14.1 Introduction

Cold seeps have been discovered since the 1980s (Paull et al. 1984) and have become a research focus because they play a significant role in the geochemical cycle as a mass and energy exchange path between the lithosphere, biosphere, and hydrosphere. Investigating the cold seep system is significantly beneficial to monitoring the marine environment and the element cycle in the deep ocean. Recovering and transporting fluid samples back to the laboratory for geochemical analysis has been a commonly used method for studying cold seep systems in recent decades (Du et al. 2020). However, the compositions in cold seep fluids, which are sensitive to ambient conditions (such as P, T, pH, and Eh), may decompose during the sample acquisition and recovery process (Lilley et al. 2003). In situ detection has become a growing trend for studying cold seep systems. Here, we briefly review the application of the in situ detection and seafloor observation of the Site F cold seep of the South China Sea.

14.2 Fine Topographic and Geomorphologic Detection

The Site F active cold seep is located southwest of Taiwan (Fig. 14.1). It is part of the passive continental margin and the leading edge of the southwestern Taiwan deformation zone (Liu et al. 2010). The ridge line is NW‒SE−trending, the water depth at the top of the ridge is approximately 1120 m, the longitudinal length along the ridge line is approximately 30 km, and the widest transverse part of the vertical ridge line is approximately 5 km. The topographic contours on the ridge are closed, and the slopes of the two wings are steep, with a maximum of more than 30° (Berndt et al. 2014). Typical seabed features of cold seep systems, such as large−scale exposed authigenic carbonate rocks, reduced sediments, chemosynthetic communities, and gas flares, have also been found (Lin et al. 2007).

Fig. 14.1
A contour map of site F located between Taiwan and the Chinese mainland. The map below it elaborates on the cold seep, with the central region plotting negative 1130 and the surrounding regions moving from negative 1180 meters and eventually plotting negative 1330 meters.

The location of the Site F active cold seep

According to the submarine topographic map formed by shipborne multibeam bathymetric data and the topographic profile close to the parallel ridge, Formosa Ridge is surrounded by submarine canyons on the N, W and E sides, making it a nearly independent positive topography. Formosa Ridge has two ridges, the north and the south ridges, with an interval of approximately 2 km. The water depths of the north and south ridges are approximately 1100 m and 1125 m, respectively. The seafloor topography tilts from north to south. On the topographic profile close to the vertical ridge line, the ridges and canyons are arranged parallel and alternately in the direction of NNW‒SSE, and the combination forms a “W” topographic feature. The height differences between the top of the ridge and the bottom of the surrounding canyon are 300−850 m, from NW to SE, and the height difference tends to increase at first and then decrease. On the same section of the vertical ridge line, the height difference in the southwest is greater than that in the northeast, and the maximum height difference appears between the top of the south ridge and the bottom of the southwest canyon, which is approximately 850 m. The width of the ridge is basically the same to the north of the top of the ridge in the south, i.e., approximately 5 km (Fig. 14.1), while in the south, it tends to become larger and then smaller, and the widest point is approximately 7 km. The slopes of the two wings of the ridge are similar and steep, the overall slope angle is more than 10°, and the slope of the southwest wing is larger than that of the northeast wing. Along the ridge from NW to SE, the slope of the two wings increases gradually, and the slope of the southwest wing reaches 34° in the 2D section. Sporadic secondary canyons are present on both wings of the ridge and are mainly concentrated on the west side of the north ridge top and the east side of the south ridge top. The secondary canyon converges into the submarine canyon at the foot of the two slopes, and there is no obvious submarine fan accumulation at the mouth of the canyon. Researchers speculate that the sediment transported by these secondary canyons is carried away by the turbidity current or undercurrent flowing through the main submarine canyon.

