Open Access
Add to BibSonomy
Abstract
Probing suspended particles in seawater, such as microalgae, microplastics and silts, is very important for environmental monitoring and ecological research. We propose a method based on polarized light scattering to differentiate different suspended particles massively and rapidly. The optical path follows a similar design of a commonly used marine instrument, BB9, which records backscattering of non-polarized light at 120°. In addition, polarization elements are added to the incident and scattering path for taking polarization measurements. Experiments with polystyrene microspheres, porous polystyrene microspheres, silicon dioxide microspheres, and different marine microalgae show that by carefully choosing the incident polarization state and analyzing the polarization features of the scattered light at 120°, these particles can be effectively differentiated. Simulations based on the Mie scattering theory and discrete dipole approximation (DDA) have also been conducted for particles of different sizes, shapes and refractive indices, which help to understand the relationship between the polarization features and the physical properties of the particles. The laboratory system may serve as a prove-of-concept prototype of new instrumentations for applications on board or even with submersibles.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Microalgae are the foundation of the aquatic food web, the primary producers, feeding everything from microscopic animal-like zooplankton to multi-ton whales [1]. Microalgae can also be the harbingers of death or disease. Certain species of microalgae produce powerful biotoxins, making them responsible for the so-called “red tides”, or harmful algal blooms. These toxic blooms can kill marine life and people who eat contaminated seafood [2]. Microplastics have accumulated in oceans and sediments worldwide in recent years, which may be ingested by low trophic fauna, with uncertain consequences for the health of the organism [3]. As one of the main particles in the ocean, silt has a strong adsorption capacity for marine pollutants and nutrients [4], and also affects the photosynthesis of marine microalgae and other primary producers through the influence of light propagation [5]. Probing these different particulates is a very important issue for the marine science community globally as well as for the daily life use such as food supply.
There are many scattering-based optical instruments for the marine applications. A globally deployed commercial instrument, ECO-BB9 developed by WET Labs, measures the backscattering coefficients at 120° [6], because scattering of the bulk water at 120° is almost not sensitive to the size distribution and composition of particles, leaving the dominant sensitivity attributable to the concentration [7,8]. Recent research shows that the backscattering of the seawater has a relationship to the concentration of the suspended particle mass [9,10], particulate organic carbon [11], phytoplankton carbon [12] and phytoplankton growth rates [13,14]. In addition, the backscattering of the water is independent of the bulk composition of particles [9]. ECO-BB9 has been a popular instrumentation for ocean color research and has been deployed in many marine observatory platforms and stations [15,16]. A large volume of data is already gathered and archived [17].
Polarized light scattering provides more information to identify and classify complex microstructure [18] and suspended particles in the seawater [19] or air [20]. Polarization is an inherent property of the light and is sensitive to the microstructure of the samples [21]. A Stokes vector S as defined in Eq. (1) is usually used to represent the polarization states of light [22],
If the Stokes vector of the incident light is Sin, the polarization state of the light scattered by the particles is Sout, and they can be connected by a Mueller matrix M, which is shown in Eq. (2),The Mueller matrix M characterizes the polarization properties of the particles and is correlated with the size, shape, refractive index, structure of the particle [22]. If we only consider the polarization parts of the Stokes vector, we can define the scattered polarizations by q, u, v, given in Eq. (3),
Compared with optical techniques using unpolarized light, polarimetry techniques have some distinctive advantages. First, since polarization modulation devices such as polarizers and wave plates may not alter the direction of light, traditional optical techniques can be upgraded to take polarization measurements, even the complete Stokes vectors or Mueller matrices by adding the polarization states generator (PSG) and analyzer (PSA) to the existing optical paths. Polarimetry techniques can provide much richer information on the optical properties and the microstructural features of the sample, but still include those available by the non-polarization techniques. Data obtained by the upgraded polarimetry instrumentations can directly correlate to those obtained by the corresponding non-polarization instrumentations.
Polarized photon scattering at different angles may be sensitive to different features of the particles [23,24], in this work, we established an experimental setup for measuring the Stokes vector of the scattered light at the 120° scattering angle. This angle is the same with the popular marine instrumentation ECO-BB9, which uses non-polarized light scattering to obtain the bulk backscattering of the water body including various particulate components, such as microalgae, silt, and microplastics. We also examined the detailed polarized light scattering behavior of particles at the 120° direction by the simulations to correlate and interpret the experimental results. Both the experimental and simulated results of polarized photon scattering at the 120° provide a possibility to correlate the archived marine data obtained by the commercial instrument ECO-BB9 which has already been deployed in many marine observatories or used in ocean expeditions.
