Detecting the Presence of Different Types of Oil in Seawater Using a Fluorometric Index

This study analyzed the fluorometric laboratory tests for the detection of the presence of oil in seawater in cases corresponding to the real situation in the sea: when the point of seawater sampling is not in the same place as the appropriate sensor. The phenomenon of fluorescence exhibited by both natural and alien substances (oil) in the sea was used. The possibility of oil detection in the water column based on a fluorometric index (FI) extracted from the excitation–emission matrix (EEM) was studied. Laboratory tests were carried out on water taken from the Gulf of Gdańsk (Baltic Sea). Seawater samples were contaminated with small amounts of various types of oil (the lowest oil-to-water ratio was 0.5 × 10−6). A statistically significant difference was found between FI values for uncontaminated seawater and seawater exposed to various kinds of oil (i.e., crude oils, lubricant oils, and fuels).


Introduction
Oil pollution is an ongoing problem for the marine environment. However, this issue was significantly reduced as a result of the MARPOL Convention [1]: tanker construction was improved, strict standards for oil discharges were established, a legal system of deterioration was developed, and methods for the early detection of oils on the sea surface were improved. Instrumental analytical methods for detecting the perpetrators of spills were also regulated [2]. The technical and organizational basis for combating spillages has significantly improved [3,4]. However, oil spills still appear [5]. Currently, in the scientific literature, the issues of remote detection of oil stains appearing at sea are widely discussed [6][7][8]. Both optical and radar methods [3,[9][10][11][12][13][14] are well developed in this field. Activities aimed at limiting the share of oil substances polluting the marine environment also include the possibility of early detection before they spread in the sea. This is particularly difficult if these substances exist in the depths of the sea in a dispersed form or they penetrate deeply into the environment beneath the surface of the water from tanks, pipelines, wrecks, mining equipment, or bottom seepage. In cases when the oil has spread in the sea and flows deeper into the sea under the influence of the wind and waves, different methods based on UV-visible, IR, and Raman spectroscopy, or those based on oil fluorescence [15][16][17][18], are required for the detection of oil substances. Moreover, for underwater oil pollution detection, different sensors can be used [19][20][21][22][23].
Both seawater and some hydrocarbon substances leaching into the water from the oil are capable of fluorescence induced by ultraviolet light. If seawater did not exhibit fluorescence, it would be easy to detect substances derived from oil. However, since oil substances and substances naturally occurring in the sea fluoresce similarly [18], a universal method for interpreting the fluorescence spectra in terms of detecting the presence of oil in seawater is required. Preliminary studies on seawater samples contaminated with only one type of oil (crude oil Petrobaltic), but sampled in various terms [24], have shown the possibility of using a numerical indicator extracted from the excitation-emission matrix (EEM) spectrum, whose value indicates the lack of exposure of water to oil pollution or an oil hazard in the vicinity of the point of water sampling or a possible sensor location. This paper outlines the results of a wider analysis of the possibility of signaling the presence of various types of oil in seawater, using a similar procedure for the interpretation of EEMs in all cases. The results presented in this paper are promising, but in order to finally implement the proposed method in real conditions and further in various sea regions, additional verification analyses are needed.

Seawater Samples
Seawater samples were collected in the Orłowo pier in Gdynia (Southern Baltic Sea, Poland). Seawater was sampled from a 1 m depth into 1 L glass bottles, three times in 2019: March (T = 4 • C and salinity 7.4 PSU), April (T = 6 • C and salinity 7.6 PSU), and May (T = 10 • C and salinity 7.5 PSU).

Oil Samples
Several kinds of oils were used for the laboratory contamination of seawater.
Flotta-extracted from the North Sea (Orkney), medium crude with API gravity 35.4 • and sulfur content 1.22%.
Gullfaks-extracted from the North Sea (off-shore), light crude with API gravity 37.5 • and sulfur content 0.22%.
(2) Lubricant oils: Marinol 1240-as a representative of lubricant oil used for the lubrication of inverter-type marine engines operating on light fuel. The oil base contains less than 3% polycyclic aromatic hydrocarbons (PCA).
Cyliten 460-as a representative of lubricant oil used in marine ship engine systems and used for the lubrication of single-and multi-cylinder reciprocating compressors of synthesis gas. The oil base contains less than 3% PCA.

