1. Introduction
Submarine hydrothermal and cold springs are two special ocean bottom areas. As the name suggests, the temperature of hydrothermal fluid is relatively high, while the temperature of cold springs is not much different from that of the surrounding seawater. Hot spring areas are widely distributed in areas where volcanic activity is active, such as midocean ridges and back-arc basins. Most of their depths are between 2200 and 2800 m (German and Baker 2004). The mechanism of its formation is as follows: the expansion of the oceanic crust in the midocean ridge zone forms a large number of small cracks in the rock. The cold seawater penetrates down the cracks and reacts violently when it encounters the high-temperature rock heated by magma. The seawater carries it. The heat and a large amount of metal elements in the rock are ejected from the cracks, forming a hot spring with a temperature of up to 350°–400°C. The deep-sea hydrothermal fluid is the window of material and energy exchange between the ocean and Earth’s crust, and its activities have a high research value on the impact of the ocean physical and chemical environment, ocean circulation, and global climate (Pelayo et al. 1994; Zhai et al. 2006; Love et al. 2017).
The cold spring area, similar to hydrothermal activities, also has high research value in terms of biological resources and geological activities. The special environment of the cold spring area makes the microorganisms in this area have special metabolic pathways and functions. In recent years, researchers have conducted more in-depth studies on the organisms and microorganisms in the methane leakage area (Yu et al. 2019; Cui et al. 2019; Case et al. 2018). In addition, the formation of the cold spring area is closely related to methane leakage, and the strong methane leakage area usually develops high saturation hydrate (Henry et al. 1996; Chen et al. 2011). It is generally accepted that organic matter content and accumulation rate at the seafloor attenuate with increasing water depth. However, hadal trenches, which represent the deepest portion of the hydrosphere on Earth, do not abide by this universal rule (Luo et al. 2017).
Compared with in situ measurement, bringing the hydrothermal sample back to the laboratory for analysis can obtain higher measurement accuracy and a wider range of detection parameters. The collection of high-quality deep seawater requires an excellent water sampler. The earliest water harvesting device can be traced back to the nineteenth century, and various water harvesting devices emerged in endlessly until the 1950s (Sheryll 2009). On the one hand, the improved model of the early sampler is still in large use due to its simplicity, reliability, mature technology, and high cost performance, for example, the Niskin water harvester used with conductivity–temperature–depth (CTD) sensors (Niskin 1962) array. The remote control sequence sampler (Wu et al. 2015) of Wu Shijun and others have sequence triggering methods. ANEMONE-11 proposed by Okamura et al. (2013) has 128 sampling channels. It is mainly mounted on autonomous underwater vehicles (AUVs) and used to monitor the chemical and biological components of the water body. The groove wheels on the inlet and outlet valves of the WHATS series of third-generation water harvesters in Japan are designed so that when the blocking lever moves from right to left, each ball valve just completes an opening or closing action in turn. Thus, the sequence control of a total of eight valves at both ends of the four sampling chambers is realized by one moving part (Miyazaki et al. 2017; Saegusa et al. 2006). A novel sampler was designed by Hunan University of Science and Technology to obtain intact sediment samples at full ocean depth (He et al. 2021). In addition, they reviewed and analyzed the key technologies of various deep-sea sediment samplers such as sealing, pressure and temperature retaining, low-disturbance sampling, and no-pressure drop transfer. Then, they identified the deficiencies of key technologies for deep-sea sediment sampling (He et al. 2020).