2. Experimental setup and methods
2.1 Experimental setup
The experimental setup is shown in Fig. 1(a). A 0.532μm laser with 200mW output power is used as the light source. A variable attenuator (ATT) controls the illumination power. A polarization state generator (PSG) controls the polarization state of the incident light. It consists of a 45° fixed quarter wave-plate (QW1), a rotatable linear polarizer (P) and a rotatable quarter wave-plate (QW2). QW1 is used to change the inherent vertical (V) linear polarization state of the light from the laser to circular polarization states, which can be transformed by P and QW2 to the needed polarization states.
Fig. 1 The schematic of the experimental setup (a) and polarization state analyzer, PSA (b). S, light source; ATT, attenuator; PSG, polarization state generator; QW1, 45° fixed quarter wave-plate; P, rotatable linear polarizer; QW2, rotatable quarter wave-plate; DP, diaphragm; L1 L2 and L3, lens; PH, circular pinhole; PSA, polarization state analyzer; P1, 0°linear polarizer; P2, 90°linear polarizer; P3, 45°linear polarizer; P4, 135°linear polarizer; QW, 135°-fast-axis quarter wave-plate; PMT, photomultiplier tube. The combination of QW and P2 is a left circular analyzer.
Seawater contains various particulate components. To differentiate different particulate components, we need to measure the scattering by individual particles rather than by the bulk water volume. Also, taking polarization measurements of individual suspended particles can achieve better signal to noise ration than taking the average of all the particles in the bulk [25]. The current experimental setup is designed for detecting the scattering signals by the individual suspended particles. By reducing the illumination and detection areas and their overlaps, we are able to ensure that the detected scattering photons come mainly from individual particle which passes through the detection area at that time, i.e., individual particle detection. In the meantime, the small detection volume also helps to reduce the background signal due to reflections on the surfaces and scatterings by water. Pulses due to the bursts of scattered photons by individual particle should not overlap. A diaphragm (DP) and a lens (L1) are used to control the shape and volume of the focused illumination light. In the detection arm, the light scattered by the particles is collected by a 50mm focal length lens (L2) and focused on a circular pinhole (PH) of 100μm diameter. The spatial location of PH is adjusted carefully to ensure that the detection volume defined by L2 and PH includes the focal spot of L1. The intersection volume of these two optical paths is the scattering volume [22].
Note that, even though we measure the scattering by individual particles, we collect at the same time the intensity and the polarization components of the scattered light by the small water body within the scattering volume. If we integrate the signal over certain time duration, the backscattered intensity of the water body should be the same as the results from ECO-BB9 at the same wavelength targeted to the same water body. Therefore, results from the polarization measurements, which contains much more details microstructural information for efficient differentiation of diverse types of particles, should be directly correlated to the backscattering data from ECO-BB9.
The suspended particles are contained in a glass beaker and stirred by a magnetic stirrer at a speed of 200 rounds per minute. The curved surface of the glass beaker effectively acts as a cylindrical lens for both the illumination and scattered lights. To remove these effects, the glass beaker is placed at the center of a glass dodecagon cuvette filled with water. Backscattered light at 120° is collected through the corresponding flat facet of the cuvette.
The scattered light passing through the PH is then collimated by a lens (L3) and its polarization states are measured by the polarization state analyzer (PSA). As the suspended particles are moving constantly, polarization components of the scattered light have to be taken simultaneously [25], as shown in Fig. 1(b). The PSA contains 3 non-polarizing beam splitter cubes which split the collimated scattered light into four channels with constant proportions. Four polarization components are obtained through three linear analyzers at 0° (P1), 45° (P3), 135° (P4) respectively and a left circular analyzer which consists of a 135°-fast-axis quarter wave-plate (QW) and a 90° linear polarizer (P2). The four polarization components of the scattered beam are detected and digitized simultaneously by four photomultiplier tubes (PMT, CH253, made by HAMAMATSU, Japan) and a multi-channel data acquisition card (DAQ, FCFR-USB2066, made by FCCTEC Technology, China).
A 4*4 instrument matrix A is used to calculate Sout from the four polarization components, using
Where = [I1; I2; I3; I4] and I1, I2, I3, I4 are the four polarization components, respectively. To obtain A, we use PSG to generate a series of known polarization states, where i ranges from 1 to N, given N polarization states. We directly input into the PSA as shown in Fig. 1(b) and then obtain the related. So according to Eq. (4), we have,To obtain A correctly, N and should be carefully selected [26,27]. In this paper, we rotate the QW2 of the PSG as shown in Fig. 1(a) 180 times with 1° for each time to generate .We use the series of the odd numbers to obtain A and those of even numbers to validate it. Finally, we can obtain q, u and v in Eq. (3) correctly and reduce the measurement error to less than 0.04.2.2 Samples
Several types of particles with simple morphological structures as well as marine microalgae are used in the experiments. We use particles of different sizes, refractive indices and surface morphology to mimic marine particles of various types, shapes and internal structures, including solid polystyrene (PS) microspheres with average diameter 10.0μm and refractive index 1.59, porous polystyrene (P-PS) microspheres with average diameter 10.0μm, refractive index 1.59 and tiny holes (diameter around 0.1μm) on the surface, and solid silicon dioxide (SD) microspheres with average diameter 10.0μm and refractive index 1.46. All these samples are supplied by NanoMicro Company, Suzhou, China.
