
Chen K, Deng M, Luo X H, et al. 2017. A micro oceanbottom E-field receiver. Geophysics, 82(5): E233-E241. DOI:10.1190/geo2016-0242.1
Chen K, Wei W B, Deng M, et al. 2015a. A new marine controlled-source electromagnetic receiver with an acoustic telemetry modem and arm-folding mechanism. Geophysical Prospecting, 63(6): 1420-1429. DOI:10.1111/1365-2478.12297
Chen K, Wei W B, Deng M, et al. 2015b. A seafloor electromagnetic receiver for marine magnetotellurics and marine controlled-source electromagnetic sounding. Applied Geophysical, 12: 317-326. DOI:10.1007/s11770-015-0494-0
Chen Y, Yang J C, Dada O A, et al. 2020. Magnetic properties indicate the sources of hadal sediments in the Yap Trench, northwest Pacific Ocean. Journal of Oceanology and Limnology, 38(3): 665-678. DOI:10.1007/s00343-019-8370-z
Constable S. 2013. Review paper: instrumentation for marine magnetotelluric and controlled source electromagnetic sounding. Geophysics Prospecting, 61(S1): 505-532.
Constable S C, Orange A S, Hoversten G M, et al. 1998. Marine magnetotellurics for petroleum exploration part I: a sea-floor equipment system. Geophysics, 63(3): 816-825. DOI:10.1190/1.1444393
Constable S C, Srnka L J. 2007. An introduction to marine controlled-source electromagnetic methods for hydrocarbon exploration. Geophysics, 7(2): WA3-WA12. DOI:10.1190/1.2432483
Dong Z, Zhang J, Yang G Y, et al. 2021. Micro-ocean-bottom electromagnetic receiver for controlled-source electromagnetic and magnetotelluric data acquisition. Review of Scientific Instruments, 92(4): 044705. DOI:10.1063/5.0044412
Jiang X D, Yu Z H. 1998. Study on the fractal character of magnetic anomaly fields and the delineation of tectonic elements in the South China Sea. Chinese Journal of Oceanology and Limnology, 16(1): 28-35. DOI:10.1007/BF02848214
Johansen S E, Panzner M, Mittet R, et al. 2019. Deep electrical imaging of the ultraslow-spreading Mohns Ridge. Nature, 567(7748): 379-383. DOI:10.1038/s41586-019-1010-0
Kasaya T, Goto T N. 2009. A small ocean bottom electromagnetometer and ocean bottom electrometer system with an arm-folding mechanism (Technical Report). Exploration Geophysics, 40(1): 41-48. DOI:10.1071/EG08118
Key K, Constable S, Liu L J, et al. 2013. Electrical image of passive mantle upwelling beneath the northern East Pacific Rise. Nature, 495(7742): 499-502. DOI:10.1038/nature11932
Matsuno T, Baba K, Utada H. 2020. Probing 1-D electrical anisotropy in the oceanic upper mantle from seafloor magnetotelluric array data. Geophysical Journal International, 222(3): 1502-1525. DOI:10.1093/gji/ggaa221
Naif S, Key K, Constable S, et al. 2013. Melt-rich channel observed at the lithosphere-asthenosphere boundary. Nature, 495(7441): 356-359. DOI:10.1038/nature11939
Oehler J F, Rouxel D, Lequentrec-Lalancette M F. 2018. Comparison of global geomagnetic field models and evaluation using marine datasets in the north-eastern Atlantic Ocean and western Mediterranean Sea. Earth, Planets and Space, 70: 99. DOI:10.1186/s40623-018-0872-y
Ogawa K, Matsuno T, Ichihara H, et al. 2018. A new miniaturized magnetometer system for long-term distributed observation on the seafloor. Earth, Planets and Space, 70: 111. DOI:10.1186/s40623-018-0877-6
QuasarGeo. 2020. QMax EM3, Quasar Geophysical Technologies.
Tada M, Baba K, Utada H. 2014. Three-dimensional inversion of seafloor magnetotelluric data collected in the Philippine Sea and the western margin of the northwest Pacific Ocean. Geochemistry, Geophysics, Geosystems, 15(7): 2895-2917. DOI:10.1002/2014GC005421
Takumi U, Mitsuhata Y, Toshihiro U, et al. 2014. A new marine magnetotelluric measurement system in a shallow-water environment for hydrogeological study. Journal of Applied Geophysics, 100: 23-31.
Topal U, Can H, Çelik O M, et al. 2019. Design of fluxgate sensors for different applications from geology to medicine. Journal of Superconductivity and Novel Magnetism, 32(4): 839-844. DOI:10.1007/s10948-018-4781-x
Weitemeyer K A, Constable S, Tréhu A M. 2011. A marine electromagnetic survey to detect gas hydrate at Hydrate Ridge, Oregon. Geophysical Journal International, 187(1): 45-62. DOI:10.1111/j.1365-246X.2011.05105.x
Wu M X, Ryall P J C. 1987. An application of marine geomagnetic survey to shallow sea engineering geology. Chinese Journal of Oceanology and Limnology, 5(1): 51-58. DOI:10.1007/BF02848522