The sampler needs to have an excellent ability to retain the gas components in the seawater, and it needs to be difficult to inhale other impurities. The loss of the original pressure of the sediment will cause the decomposition of organic matter and the death of microorganisms in the sample (Huang et al. 2018). A sampler can be deployed in the deepest trenches in the ocean as a result of its ability to resist ultrahigh pressure and its good bidirectional sealing performance. Three sets of this new sampler were constructed and field tested in the Mariana Trench (Wu et al. 2018). A pressure-retaining sampler (PRS) can maintain seawater samples under in situ pressure during recovery (Peoples et at. 2019). The State Key Laboratory of Fluid Power and Mechatronic Systems of Zhejiang University has developed a novel pressure-tight sampler with a metal seal capable of a acquiring high-purity fluid samples from deep-sea hydrothermal vents. All parts of the sampler in contact with the sampler are made of titanium, which can reduce the organic carbon blank and makes it suitable for collecting organic samples (Wang et al. 2020). The University of South Carolina has designed a hydrothermal organic geochemical (HOG) sampler for deployment on a deep-sea submersible that minimizes entrainment of surrounding seafloor water when sampling (Lang and Benitez-Nelson 2021). On the other hand, the cost of ocean exploration and deep-sea diving is very high. The water harvester has sequential sampling, that is, the ability to sample multiple times in one dive is very necessary, and each sampling should not take too long. In response to the above problems, this paper proposes a horizontal flowthrough water sampling device. The sampling chamber is arranged in a U shape, the inlet and the outlet are on the same side, only one trigger mechanism is arranged on one side to control the opening and closing of the inlet and outlet valves at the same time, which solves the defects of the traditional trigger mechanism that the structure is complicated and the space required is large. The advantage of using the displacement mechanism of the target sample and prefilled pure water for sampling is that the sampling channel can be flushed during the sampling process to reduce pollution and increase the sampling concentration. For different flow rates and the structures of the sampling cavity, numerical simulation was used to simulate the working process of the sampler. The results showed that the structure of the sampling cavity had a significant impact on the displacement efficiency at different flow rates. Through the analysis of the results, the optimal structure of the sampling cavity was obtained. The optimized flowthrough sampling chamber can effectively improve the purity of sampling and reduce cross contamination between multiple samplings and, at the same time, has a high displacement efficiency. With trigger technology, the sequence trigger control of multiple sampling channels is realized in a compact structure.
The significance of deep-sea sampling technology research and the research results at home and abroad are expounded in the first section. The setting and calculation equations of the water collector simulation model are introduced in the second section, the displacement principle is introduced, and the simulation results are obtained by numerical calculation. The structural design of the water collector is introduced in section 3. Experiments are designed in section 4 and the experimental data are compared with the numerical results. The conclusions of the study are summarized in section 5.
2. Numerical calculation and analysis
a. Simulation model
For the displacement performance of the horizontal sampling chamber, the influence of the inlet and outlet positions on the displacement performance is mainly considered. In the study on the displacement efficiency of the electrode cavity for deep-sea exploration by Fan et al. (2010), the SS-I cavity is an approximate cylindrical shape, and the inlet and outlet are on the horizontal cylindrical axis. Both simulation and experiment are proved that under this arrangement, after the light liquid displaces the heavy liquid in the upper part of the cavity, it cannot well displace the heavy liquid in the lower half, thus forming a displacement blind zone. Based on the above reasons, the inlet and outlet of this horizontal sampler are initially planned to be arranged at different heights, and the specific effects will be discussed through simulation.
The selected simulation physical model of the horizontal sampling cavity is shown in Fig. 1. Figure 1a shows two capsule-shaped cavities with the inlet and outlet in the middle, Fig. 1b shows two capsule-shaped cavities, with the inlet at the bottom and the outlet at the top, both close to the wall of the sampling cavity, Fig. 1c is cylindrical, the structure and the layout of the entrance and exit are the same as Fig. 1b, and in Fig. 1d there is an inclination angle of 10° on the basis of Fig. 1b. The diameter of a single sampling cavity in the four layout schemes is 30 mm, and the volume is 50 mL.
Simulation calculations are carried out for a total of eight working conditions from W1 to W8, the density of the prefilled liquid is 1000 kg m3, and the key parameters are shown in Table 1.
Working conditions selected for horizontal throughflow sampling channel simulation.