Wang et al. (2021) systematically identified and analyzed the seabed features of the cold seep area of Site F through the massive high−definition imaging data obtained by ROV geophysical data and a 3D laser scanning system. The results show that the seabed features of this cold seep area mainly include 6 types: cold seep vents with gas flares, carbonate rocks of different shapes, dense benthic chemosynthetic communities, mussel beds with low population densities, shells and/or shell debris areas, and reduced sediments. All the abovementioned seabed features are distributed in an approximately zoned pattern with gas plumes and benthic communities in the innermost part of Site F and reduced sediments in the outermost part (Wang et al. 2021).

A comprehensive analysis reveals that two main sets of dominant fluid transport channel systems have developed in the authigenic carbonate mound at Site F, including the large−scale fluid channel under the flourishing biological area in the south and the transport system composed of multiple dominant channels in the luxuriant biological region in the north (as shown in Fig. 14.2). This discovery reflects the interaction between the cold seep fluid transport channel and the sediment type: the location of the dominant cold seep fluid transport channel and the intensity of fluid activity control the biological distribution characteristics on the authigenic carbonate mound; the development of chemosynthetic communities reacts to the activity of seepages; and authigenic carbonate rocks can converge longitudinally and channel laterally to cold seep fluids.

Fig. 14.2
A conceptual model of the site F. It consists of regions with reduced sediments, mussel, and shell beds, fluid conduits with authigenic carbonate mounds, shell beds, flourishing communities, migration direction of cold seep fluids, authigenic carbonates, methane bubbles, gas hydrate, and gas flares.

A conceptual model of the cold seep fluid conduits in the shallow subsurface at Site F based on the distribution patterns of the seabed features

14.2.1 In Situ Detection and Findings

Since the discovery of cold seep systems on the seafloor in the 1980s, scientists have gradually realized that cold seep systems connect the oceanic lithosphere, marine hydrosphere and benthic biosphere. These systems are transport pathways for deep−sea matter and energy. The phenomenon of deep−sea seepage is of great significance to the reaction mechanism of biogeochemical processes and the interaction of different deep−sea layers.

The components of the fluids erupting from the cold seep were studied by sampling and then transported to the laboratory for analysis. However, changes in the temperature, pressure and redox conditions during the sampling process led to the escape or chemical reaction of the components in the fluid, resulting in the failure of the laboratory analysis results to obtain the real in situ concentration of the components in the fluid.

With the development of Raman spectroscopy, it has been gradually used for in situ detection in the deep ocean. The Monterey Bay Aquarium Research Institute (MBARI) claimed the lead in applying laser Raman spectroscopy to deep sea in situ detection and developed the first international deep ocean Raman in situ spectrometer (DORISS), which has been updated to DORISS II (Brewer et al. 2004; Zhang et al. 2012). The Ocean University of China (OUC) has developed a deep ocean Raman spectrometer with two excitation lasers (Du et al. 2015). The Institute of Oceanology, Chinese Academy of Sciences (IOCAS) has developed a deep ocean Raman spectrometer with serial insertion probes (Zhang et al. 2017b), and the Dalian Institute of Chemical Physics, Chinese Academy of Sciences (DICP) has developed a deep ocean Raman spectrometer with a ultraviolet laser. The Technical University of Berlin (TUB) has developed Raman spectrometers for field applications based on SERS technology (Schmidt et al. 2004). Depending on the system performance, the Raman spectrometers mentioned above have been used to detect different kinds of targets from shallow water to the deep sea.

14.2.2 In Situ Detection of Fluids in the Cold Seep Area

For most cold seep systems, the cold seep fluid is mainly composed of seawater and CH4 and may include liquefied fine−grained sediments and other components. There are two main sources of CH4 components in the fluid: biogenic gas and thermal degradation gas (Stakes et al. 1999; Judd and Hovland 2009). The former is the gas formed by the degradation of organic matter in shallow sediments with the participation of microorganisms, while the latter is the gas formed by thermal degradation of kerogen in deep strata under the action of microorganisms. Of course, the methane component in the fluid may also be a mixed product of the two genetic gases. For these two genetic gases, their CH4 content and carbon isotope characteristics are obviously different, so they can be used as an important index to assess the source of methane. The study of cold seep systems often begins with the analysis of various indicative components in fluids.

