Marine environmental problems have attracted increasing attention in recent decades [1
]. Petroleum products play an important role in modern society as a source of energy and chemical feedstock, as a result, oil spills inevitably occur during the production, use, transport and storage of petroleum products. One of the most complete studies of the spread and impact of oil spills on human activity, environmentally sensitive shorelines and offshore regions is the research of Alves et al. on the Mediterranean Basin [3
]. In particular, oil production platforms and oil storage tanks in coastal areas are major sources of oil spills, which can affect various aspects of daily human activities [6
]. Therefore, routine surveillance in major ports is important for the early warning, prevention and control of oil spills. There are various modelling techniques applied to simulate the behavior of oil spills. Alves et al. modelled the behavior of oil from spills located at various depths below the sea surface [8
]. In addition, excellent laboratory and remote sensing techniques exist for the remote detection of oil spills [10
]. For such applications, techniques and sensors are required that permit high-frequency monitoring in the field, at a low cost per analysis. For example, Chase et al. have developed an oil spill detection and alarm system that can detect trace oil and micron-thin oil film on the surface of seawater, in real time [15
]. It is difficult to achieve stable and low-cost measurements when continuous monitoring is desired. Further, producing sensor with anti-corrosion properties and a low explosive risk is challenging, given the hazardous nature of old oil storage tanks.
Optical measurement techniques are well suited to this purpose, since they permit continuous surveillance [16
]. Excitation-emission matrix spectroscopy (EEMS) is an effective tool for the detection and quantification of oils on the surface of seawater [17
]. In EEMS, ultraviolet (UV) lasers are used as the excitation source in the spectral region between 240 and 355 nm and emission spectra are collected in the wavelength region between 300 and 500 nm. EEMS require large, specialized laboratory equipment to acquire the fluorescence spectra, which requires a dedicated power source. Thus, EEMS are not practical for many field applications. Among the various optical principles available for sensing, UV-induced fluorescence methods have an advantage in that almost all oil types have characteristic fluorescence properties and fluorescence can be used to obtain valuable information required to detect oil on various backgrounds, including seawater, soil, ice and snow [18
]. For example, fluorescence spectra and light absorption properties have been obtained by UV-induced fluorescence methods to detect trace oil on seawater. These advantages have led to the development of a variety of laboratory and field instruments [20
Fiber optical sensors can be used with UV lasers for the fluorescence detection of organic pollutants in water, especially oil products [22
]. A multi-channel receiver, fluorescence LiDAR, or photomultiplier tube (PMT) can be used to detect and record the emitted fluorescence spectrum. When combined with fiber optics, in situ applications and long-distance surveillance systems are feasible [25
]. However, for short wavelengths in the UV spectral range, laser transmission is limited by increased fiber attenuation in the fiber material. The transmission quality will decrease when intense UV light propagates through the commonly used fused silica fibers [26
]. In addition, fiber optics need more maintenance for day and night surveillance. These factors set limitations on the long-distance detection and monitoring of oil spills.
In this study, we considered the UV-induced fluorescence of both clean and oil-bearing seawater to examine the efficacy of applying the coastal-mounted fluorescence filter system-based (FFS-based) method for the detection and monitoring of oil spills on the surface of seawater. Our goal was to develop a compact, sensor-based instrument with day-long operability that could be mounted at a coastal site such as a harbor or port. This sensor-based instrument complements ‘late’ spill modelling, once the oil spill has reached the shore or is very close to it. This paper describes the design of the coastal-mounted FFS instrument and the experiments to evaluate its performance.
2. Design and Implementation of FFS-Based Coastal-Mounted Sensor
presents schematic diagrams of the coastal-mounted sensor for monitoring oil spills. The continuous wave source is a xenon lamp (200–300 nm, L4634-01 synthetic silica glass, Hamamatsu Photonics, Hamamatsu, Japan) with a dedicated power source (C13315, Hamamatsu Photonics). For this study, we chose a PMT (H10723-110, Hamamatsu Photonics) as the real-time detection unit. The sensor was also required to meet a number of constraints, including a compact size, low power consumption and independence from outside water and power requirements. These subassemblies are compact and integrated within a stainless-steel case (roughly 45 × 40 × 20 cm) which has anti-corrosion and anti-explosion properties. The coastal-mounted sensor can detect micron-thin oil films in the laboratory and can operate at range in excess of 5-m above the seawater surface.