ANSYS Fluent is used for simulation calculations and the k–ε equation is used to solve the simulation model. To describe the flow of the near-wall flow field better, the enhanced wall treatment method is used and the full buoyancy effects option is turned on to emphasize the buoyancy in the flow field. In the simulation, the heat exchange between the two liquids is not considered, that is, there is no chemical reaction or heavy metal precipitation between the sample and the prefilled pure water. On the other hand, although there is a certain density difference between the sample and the prefilled pure water, this density difference is very small. To facilitate convergence, it can be considered that the Boussinesq assumption is valid, that is, the density change caused by the uneven concentration of the substance is very small. When solving the Navier–Stokes equation, in addition to the gravity term, the density in other terms is regarded as a constant.
CFD is the English abbreviation of computational fluid dynamics, which is the use of numerical calculation methods to solve the basic governing equations of fluid mechanics, mainly including mass equation, momentum equation, and energy equation.
b. Displacement rate
c. The result of the calculation
As shown in Fig. 2, the heavy liquid enters from the entrance located in the middle and then slides downward under the action of gravity. The forward of the heavy liquid sneaks to the left along the lower wall, and there is a certain layering phenomenon between the heavy liquid and the light liquid. Above the dividing line, the remaining light liquid has become a blind spot for displacement. After a long time (600 s), there is still a large light liquid area in the sampling chamber.
When the inlet and outlet positions of the sampling cavity are located on the upper and lower sides, respectively, the overall situation of the displacement is obviously different. When the heavy liquid first enters the sampling cavity, its lower side is close to the wall of the sampling cavity and moves to the left as a whole. There will also be a certain layering phenomenon between the heavy liquid and the light liquid. The heavy liquid precipitates in the lower part and continuously raises the interface. The last small part of the light liquid can eventually be replaced.
After increasing the inlet flow rate, it can be found that both of them did not produce significant stratification (Fig. 3). The heavy liquid and the light liquid were mixed vigorously, and W4 was slightly behind.
It can be clearly seen from Fig. 4a that at low speeds, W1 working conditions will produce displacement blind areas. Although the initial trend is similar to that of W3, the displacement rate will soon show a slow growth situation, and the displacement effect poor. The displacement rate of W3 working condition has increased steadily with time. At low speeds, the U2 scheme with the inlet and outlet positions on both sides is obviously better than the U1 scheme with the inlet and outlet positions in the center.
At high speeds, the two behave similarly, and the W2 operating condition is slightly better than the W4 operating condition. In the case of high flow rate, the displacement process is dominated by the mixing process.
Figure 4b is the relationship curve between body efficiency E(e) and inflow volume ratio e. At low speeds, the slope of W1 decreases earlier due to the existence of the displacement blind zone, the linear growth phase of W3 is maintained for a longer time, and the inflow volume ratio when the displacement is completed is also small.
Although low-speed sampling takes a long time to complete the sampling, it is more efficient to collect samples that flow out slowly, and it is not easy to inhale the surrounding seawater.
Figures 5a and 5b are the displacement effect diagrams under W5 and W6 working conditions, respectively. The trend of the overall displacement process is basically the same as that of the capsule-type sampling chamber. The main purpose is to verify whether the edge of the cylindrical sampling cavity will become a blind spot for displacement.
Figure 6 is the displacement performance curve of the cylindrical sampling chamber and the capsule-type sampling chamber at low speed and high speed. It can be seen from this that, when the volume and diameter are the same, there is almost no difference in displacement performance between the two. However, in actual situations, there may be some solid particles in the water samples, which are easy to accumulate in the side seams of the cylindrical sampling cavity, which affects the reliability of the samples. Therefore, although the processing of the capsule-type sampling cavity is more difficult, it has higher reliability.
There is an inclination angle of 10° between the sampling cavity and the horizontal in W7 and W8 working conditions. In the W3 working condition, it can be seen that there is light liquid accumulation in the sampling cavity at 120 s. It is expected that the sampling cavity can be improved by placing the sampling cavity inclined.