The RiP was used for in situ detection of the cold seep fluids at Site F in the South China Sea. By detecting the spectra of sediment pore water at different depths (as shown in Fig. 14.3), the concentration profiles of methane, sulfate and hydrogen sulfide in sediment pore water were obtained. A total of 22 profiles were acquired during past scientific expeditions. The results show that the sulfate−methane interface is shallow in the reduced sediment area, with sufficient methane supply, while the sulfate−methane interface is deeper in the common sediment area.

Fig. 14.3
A multi-line graph plots intensity versus Raman shift. It plots 7 curves with increasing to decreasing values. All curves plot stable values until 3000 delta centimeters superscript negative 1, after which all curves plot their peak values at 3500 delta centimeters superscript negative 1. 50 centimeters plot is the highest among the curves. 3 insert graphs that plot C H subscript 4 and H subscript 2 S.

The in situ Raman spectra of the sediment pore water at different depths

The cold seep fluids erupted from the active vent at Site F, and the fluids in the chemosynthetic communities near the active cold seep vent were also detected in situ using the RiP system (Du et al. 2018a). The in situ Raman spectra revealed that the presence of CH4 and C3H8 in the fluids erupted from the cold seep vent, consistent with the results of gas chromatography analysis. Finding C3H8 provides a new possible explanation for the carbon source of this cold seep area. The in situ Raman spectra of the fluids at different depths of the chemosynthetic communities show that the concentration of SO42− decreases with increasing depth, whereas the concentrations of CH4 and S8 increase in fluids in the chemosynthetic communities but without H2S (Fig. 14.4). These results indicate that CH4 has been oxidized by SO42− and that S8 is formed. This process usually occurs in reduced sediments as anaerobic oxidation of methane. Overall, the findings provide new insight into the geochemical analysis of cold seep vent fluids and in situ proof of the oxidation of methane in chemosynthetic communities near cold seeps.

Fig. 14.4
2 photos and 2 illustrations. Photographs a and c display a piece of machinery with a metal rod penetrating into the ocean floor. b, the sediments like C H subscript 4, C subscript 3 H subscript 8, H subscript 2 S below the ocean floor coming up through the cold seep vent. d, a layer of cold seep hemosynthetic communities between the seawater and sediments.

The in situ Raman quantitative detection and results of cold seep vents and fluids in chemosynthetic communities near an active cold seep vent in the South China Sea

14.2.3 In Situ Detection of the Solid Targets

14.2.3.1 Detection of Authigenic Carbonates

Raman spectroscopy shows great advantages in the in situ detection of deep−sea fluids, gases and gas hydrates and authigenic carbonates because the formation and evolution of authigenic carbonates are closely related to cold seep fluids. In situ detection of authigenic carbonate rocks in cold seep areas is helpful to accurately understand the formation environment of authigenic carbonate rocks and the interaction between fluids and rocks. The application of laser Raman spectroscopy to the in situ detection of deep−sea rocks can reveal preliminary information, such as the rock mineral composition and relative content, which makes our sampling work directional and purposeful. Previous studies on Raman spectra of carbonate minerals have shown that carbonate minerals are very sensitive to Raman scattering (Wang et al. 1995). White (2009) effectively distinguished the difference between calcite and aragonite in a deep−sea extreme environment (White 2009) based on Raman spectroscopy. The Raman spectra of biogenic carbonate minerals, such as calcite, dolomite and aragonite, have been systematically studied, laying the foundation for qualitative identification (Herman et al. 1987; Urmos et al. 1991; Buzgar and Apopei 2009). Raman spectroscopy is also used to study the mechanism of carbonate formation (Bonales et al. 2013). The Raman shift of carbonate can indicate the difference of the magnesium content in magnesium calcite (Borromeo et al. 2017). The above studies show that Raman spectroscopy can effectively reflect the carbonate minerals in authigenic carbonate rocks, which lays a foundation for deep−sea in situ exploration of authigenic carbonate rocks. The Raman shift of the main carbonate peak of Site F is 1084−1087 cm−1, which is three wavenumbers less than that of Site P (f the Pear River Mouth basin) (Fig. 14.5). The results show that there is a significant component difference between them, indicating that the difference in the element composition affects the Raman shift.