Exposure of Seawater Samples to Oil
For laboratory measurements, seawater samples were contaminated by each oil. Small amounts of oil were placed on a slice of aluminum foil and then weighed and inserted into a seawater sample ( Figure 1) to reach the desired oil-to-water ratio. Four samples polluted by each kind of oil with an oil-to-water ratio (r o/w ) in the range of 0.5 × 10 −6 to 500 × 10 −6 were prepared. Natural seawater was exposed to the added oil for one day.

Measurement and Apparatus
Instrument A Hitachi F-7000 FL spectrofluorometer was used to determine the EEMs using a 1 × 1 cm quartz cuvette.

Measurement Parameters
The excitation wavelength was changed from 200 to 480 nm with an excitation sampling interval of 5 nm. The emission wavelength was changed from 260 to 600 nm with a 5 nm emission sampling interval, a 10 nm excitation slit, and a 10 nm emission slit. The integration time was 0.5 s, and the photomultiplier tube voltage was 400 V.

Measurement of Seawater and Oil-Contaminated Seawater Samples
1. The EEMs for samples of environment seawater were previously measured three times, allowing for a reliable, artefact-free background to be obtained. 2. Next, the EEMs of seawater exposed to each kind of oil for various ro/w were measured. 3. Measurements for all samples were performed at a stabilized temperature of 20 °C. 4. Rayleigh scattering to yield a digital matrix of EEMs was removed (if the excitation wavelength and emission wavelengths were equal and the emission wavelength was two times higher than the excitation wavelength).

Fluorometric Index (FI) Calculations
The studies took into account 84 EEMs for artificially contaminated samples of seawater and nine EEMS for uncontaminated samples ( Figure 2). Based on the EEMs of seawater samples and seawater contaminated by different kind of oils, an FI was proposed for oil detection. The proposed FI was extracted from the EEMs and used only specific values of fluorescence intensity described by the excitation wavelength maximum (λEx) corresponding to the emission wavelength maximum (λEm) for selected wavelengths for natural seawater and seawater polluted by oil. The FI was defined as a quotient of the fluorescence intensity at the emission wavelength for seawater polluted by oil to the intensity at the emission wavelength for natural seawater corresponding to the detected excitation maximum for both natural seawater and seawater polluted by oil described as FIo/w by Formula (1). This definition of FIo/w indicates the growth of its value in the presence of oil in water.
The effectiveness of oil detection in seawater based on the proposed indicator FIo/w was checked, taking into account the kinds of oils (Section 2) and oil-to-water ratios.

Measurement and Apparatus
Instrument A Hitachi F-7000 FL spectrofluorometer was used to determine the EEMs using a 1 × 1 cm quartz cuvette.

Measurement Parameters
The excitation wavelength was changed from 200 to 480 nm with an excitation sampling interval of 5 nm. The emission wavelength was changed from 260 to 600 nm with a 5 nm emission sampling interval, a 10 nm excitation slit, and a 10 nm emission slit. The integration time was 0.5 s, and the photomultiplier tube voltage was 400 V.

1.
The EEMs for samples of environment seawater were previously measured three times, allowing for a reliable, artefact-free background to be obtained.

2.
Next, the EEMs of seawater exposed to each kind of oil for various r o/w were measured.

3.
Measurements for all samples were performed at a stabilized temperature of 20 • C.

4.
Rayleigh scattering to yield a digital matrix of EEMs was removed (if the excitation wavelength and emission wavelengths were equal and the emission wavelength was two times higher than the excitation wavelength).