Combining Figs. 7 and 8, it can be seen that the overall trend of the displacement process of W7 and W8 is similar to that of W3 and W4. The difference is that, at low speeds, when the sampling chamber is placed obliquely, the light liquid on the upper part can float and exit from the outlet along the slope of the upper surface of the sampling chamber under the action of buoyancy, so there is no tiny blind zone. On the other hand, the slightly elevated position of the outlet also makes the interface reach the outlet later, so it has a longer linear growth period of displacement.
3. Water sampler design
a. Sampling channel
The sampling channel of the vertical throughflow water sampler is generally relatively slender. Therefore, the interior of the water sampler is relatively open as a whole, and the space utilization rate is not high. To improve the space utilization rate and further increase the reuse rate of the mechanism, the sampling channel of the vertical throughflow water sampler is folded into a U shape from the middle. Compared with the vertical throughflow water sampler, its advantage is that the length of the sampling channel is shortened to about half of the original, so that the overall length of the device is reduced accordingly, and at the same time, the sampling cavity can also be appropriately larger. The outlet ball valve and the inlet ball valve are now located on the same side. The vertical throughflow water sampler needs to arrange two sets of trigger mechanisms on both sides, but now it only needs to arrange one set on one side.
As shown in Fig. 9, similar to the vertical water sampler, its main components include the sampling channel, trigger mechanism, and trigger mechanism drive unit. The main difference is that in Fig. 10, the sampling channel is U-shaped and arranged horizontally. There is only one set of trigger mechanism, and the sampling cavity has a larger volume.
Starting from the sampling inlet, it is the inlet ball valve, the sampling cavity 1, the middle ball valve, the connecting pipe, the middle ball valve, the sampling cavity 2, and the outlet ball valve. Different parts are connected by tube.
b. Trigger mechanism
The inlet and outlet of the horizontal throughflow sampling channel are arranged on the same side. In this regard, a new type of incomplete bevel gear mechanism is designed. As shown in Fig. 11, it is mainly composed of a gear plate, pinion A, pinion B, and other auxiliary structures. The gear plate is connected to the output shaft of the motor. Pinion A is connected with the rotating shaft of the inlet ball valve, and pinion B is connected with the rotating shaft of the outlet ball valve.
The four sampling channels are arranged counterclockwise, with the outlet ball valve in front and the inlet ball valve in the back. When the water collector starts sampling, the motor drives the output shaft to drive the gear plate. The gear plate rotates counterclockwise. The trigger mechanism moves for the first time. The pinion gear A rotates 90° to open the outlet ball valve, and then the sampling displacement process is carried out. After the displacement is completed, the gear plate continues to rotate, and the trigger mechanism acts for the second time until pinion A and pinion B are about to mesh with the gear plate, which is a transitional stage. Then, the chainring continues to rotate counterclockwise to a certain angle, driving pinion B and pinion A to rotate 90°, so that both the inlet ball valve and the outlet ball valve are closed, and the entire operation process of such a sampling channel is completed. By analogy, the trigger process of multiple sampling channels can be completed.
c. Trigger mechanism drive unit
The rotary shaft of the driving motor needs to realize a reliable dynamic seal while outputting the rotary motion. At present, the more widely used sealing method on the dynamic seal is mainly the O-ring seal. To alleviate the problem of friction and heat generation between the rotating shaft and the O-ring, a PTFE slip ring is used on the basis of the O-ring seal.
According to the preliminary estimation of the trigger mechanism’s requirements to meet the output torque, rated speed, controllability, reliability and other factors, this article selected the 36GP-42H250B08 planetary reduction stepper motor based on the limited consideration of weight and volume. Synchronous motors generally need to be used with drives. To control the overall volume, a small drive that can be integrated with the motor is selected, and its model is 7TPSM4220.
The overall structure of the drive motor unit is shown in Fig. 12. The motor is fixed by the mounting plate and the upper end cover, and the motor outputs the power to the output shaft through the coupling.