Fig. 14.5
A graph of intensity versus Raman Shift. The curves for Site F and Site P plot stable values parallel to the x-axis until delta centimeters negative 1 1080, after which both curves plot their respective peak values at 1086 and 1089 A.U for Sites F and P, respectively. They go back to plotting stable values immediately after.

The in situ Raman spectra of the authigenic carbonate acquired at the Formosa Ridge (Site F) and the eastern region of the Pear River Mouth basin (Site P)

In situ Raman detection of the authigenic carbonates and fluids at different locations with different fauna has been accomplished simultaneously. The in situ Raman quantitative analyses of the fluids in the dense chemosynthetic communities (total of 30 cm) show that the concentrations of Cl and SO42− decreased while CH4 increased with increasing depth (Fig. 14.6). On the other hand, the minerals at the different positions with different faunal densities also show significant differences. The in situ Raman spectra of the authigenic carbonates show that the full width at half maximum (FWHM) of the Raman peak of aragonite increased from the fauna−rich area to the desert edge area, suggesting that the degree of destruction of the aragonite increased due to the effects of the surrounding fluids.

Fig. 14.6
3 line graphs plot depth versus S O subscript 4 superscripts 2, C H subscript 4, and w t percentage N a C l. Graph a and c plots an increasing trend for Fauna rich areas while graph b plots a decreasing trend.

The in situ Raman quantitative analyses of the fluids from different locations. a Quantitative analysis of SO42− concentration in the fluid at different locations. The SO42− concentrations in seawater at different locations ranged from 28.5 to 30 mM. b Quantitative analysis of CH4 concentration in the fluid at different locations

The changes in the components suggested that complex biochemical processes occurred in the communities (Du et al. 2018b). Xi et al. proposed that one of the reactions may be the sulfate−driven anaerobic oxidation of methane (Xi et al. 2020). The reactions also contribute to the formation of authigenic carbonates and the regular aragonite structure in the shallow seafloor (Luff and Wallmann 2003):

$${CH}_{4}+S{O}_{4}^{2-}\to {HCO}_{3}^{-}+{HS}^{-}+{H}_{2}O$$
(14.1)
$$2{HCO}_{3}^{-}+C{a}^{2+}\to {CaCO}_{3}+{CO}_{2}+{H}_{2}O$$
(14.2)

Furthermore, compared with the reducing sediments in the seafloor, the high fluid perturbance in the chemosynthetic communities increased the dissolved oxygen content in the communities, which contributed to the oxidation of methane:

$$C{H}_{4}+2{O}_{2}\to C{O}_{2}+{2H}_{2}O$$
(14.3)

With the upward diffusion of the methane−rich fluid in the communities, the methane content showed a decreasing trend due to reactions (14.1) and (14.3). As an important reactant, SO42− is efficiently reduced in reaction (14.1) due to the addition of upward methane. Moreover, as the product of the above three reactions, the addition of H2O efficiently lowered the salinity of the fluid in the chemosynthetic communities, especially at the bottom of the communities. The fluid with low Cl and SO42− can inhibit the corrosion of authigenic carbonates to a great extent under long−term erosion (Xi et al. 2020).

Based on common features of the fluids and authigenic carbonates covered by chemosynthetic communities with different densities and the high−resolution videos of the cold seep system, Xi et al. (2020) proposed a detailed model of the system at Site F to study the evolution of a seepage system by the differences in the spatial distribution characteristics of the authigenic carbonate and overlying chemosynthetic communities.

In addition, combined with the advantages of quantitative analysis of Raman spectroscopy, the compositional changes of fluids and the Raman spectral characteristics of authigenic carbonate minerals in the corresponding fluids are quantitatively analyzed. This analysis not only provides effective data and a new understanding for the simultaneous analysis of the interaction between cold seep fluid and authigenic carbonate rocks but also expands the scope of laser Raman spectroscopy in deep−sea in situ detection.