Fluorometric Index (FI) Calculations
The studies took into account 84 EEMs for artificially contaminated samples of seawater and nine EEMS for uncontaminated samples ( Figure 2). Based on the EEMs of seawater samples and seawater contaminated by different kind of oils, an FI was proposed for oil detection. The proposed FI was extracted from the EEMs and used only specific values of fluorescence intensity described by the excitation wavelength maximum (λEx) corresponding to the emission wavelength maximum (λEm) for selected wavelengths for natural seawater and seawater polluted by oil. The FI was defined as a quotient of the fluorescence intensity at the emission wavelength for seawater polluted by oil to the intensity at the emission wavelength for natural seawater corresponding to the detected excitation maximum for both natural seawater and seawater polluted by oil described as The effectiveness of oil detection in seawater based on the proposed indicator FI o/w was checked, taking into account the kinds of oils (Section 2) and oil-to-water ratios.
For oil detection in seawater, it is necessary to take into account the presence of natural seawater components. Therefore, to obtain information on the presence of natural seawater components, the EEMs of natural seawater were determined three times in 2019. Next, the influence of oil presence in seawater on the EEMs was analyzed. Figure 2 presents EEMs for natural seawater (no oil) and seawater polluted by different types of oils selected for an oil-to-water ratio of 50 × 10 −6 on 15 March 2019. The EEMs of natural seawater (no oil) in Figure 2 indicate the presence of the main peak in the UV-range centered at 225 nm of an excitation wavelength and corresponding to an emission wavelength centered at 360 nm (λ Ex /λ Em = 225/360) ( Table 1). The detected peak is well linked to the tryptophan-like seawater component. Figure 2 clearly presents the influence of oil on the EEMs. In the EEMs of seawater polluted by oil, for all oils, the main peak is centered at 340 nm for an emission wavelength corresponding to an excitation wavelength of 225 nm (λ Ex /λ Em = 225/340) ( Table 1) with an accuracy of ±5 nm. Moreover, for all types of oils, a second peak was centered at 330 nm for an emission wavelength corresponding to an excitation wavelength of 275 nm (λ Ex /λ Em = 275/330) ( Table 1) with an accuracy of ±5 nm, but with a lower fluorescence intensity than the main peak at 225/340. The results of the determined peaks for the EEMs for individual oils are presented in Table 1, and it is visible that, for the selected oil in the EEMs, other specific peaks depending on the type of oil and oil-to-water ratio are determined. The EEMs for all considered oils ( Figure 2) revealed bands of fluorescence intensity between 300 and 400 nm, with the main band at 340 nm. These bands are probably caused by the presence of fluorophores, which are usually centered in one-to three-ring aromatic structures in the petroleum substances [25][26][27][28] For oil detection in seawater, it is necessary to take into account the presence of natural seawater components. Therefore, to obtain information on the presence of natural seawater components, the EEMs of natural seawater were determined three times in 2019. Next, the influence of oil presence in seawater on the EEMs was analyzed. Figure 2 presents EEMs for natural seawater (no oil) and seawater polluted by different types of oils selected for an oil-to-water ratio of 50 × 10 -6 on March 15, 2019. The EEMs of natural seawater (no oil) in Figure 2 indicate the presence of the main peak in the UV-range centered at 225 nm of an excitation wavelength and corresponding to an emission wavelength centered at 360 nm (λEx/λEm = 225/360) ( Table 1). The detected peak is well linked to the tryptophan-like seawater component. Figure 2 clearly presents the influence of oil on the EEMs. In