4. Prototype and field test
a. Prototype
According to the results of numerical simulation, a capsule-shaped sampling cavity is selected. To improve the processability, the sampling cavity is designed as an outer cylinder and an inner capsule. The corner area of the cylinder is removed to reduce weight.
The assembled prototype is shown in Fig. 13. Since the prototype development is mainly based on principle verification, only one sampling channel was made.
b. Tightness test
The sealing performance test is carried out outside the pressure chamber, and the test method is internal pressure. The test platform is shown in Fig. 14, and the pressure medium is silicone oil. At the beginning of the test, it is necessary to inject silicone oil into the sampling channel. When the silicone oil starts to flow out of the outlet ball valve, close the outlet ball valve and continue to fill with silicone oil. When the pressure gauge shows that the pressure reaches about 2 MPa, perform an exhaust operation through the pressure test pump, and repeat until all the gas in the sampling channel has been exhausted. After the exhaust is completed, fill with silicone oil until the target pressure and keep it for a certain period of time to observe the sealing effect of the sampling channel.
The results show that when the pressure is 20 MPa, the pressure drop is less than 0.5 MPa within one hour, and the pressure drop is within 3%, indicating that the sampling channel has a sealing capacity of 20 MPa, which can better meet the requirements of kilometer-level operations.
c. Displacement performance test
When the displacement performance test is performed on the sampling channel, one end of the sampling channel is connected to the liquid inlet tank through a peristaltic pump, and one end is connected to the waste liquid tank. The setting of the experiment refers to the work of Fan et al. (2010).
The test platform of the displacement efficiency is shown in Fig. 15. At the selected time point, the inflow is stopped, the on–off valve at the inlet of the sampling cavity is closed, and then the sampling cavity is removed and the liquid in the cavity is drawn out, and fully stirred to make it evenly mixed. A salinometer is used to measure the concentration of the diluted solution to obtain the displacement rate.
The result of the test is shown in Fig. 16. The sampling channel can complete the sampling task well. The overall trend of the test is similar to the numerical simulation, but slightly slower than the simulated value. The reason that the size of the sampling access inlet is slightly smaller than the simulated value, resulting in a high Reynolds number, which aggravates the mixing of liquids, or because the inlet shape is irregular, the inflow is not completely straight, which aggravates the instability of the stratification. It is also possible that the inflow speed of the peristaltic pump has certain fluctuations, which aggravates the mixing of liquids.
5. Conclusions
-
When the flow rate at the inlet of the sampling channel is high, the displacement effect of the sampling cavity is less sensitive to factors such as the position of the inlet and outlet, the shape of the sampling cavity, and the inclination of the installation of the sampling cavity due to the fierce fluid movement.
-
However, at low speeds, various conditions have a significant impact on the displacement effect of the sampling cavity. When the inlet and outlet of the sampling cavity are installed in the center, obvious displacement blind areas appear in the upper half of the two sampling cavities. The inlet is installed below and the outlet is installed above, which can effectively improve the displacement effect at low speed.
-
Cylindrical and capsule sampling cavities with the same volume and diameter have almost the same displacement performance. However, considering that the side seams of cylindrical sampling cavities will cause the accumulation of particulates in water samples, capsule sampling cavities are better than cylindrical sampling cavities.
-
When installing the sampling cavity, setting a slight inclination angle helps to improve the displacement effect of the sampling cavity at low speeds.
-
If the water sample is a hot liquid with slow eruption, a cold spring, or a sample with a small branch area, a low sampling speed should be selected to avoid inhalation of impurities and at the same time have a higher sampling efficiency. When the sample subdivision area is large, a higher sampling speed can be used to shorten the time required for the sampling operation.
-
The development of this prototype has successfully optimized the triggering device of the traditional sampling chamber for the vertical arrangement of the water sampler, and has saved space. The sealing level of the sampling channel of the prototype can reach at least 20 MPa, and it has a strong sample storage capacity at the kilometer level. However, this prototype still needs to be improved in its ability to hold the sample pressure. Adding a pressure holding structure on the basis of the original design may further enhance its pressure holding capacity.