14.2.3.2 Detection of Gas Hydrates

The properties of gas hydrate have been studied in the laboratory by simulations in past decades. The structures of the methane hydrate are transformed with changes in the pressure. The methane gas hydrate in natural fluid inclusions has been used to identify the formation conditions in fluid inclusions. With the development of submarine vehicles and in situ detection technologies, the deep sea has become a natural testing ground for studying the physical and chemical properties of natural gas hydrate (NGH). Raman spectroscopy has been increasingly applied for the in situ detection and analysis of natural gas hydrate (NGH). Zhang et al. (2017a) conducted the first in situ detection of gas hydrates exposed on the seafloor at Site F in the South China Sea using the Raman insertion probe for gas hydrate (Fig. 14.7). The results reveal that the hydrates are primarily Structure I hydrates but contain significant amounts of C3H8 and H2S, indicating that they may have originated from thermogenic sources. Free gas can exist in hydrate fabric during very rapid hydrate formation, and a new growth model for the early stages of crystalline gas hydrate formation was proposed based on the in situ data. Du et al. (2018b) conducted in situ detection of synthetic gas hydrate (SGH) that formed quickly by cold seep fluids erupting from active cold seeps at Formosa Ridge in the South China Sea. Authigenic carbonate debris, which is found in the in situ Raman spectra of SGH samples for the first time, may be one factor that can promote the formation of synthetic gas hydrate. The gas–water interface also contributes to the rapid formation of SGH samples. Time series Raman spectra of the four SGH samples demonstrate that methane large−to−small cage occupancy ratios of the hydrates vary from 1.01 to 1.39, and the ratios of methane large−to−small gas hydrates increase from 0.53 to 1.55. The results indicate that gas hydrate is continuously formed by the occluded gaseous CH4 after quick formation. The Raman spectra of the SGH sample, which was surprisingly found and recovered after 410 days on the seafloor, reveal that a methane−rich environment is a necessary condition for the stable existence of synthetic gas hydrate (Fig. 14.8).

Fig. 14.7
3 photos and a graph. 3 photos of a to c display the seafloor and highlight pieces of gas hydrate. d plots intensity versus Raman Shift. A line follows an increasing and decreasing fluctuating trends. 3 inserts of increasing portions of the line. 1, H subscript 2 O has the highest at 1630. 2, H subscript 2 S has the highest at 2585. 3, C H subscript 4 has the highest 2917.

The in situ optical and Raman observations of the gas hydrates exposed on the seafloor at the SCS−SGH02 site. a The exposed hydrate inside the channel of a cold seep vent surrounded by lush chemosynthetic communities. b A close−up view of the exposed hydrate attached to the inner wall of the channel. c The release of small pieces of gas hydrate from the vent during the Raman observations performed by the RiP−Gh probe, which is held by the arm of the ROV Faxian. d In situ Raman spectrum of the natural gas hydrate from 900 to 3,800 cm−1 at the SCS−SGH02 site

Fig. 14.8
2 diagram illustrates the how the S G H is formed, and presence of hydrate gas. a. It consists of a layer of S G H and seawater in a cell with an insert of single gas bubble. b has clump weight sandwiched between a layer of closed space and semi-open space. All 3 parts are called a push core sampler.

a The proposed model for the quick formation of SGH; b a schematic diagram of the preliminary model for the stable existence of gas hydrate