The result of experiments in initial design and upgrade of sensor indicated that the xenon lamp and PMT can be proved to be highly effective for detection of micron-thin oil film (the minimum oil film thickness is 1.0 μm) from a distance of 5-m above the target surface area. The key limitation for detection range was that the excitation intensity of xenon lamp is whether enough to enable oil film to be detected by the PMT. Besides, this 5-m limit is the approximate upper bound for reliable detection. The coastal-mounted sensor can achieve reliable detection once per second (maximum frequency) determined by the sensor’s control unit. Wireless transmission module is required for the coastal-mounted sensor, which has been designed to use a basic RS232/RS485 protocol for integration with industrial process control systems. The main parameters of the coastal-mounted sensor are showing in Table 1
Previous studies on oil-on-seawater monitoring showed that fluorescence signals were affected by solar radiation and the choice of continuous wave source. These fluorescence signals were unable to detect trace oil and micron-thin oil films, unless the signal was filtered and processed [27
]. In this study, a FFS was designed to improve the accuracy and reduce noise effects in the coastal-mounted sensor (Figure 1
, No. 3). The FFS is composed of two-stage band-pass optical filters (300–400 nm) and a convex lens, as shown in Figure 2
Selection of the 300–400 nm band-pass filter was made considering the impacts of solar radiation, as well as the fact that most oils have a fluorescence peak within this wavelength band. Further, the reflection coefficient of the ocean surface to the light in this band is less than that for the 400–600 nm band [29
]. The first-stage optical filter is larger in size to expand the viewing angle. The convex lens is used to concentrate the fluorescence signals and thus converge on a single point in the center of the second optical filter. For detection, a spectrograph’s (laboratory) or PMT’s (field) fiber optics are connected to the center of the second optical filter (Figure 2
a). To ensure the FFS receives fluorescence that is not affected by the continuous wave source, it is necessary to install a band-stop filter (300–400 nm) at the continuous wave source (Figure 2
3.1. Seawater and Oil Samples
Seawater samples were collected in 80 mL glass bottles in September 2017 at the end of the largest berth in the port of Lingshui, China, on the northern Yellow Sea. Considering the possible presence of fluorescent substances (e.g., phytoplankton, algae) in clean seawater, enough seawater was collected to obtain background values.
Three main types of oils must be considered in studies of petroleum transport and storage: light fuel oil (diesel), heavy fuel oil and crude oil. We used −10# diesel oil, 0# diesel oil; 180# fuel oil, 380# fuel oil; Saudi crude oil and Brazilian crude oil, to represent these different oil types (Table 2
To simulate an oil spill on seawater, a small amount of each oil type was individually placed into six 40 mm × 25 mm glass bottles, along with 5 mL seawater, to form a 5
thick oil film (Figure 3
). The amount of each oil was calculated using:
is the diameter of the vial (40 mm) and
is the desired computational thickness of the oil film (5 μm). These parameters afford the requisite oil volume of 6.28 μL. Bottles and lids with ground glass joints were selected to prevent oil volatilization. Additionally, a 40 mm × 25 mm glass bottle containing only seawater was used to determine the background fluorescence value for clean seawater.
3.2. Measurement and Apparatus
Both laboratory and field experiments were conducted. In the laboratory, the sensor’s FFS was connected to the fiber-optics of a 163-mm focal length Czerny-Turner spectrograph (Shamrock 163 Andor Technology, Northern Ireland, UK) and fluorescence was detected over a spectral range of 180–850 nm with a resolution of 5 nm (laboratory-mounted sensor). Ambient natural light was used as the environmental background in order to simulate the oil spill detection environment of the coastal areas in the laboratory. The detection distance was set at 1 m and all measurements were conducted under the same conditions.
For the field experiments, the sensor’ FFS was connected to a PMT (H10723-110, Hamamatsu Photonics) before installation in Lingshui Port, China (coastal-mounted sensor). This sensor for monitoring oil spills was installed on the bank of the largest berth at a height of 1 m from the highest water line. Then, the coastal-mounted sensor was employed in simulated oil spill experiment as well as a continuous operation trial. The monitoring data were stored and analyzed.
In this study, we used a continuous wavelength (200–300 nm) source with a xenon lamp and 300–400 nm band-pass filters to determine the UV-induced fluorescence spectra of clean sea water and six oil samples, representing three characteristic oil types. The RFIs were determined using the FFS-based coastal-mounted sensor and spectrograph from 180–300 nm (fluorescence maximum for diesel oils) and from 300–400 nm (fluorescence maximum for heavy fuel and crude oils). The RFI values for all oil types were substantially higher than for clean sea water, demonstrating that the FFS-based coastal-mounted sensor using the RFI can be effective for the daily routine monitoring of surface oil spills along coasts and in harbors, even when different types of oil are present. Furthermore, based on the FFS and PMT, the coastal-mounted sensor can effectively inhibit background light interference. Even in environments with strong solar radiation, the sensor accurately differentiated between oil spills and the clean sea water surfaces.
In addition to monitoring oil spills in harbor areas, this effective system could be used to continuously monitor reservoirs, drains, sewage tanks and other aquatic environments where oil products may leak, overcoming the limitations of traditional monitoring technologies.