Acknowledgments.
We are grateful to Professor Jiawang Chen and his students for their help in the experiment. We would also like to thank Daohua Chen and Yinan Deng from Guangzhou Marine Geological Survey (GMGS) for their help, and the support of Key Special Project for Introduced Talents Team of Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou) (GML2019ZD0506) from Southern Marine Science and Engineering Guangdong Laboratory (Guangzhou), 511458, and Research on Integrated Technology of Deep-sea in-situ Filtration and Fidelity Sampling (2021Szvup013) from Shenzhen Research Institute of Dalian Maritime University.
Data availability statement.
No datasets were generated or analyzed during the current study.
REFERENCES
Case, D. H., A. Ijiri, Y. Morono, P. Tavormina, V. J. Orphan, and F. Inagaki, 2018: Aerobic and anaerobic methanotrophic communities associated with methane hydrates exposed on the seafloor: A high-pressure sampling and stable isotope-incubation experiment. Front. Microbiol., 8, 2569, https://doi.org/10.3389/fmicb.2017.02569.
Chen, Z., W. Yan, M. H. Chen, S. H. Wang, S. B. Xiao, J. Lu, and H. P. Yang, 2011: Advances in gas hydrate dissociation and fate of methane in marine sediment. Diqiu Kexue Jinzhan, 21, 394–400.
Cui, H. P., and Coauthors, 2019: Microbial diversity of two cold seep systems in gas hydrate-bearing sediments in the South China Sea. Mar. Environ. Res., 144, 230–239, https://doi.org/10.1016/j.marenvres.2019.01.009.
Fan, W., H. C. Pan, and Y. Chen, 2010: Mechanism of high-performance liquid-liquid displacement for deep-sea sensor chamber. J. Mech. Eng., 46, 143–149, https://doi.org/10.3901/JME.2010.04.143.
German, C. R., and E. T. Baker, 2004: On the global distribution of hydrothermal vent fields. Mid-Ocean Ridges: Hydrothermal Interactions between the Lithosphere and Oceans, Geophys. Monogr., Vol. 148, Amer. Geophys. Union, 245–266.
He, S., Y. Peng, Y. Jin, B. Wan, and G. Liu, 2020: Review and analysis of key techniques in marine sediment sampling. Chin. J. Mech. Eng., 33, 66, https://doi.org/10.1186/s10033-020-00480-0.
He, S., Y. Peng, Y. Jin, J. Yan, and B. Wan, 2021: Design and experimental study of a novel full-ocean-depth pressure-retaining sediment sampler. J. Atmos. Oceanic Technol., 38, 1715–1726, https://doi.org/10.1175/JTECH-D-20-0202.1.
Henry, P., and Coauthors, 1996: Fluid flow in and around a mud volcano field seaward of the Barbados accretionary wedge: Results from Manon cruise. J. Geophys. Res., 101, 20 297–20 323, https://doi.org/10.1029/96JB00953.
Huang, H., L. Huang, W. Ye, S. Wu, C. Yang, Y. Chen, and H. Wang, 2018: Optimizing preloading pressure of pre-charged gas for isobaric gas-tight hydrothermal samplers. J. Press. Vessel Technol., 140, 021201, https://doi.org/10.1115/1.4038901.
Lang, S. Q., and B. Benitez-Nelson, 2021: Hydrothermal Organic Geochemistry (HOG) sampler for deployment on deep-sea submersibles. Deep-Sea Res. I, 173, 103529, https://doi.org/10.1016/j.dsr.2021.103529.
Love, B., M. Lilley, D. Butterfield, E. Olson, and B. Larson, 2017: Rapid variations in fluid chemistry constrain hydrothermal phase separation at the Main Endeavour Field. Geochem. Geophys. Geosyst., 18, 531–543, https://doi.org/10.1002/2016GC006550.