14.2.3.3 Detection of Biological Samples

With the aid of micro−techniques, Raman spectroscopy has been used for the in situ detection of different bacteria, which is the method of direct detection in the natural environment via an environment simulator without sample preparation (Kirschner et al. 2001). The distribution pattern of bacterial aggregates over a relatively large area can also be obtained by confocal Raman microscopy (CRM) because this technology has been established. With the advantages of high lateral and axial resolution, CRM has been regarded as a powerful tool to acquire noninvasive and in situ information on microbial communities (Pätzold et al. 2006). The microbial distribution of nitrifiers and anammox bacteria are scanned and mapped in situ based on the resonance Raman effect of cytochrome c (Cyt c), which may be a promising indicator for the metabolic condition of the bacteria (Pätzold et al. 2006). CRM is used to confirm the zero−valent sulfur compounds (S0) formed during anaerobic oxidation of methane (AOM) through a new pathway for dissimilatory sulfate reduction by methanotrophic archaea (Pasteris and Wopenka 2004; Milucka et al. 2012). CRM is also deployed to classify mussel foot (Liang et al. 2020) and mussel shells from cold seep areas (Cui et al. 2020). The application of Raman spectroscopy for biological samples provides new insight for real−time investigation of the life process and strategies of adaptation to an extreme environment of deep−sea macrofaunals. The symbiotic mechanism between metazoans and chemoautotrophs is an essential factor in maintaining the stability and sustainability of life activities in deep−sea ecosystems, and it has become an important issue in deep−sea research. However, there is no in situ detection method for deep−sea extracellular metabolism. Because of the influence of the Raman detection limit, it is difficult to detect deep−sea microbial metabolites or intermediates at low concentrations. Surface−enhanced Raman scattering (SERS) has been applied to enhance the Raman detection limit. A SERS substrate of coccinella septempunctata−like silver nanospheres coated with silver nanoisland structures (Ag NI@Ag NSs) was developed by a simple annealing process (Fig. 14.9), and the detection of a low concentration of the archaeal biomarker PETA was realized under a simulated deep−sea pressure of 1100 m (11 MPa) (Wang et al. 2022).

Fig. 14.9
A 3-part illustration. a. A S E M image of a ladybug with a g nano island. b. the schematic representation of a g nano island consisting of a g nanospheres and molecules on its surface along with a Raman signal. c plots intensity versus Raman shift. P E T A is highest in hydrothermal vent fluid, with specific regions highlighted.

a SEM diagram of the SERS base (the morphology was similar to the pattern of coccinella septempunctata in the illustration); b enhancement principle of the SERS base; c Raman spectra obtained from SERS detection in complex environments

14.3 In Situ Experiment and Long−Term Detection Using Mulit−RiPs

Fluid seepages in deep sea cold seep areas link the lithosphere, the hydrosphere and the seafloor biosphere. The strong inhomogeneity of cold seep systems in terms of temporal and spatial scales requires in situ, long−term and multitarget detection. Many sensors, including deep ocean Raman spectrometers, have been developed and used for in situ detection in cold seep areas. However, the limitation of detecting a single target presents a large challenge to obtaining a more comprehensive understanding of cold seep systems.

A novel long−term ocean observation platform (LOOP) and multichannel Raman insertion probes (Multi−RiPs) were developed and deployed at Site F of the SCS during 2020, 2021 and 2022 (Fig. 14.10). Using the deep−sea cold seep area as a natural test site, the designed in situ experiment was conducted continuously for a long time (Du et al. 2022). A new steerable mode for launching and recovery with the aid of the research vessel and submarine vehicles was used during the deployments of the LOOP at Site F in the SCS. The LOOP can be operated in an online real−time control mode, allowing landing site selection and adjustment of observation parameters during the launching process, with a subsequent switch to an offline stand−alone operation mode for long−term, continuous observation (Du et al. 2023). Thus, in situ experiments can be conducted with the aid of ROV and detected continuously in situ by Multi−RiPs.

Fig. 14.10
3 photographs of big machinery placed in the seafloor in 3 consecutive years between 2020 and 2022.

The Mulit−RiPs was deployed by LOOP in three consecutive years (a: 2020; b: 2021; c: 2022) at Site F for the long−term detection of deep−sea in situ experiments

A calibration switching module (CSM) is used to connect the light path between the laser, the Raman insertion probe and the spectrometer (Fig. 14.11). Thus, Raman probes can share core components (the laser, spectrometer, etc.) that are sequentially controlled by the CSM and detect the specified targets in situ according to the preprogrammed schedule. Combined with the developed series of Raman insertion probes, the Multi−RiPs can simultaneously detect multiple targets with different phase states, such as fluid, solid and gas, in deep−sea cold seep systems.