the EEMs of seawater polluted by oil, for all oils, the main peak is centered at 340 nm for an emission wavelength corresponding to an excitation wavelength of 225 nm (λEx/λEm = 225/340) ( Table 1) with an accuracy of ±5 nm. Moreover, for all types of oils, a second peak was centered at 330 nm for an emission wavelength corresponding to an excitation wavelength of 275 nm (λEx/λEm = 275/330) ( Table 1) with an accuracy of ±5 nm, but with a lower fluorescence intensity than the main peak at 225/340. The results of the determined peaks for the EEMs for individual oils are presented in Table 1, and it is visible that, for the selected oil in the EEMs, other specific peaks depending on the type of oil and oil-to-water ratio are determined. The EEMs for all considered oils ( Figure 2) revealed bands of fluorescence intensity between 300 and 400 nm, with the main band at 340 nm. These bands are probably caused by the presence of fluorophores, which are usually centered in one-to three-ring aromatic structures in the petroleum substances [25][26][27][28]  Taking into account the determined EEM peaks for all oils, it is clear that the peak for natural seawater (λEx / λEm = 225/360) and the main peak for seawater polluted by oil (λEx/λEm = 225/340) were detected at the same excitation wavelength. However, taking into account the peaks for the EEMs determined previously for natural seawater from more than a dozen tests in 2017 [24], the peak for natural seawater is dominant for an emission wavelength at 355 nm (λEx/λEm = 225/355). This is a reason to choose the selected wavelengths for FIo/w definition. Therefore, FIo/w was calculated as the Taking into account the determined EEM peaks for all oils, it is clear that the peak for natural seawater (λ Ex /λ Em = 225/360) and the main peak for seawater polluted by oil (λ Ex /λ Em = 225/340) were detected at the same excitation wavelength. However, taking into account the peaks for the EEMs determined previously for natural seawater from more than a dozen tests in 2017 [24], the peak for natural seawater is dominant for an emission wavelength at 355 nm (λ Ex /λ Em = 225/355). This is a reason to choose the selected wavelengths for FI o/w definition. Therefore, FI o/w was calculated as the quotient of the fluorescence intensity at a 340 nm emission wavelength and an intensity at 355 nm, while the excitation wavelength remained equal to 225 nm (Formula (2)): where I(λ Em ) describes, respectively, the fluorescence intensity corresponding to the emission wavelength for polluted seawater (340 nm) and natural seawater (355 nm) linked to the same excitation wavelength (λ Ex ) for both kinds of seawater (225 nm). Table 1. Major fluorescent peaks of natural seawater and the same samples polluted by oil at oil-to-water ratios of 0.5 × 10 −6 , 5 × 10 −6 , 50 × 10 −6 , and 500 × 10 −6 with their wavelength-independent maxima (for 15 March 2019). for polluted seawater achieved higher values (above 1) than for natural seawater (below 1). This is confirmation of the good choice of wavelengths for FI o/w . Moreover, in Tables 3-5, it is clearly visible that FI o/w values depend on the type of oil and are stabilized for an oil-to-water ratio above 50 × 10 −6 .    Figure 3 presents FI o/w for unpolluted seawater and seawater polluted with various kinds of oil for different oil-to-water ratios ranging from 0.5 × 10 −6 to 500 × 10 −6 . A distinct increase in FI o/w already exists for r o/w = 0.5 × 10 −6 . Beginning with r o/w = 50 × 10 −6 , the FI o/w value ceases to depend on r o/w . The deviation of FI o/w values from other oils was observed for Marinol and Cyliten. This is probably caused by the low content of PCA. . Figure 3. Fluorometric index (FIo/w) for unpolluted seawater and seawater artificially polluted with various kinds of oil for oil-to-water ratios of 0.5 × 10 −6 , 5 × 10 −6 , 50 × 10 −6 , and 500 × 10 −6 .