Luo, M., J. Gieskes, L. Chen, X. Shi, and D. Chen, 2017: Provenances, distribution, and accumulation of organic matter in the southern Mariana Trench rim and slope: Implication for carbon cycle and burial in hadal trenches. Mar. Geol., 386, 98–106, https://doi.org/10.1016/j.margeo.2017.02.012.
Miyazaki, J., and Coauthors, 2017: WHATS-3: An improved flow-through multi-bottle fluid sampler for deep-sea geofluid research. Front. Earth Sci., 5, 45, https://doi.org/10.3389/feart.2017.00045.
Niskin, S. J., 1962: A water sampler for microbiological studies. Deep-Sea Res. Oceanogr. Abstr., 9, 501–503, https://doi.org/10.1016/0011-7471(62)90101-8.
Okamura, K., T. Noguchi, M. Hatta, M. Sunamura, T. Suzue, H. Kimoto, T. Fukuba, and T. Fujii, 2013: Development of a 128-channel multi-water-sampling system for underwater platforms and its application to chemical and biological monitoring. Methods Oceanogr., 8, 75–90, https://doi.org/10.1016/j.mio.2014.02.001.
Pelayo, A. M., S. Stein, and C. A. Stein, 1994: Estimation of oceanic hydrothermal heat flux from heat flow and depths of midocean ridge seismicity and magma chambers. Geophys. Res. Lett., 21, 713–716, https://doi.org/10.1029/94GL00395.
Peoples, L. M., M. Norenberg, D. Price, M. McGoldrick, M. Novotny, A. Bochdansky, and D. H. Bartlett, 2019: A full-ocean-depth rated modular lander and pressure-retaining sampler capable of collecting hadal-endemic microbes under in situ conditions. Deep-Sea Res. I, 143, 50–57, https://doi.org/10.1016/j.dsr.2018.11.010.
Saegusa, S., U. Tsunogai, F. Nakagawa, and S. Kaneko, 2006: Development of a multibottle gas-tight fluid sampler WHATS II for Japanese submersibles/ROVs. Geofluids, 6, 234–240, https://doi.org/10.1111/j.1468-8123.2006.00143.x.
Sheryll, R. P., 2009: New technology for uncontaminated and pressure controlled deep-sea sampling: Deep Ocean Benthic Sampler (DOBS). Ph.D. thesis, Stevens Institute of Technology, 611 pp.
Wang, X., S.-J. Wu, Z.-F. Fang, C.-J. Yang, and S. Wang, 2020: A pressure-tight sampler with flexible titanium bag for deep-sea hydrothermal fluid samples. J. Atmos. Oceanic Technol., 37, 2065–2073, https://doi.org/10.1175/JTECH-D-20-0017.1.
Wu, S.-J., C.-J. Yang, K. Ding, and C.-Y. Tan, 2015: A remotely operated serial sampler for collecting gas-tight fluid samples. China Ocean Eng., 29, 783–792, https://doi.org/10.1007/s13344-015-0055-6.
Wu, S.-J., S. Wang, and C.-J. Yang, 2018: Collection of gas-tight water samples from the bottom of the Challenger Deep. J. Atmos. Oceanic Technol., 35, 837–844, https://doi.org/10.1175/JTECH-D-17-0170.1.
Yu, J. J., M. X. Wang, B. Z. Liu, X. Yue, and C. L. Li, 2019: Gill symbionts of the cold-seep mussel Bathymodiolus platifrons: Composition, environmental dependency and immune control. Fish Shellfish Immunol., 86, 246–252, https://doi.org/10.1016/j.fsi.2018.11.041.
Zhai, S., X. Wang, and Y. U. Zenghui, 2006: Heat and mass flux estimation of modern seafloor hydrothermal activity. Acta Oceanol. Sin., 25, 43–51, https://doi.org10.1016/j.marchem.2005.09.003.
Zhu, H. N., 2018: Research on water harvester based on mobile phone remote control. M.S. thesis, Changchun University of Science and Technology, 91 pp.