Fig. 14.11
A block diagram of the multi-RiPs model. It includes power originating from the control and power module to the spectrograph, C C D, laser, fiber adapter, and the C S M. The C S M consists of a calibration switch module that starts from the laser through the laser shuttle and into an optic switch. It then branches into probes 1 through 4.

The layout diagram of key optical components and the optical switching module of the Multi−RiPs (modified from Du et al. 2022)

By analyzing the continuous long−term spectra obtained by two Raman probes placed near the cold seep vents at different depths in the chemosynthetic communities, it was found that the most significant difference at different depths in the chemosynthetic communities was the difference in elemental sulfur: elemental sulfur was not detected at the middle depth in the chemosynthetic communities, but more elemental sulfur was found at the bottom of the chemosynthetic communities.

The continuous long−term Raman spectra obtained from the in situ synthesized gas hydrate show that after the quick formation of gas hydrate, enough gaseous methane with a closed honeycomb structure causes the hydrate formation process to continue dominating the dynamic process of the hydrate system. At the same time, it is also confirmed that the stable existence of natural gas hydrate requires a methane−rich environment in addition to the stable pressure and temperature field of natural gas hydrate.

Gas flows are probably transient events. At certain times, gas flows through the sediments and eventually into the bottom water. Along the way, some gas forms hydrates along fractures and in weakening horizontal zones within the sediments. While the gas flow is episodic, the hydrates act as dynamic “capacitors” from which methane is constantly diffusing into the direction of the methane−depleted bottom water, thereby creating stable chemical gradients for microbes carrying out anaerobic oxidation of methane and, subsequently, the occurrence of chemosynthetic organisms. By placing a Raman cold−seep eruption fluid collection device attached above the cold−seep vent, natural gas hydrate, or free−form methane gas, can be captured by the collection device, and when the cold seep is not erupting, the seawater is captured by the device. The time series Raman spectra confirmed that the cold seeps are intermittently erupting, which is consistent with previous studies. The results also indicate that the gas hydrate at shallow sediment depths and exposed on the seafloor acting as a capacitor dominate the biogeochemical processes in the cold seep area.

14.4 Summary and Perspectives

The substance brought by seepages from the deep earth has a complicated composition, and the concentrations of the targets in the fluid plume even a few centimeters away from the cold seep vent decrease sharply. They have become the two barriers that limit laser spectroscopy from having broader applications in the in situ detection of seepages in the deep ocean. One limitation is the low detection sensitivity of laser spectroscopy to the targets. To improve the detection sensitivity of conventional Raman spectroscopy, Raman signal enhancement methods based on multiple reflection, liquid core optical fiber, resonance enhancement and surface enhancement have been proposed and used.

The other limitation of laser spectroscopy is that one kind of spectroscopy may be sensitive to certain types of targets (such as molecules for laser Raman spectroscopy (LRS) and atoms or elements for laser−induced breakdown spectroscopy (LIBS)). Therefore, sufficient and effective information cannot be obtained by a single measurement. To solve the problem of insufficient detection ability, combining different kinds of in situ detection technologies to achieve effective detection for different types of targets (such as acid ions and metal cations) has been proposed in past decades.

To further expand the application field of laser spectroscopy for in situ detection, reliable LRS and LIBS signal enhancement methods suitable for deep sea conditions and quantitative analysis methods with high accuracy for underwater LIBS need to be developed. Underwater multispectral in situ integrated detection systems have become a tendency for in situ detection.

Another trend for more applications of in situ laser spectroscopy is the long−time and real−time detection deployed in submarine observation networks or underwater observatory landers for seafloor ecosystems and geological systems. These two directions are also the main development trends of laser spectroscopy in the field of deep−sea in situ detection.