Statistical Analysis of the Fluorometric Index
Statistical calculations were performed to check the deviation of determined fluorometric index (FIo/w) values for three different times of sampling for natural seawater and the same seawater polluted by oil. Confidence intervals (CI) of the mean of the FIo/w values were performed using a Student's t-test at a significance level of 0.05 for each kind of oil for three water sampling terms. Figure 4 presents the fluorometric index (FIo/w) and CI of the mean for natural seawater (no oil) and seawater artificially polluted by different kinds of oil with oil-to-water ratios of 0.5 × 10 −6 , 5 × 10 −6 , 50 × 10 −6 , and 500 × 10 −6 . The FIo/w values for different kinds of oils and different oil-to-water ratios are statistically different in several cases. However, the confidence bars of contaminated water and clean water did not overlap.

Statistical Analysis of the Fluorometric Index
Statistical calculations were performed to check the deviation of determined fluorometric index (FI o/w ) values for three different times of sampling for natural seawater and the same seawater polluted by oil. Confidence intervals (CI) of the mean of the FI o/w values were performed using a Student's t-test at a significance level of 0.05 for each kind of oil for three water sampling terms. Figure 4 presents the fluorometric index (FI o/w ) and CI of the mean for natural seawater (no oil) and seawater artificially polluted by different kinds of oil with oil-to-water ratios of 0.5 × 10 −6 , 5 × 10 −6 , 50 × 10 −6 , and 500 × 10 −6 . The FI o/w values for different kinds of oils and different oil-to-water ratios are statistically different in several cases. However, the confidence bars of contaminated water and clean water did not overlap. Next, the standard deviation of the mean values of FIo/w for all kinds of oil for 21 samples from three different times of seawater sampling were determined. Figure 5 presents the average standard deviation of the mean calculated FIo/w values for natural seawater and artificially polluted seawater for different oil-to-water ratios. The highest value of the standard deviation of FIo/w value was found for the 500 × 10 −6 oil-to-water ratio. However, the standard deviation of the average FIo/w value for water exposed to oil does not overlap with the standard deviation of the average FIo/w value for uncontaminated water. Moreover, the coefficient of variations of FIo/w achieved values of approximately 10% for the oil-to-water ratios 0.5 × 10 −6 and 50 × 10 −6 , 14% for the oil-to-water ratio 50 × 10 −6 , and 19% for the oil-to-water ratio 500 × 10 −6 . This indicates that the diversity of FIo/w for all oil samples within a given oil-to-water ratio is small, and the FIo/w values indicate high similarity. Next, the standard deviation of the mean values of FI o/w for all kinds of oil for 21 samples from three different times of seawater sampling were determined. Figure 5 presents the average standard deviation of the mean calculated FI o/w values for natural seawater and artificially polluted seawater for different oil-to-water ratios. The highest value of the standard deviation of FI o/w value was found for the 500 × 10 −6 oil-to-water ratio. However, the standard deviation of the average FI o/w value for water exposed to oil does not overlap with the standard deviation of the average FI o/w value for uncontaminated water. Moreover, the coefficient of variations of FI o/w achieved values of approximately 10% for the oil-to-water ratios 0.5 × 10 −6 and 50 × 10 −6 , 14% for the oil-to-water ratio 50 × 10 −6 , and 19% for the oil-to-water ratio 500 × 10 −6 . This indicates that the diversity of FI o/w for all oil samples within a given oil-to-water ratio is small, and the FI o/w values indicate high similarity.

Conclusions
The value of the fluorometric index (FIo/w) indicates the presence of oil in seawater, although it nonlinearly depends on the oil-to-water ratio, but only for low values. The FIo/w values remain at the same level when the oil-to-water ratio exceeds 50 × 10 -6 . The influence of the type of oil does not manifest itself in an unambiguous way. However, the performed statistical calculations indicate a high similarity in the FIo/w values independent of the type of oil as well as the oil-to-water ratio. Moreover, no case was found where the value of FIo/w for seawater exposed to oil was not different from the value of FIo/w for natural seawater. At an oil-to-water ratio below 0.5 × 10 −6 , the detection of oil with the FIo/w indicator is probably possible; however, the quantitative lower limit of detection is not yet specified. In turn, to check the effectiveness of the method against weathered oils and their mixtures, separate research is needed. A possible future-oriented immersion sensor in its structure would have to reflect the idea described in this work (i.e., be equipped with an excitation channel (225 nm) and two emission channels (340 and 355 nm)). The listed wavelengths refer to the area of the Gulf of Gdańsk (Baltic Sea); therefore, in relation to other sea areas, the values of these wavelengths should be verified and possibly corrected.

Conclusions
The value of the fluorometric index (FI o/w ) indicates the presence of oil in seawater, although it nonlinearly depends on the oil-to-water ratio, but only for low values. The FI o/w values remain at the same level when the oil-to-water ratio exceeds 50 × 10 −6 . The influence of the type of oil does not manifest itself in an unambiguous way. However, the performed statistical calculations indicate a high similarity in the FI o/w values independent of the type of oil as well as the oil-to-water ratio. Moreover, no case was found where the value of FI o/w for seawater exposed to oil was not different from the value of FI o/w for natural seawater. At an oil-to-water ratio below 0.5 × 10 −6 , the detection of oil with the FI o/w indicator is probably possible; however, the quantitative lower limit of detection is not yet specified. In turn, to check the effectiveness of the method against weathered oils and their mixtures, separate research is needed. A possible future-oriented immersion sensor in its structure would have to reflect the idea described in this work (i.e., be equipped with an excitation channel (225 nm) and two emission channels (340 and 355 nm)). The listed wavelengths refer to the area of the Gulf of Gdańsk (Baltic Sea); therefore, in relation to other sea areas, the values of these wavelengths should be verified and possibly corrected.