Two species of the marine microalgae, i.e., Chattonella marina (CM) and Alexandrium tamarense (AT), are used in this paper. Both of them are red tide algae and poisonous to a human being. AT is a species of dinoflagellates known to produce saxitoxin, a neurotoxin which causes the human illness clinically known as paralytic shellfish poisoning [28]. The cells are of spherical shape with a diameter of 20~40μm as shown in Fig. 2(a).
Fig. 2 Microscopic images of the two species of marine microalgae cells. (a) Alexandrium tamarense. (b) Chattonella marina.
CM is a species of Chattonellales known to produce neurogenic, hemolytic and coagulant toxin, which causes a large number of fish deaths and results in red tides [29]. The cells are of spindle-like shape 30~74μm long and 20~30μm wide, as shown in Fig. 2(b).
2.3 Signals
As individual particles pass through the scattering volume, the polarization components of the scattered light are recorded simultaneously in the four channels. Figure 3(a) shows a fraction of the signals of P3 channel (45° linear analyzers) scattered by the suspended polystyrene microspheres, where the incident light is the 0° linearly polarized. The signals consist of a series of temporal pulses. Intensities of the pulses fluctuate seriously since the particles are illuminated by a non-uniformed light field in the scattering volume. However, since polarization parameters are all based on differential and ratiometric measurements, they are much less sensitive to fluctuations in intensities. In the experiments, signals due to polarized photons scattering by the particles are identified if signal-to-noise ratios in any of the four channels are bigger than 5. Intervals between the pulses are related to the concentration of the particles and the stirring speed of the magnetic stirrer. The full widths at half maximum of the pulses are about 0.1-0.3 millisecond (ms). The sampling frequency of DAQ is 200 kilo-Hz, which ensures enough samplings to recover the detailed shape of the pulses for accurate measurements of their intensities.
Fig. 3 The time series of the signal with 0° linearly polarized incident light and 45° linearly polarized analyzing. (a) Signals of the aquatic suspension of the polystyrene microspheres, (b) signals of the water without particles, and the two inserted plots are the details of two selected (dashed line square) fragments. Note that the scales of the vertical axes in (a) and (b) are different.
In the experiments, the concentration of the particles as well as the size of the scattering volume is controlled to maintain individual particle detection in the scattering volume. For this, we increase the concentration of the particles gradually. The number of recorded pulses increases linearly at low concentration but saturate at higher concentration. We set the concentration of the particles to less than 1.0 million particles per milliliter, to balance the individual particle detection and the detection efficiency. In real applications, the number density for the microalgae is generally less than 0.1 million per milliliter even during the red tides.
Before each experiment, we monitor the background signal level of the system under the same experimental condition but without particles in the stirred water. Figure 3(b) shows the signals without the particles, and two inserted plots show the two selected fragments. One can see that the background is very low, and only one pulse with a small peak value occurs in the same time duration as Fig. 3(a). So, test results on the background of the system show that the stray reflections on the surfaces of the glass beaker and the dodecagon cuvette, as well as the scattering by water and the air bubbles, do not cause significant influence to the detected pulses due to scattering by the individual particles.
And in the experiments, we slightly deviate the light source beam from the normal incidence on the surfaces, so the reflected lights keep away from the scattering volume, which effectively avoids the scattered intensity illuminated by the reflected light contributing to the measured signals.
2.4 Data analysis
Linear discrimination analysis (LDA) [30] is a method used to find a linear combination of the normalized elements of Stokes vectors, [q u v], for the best separation of the two sets of data, events or objects. Different linear transformations of the features of the two sets are tried to evaluate the probability distribution of the two sets after the transformation. Since LDA assumes a normal distribution, the probability distribution of the two sets after the linear transformation can be summarized with their mean (and) and standard deviation (and). Then LDA maximize a target function
Until the maximum value Lm is achieved, which corresponds to the maximum variance between the two distributions , and the minimum variance of the two sets . In the meantime, an alternative and a more commonly used way to calculate the optimal linear transformation is through Bayes' Law and probability density function, with the use of covariance matrix [30,31]. After the training, LDA gives an explicit function f in Eq. (7), which transfers [q u v] to a one-dimensional parameter x,3. Experimental results
In the experiments, the samples are suspended particles diluted to less than one million per milliliter. For each sample, we set 6 different incident polarization states for the illumination lights, i.e., horizontal (H-), vertical(V-), 45° (P-), 135°(M-) linear polarization states, right(R-) and left (L-) circular polarization states. For each incident polarization, we record the scattered light for two minutes. The scattering pulses, as shown in Fig. 3, are extracted, and the Stokes vectors of the scattered light are calculated.
The comparisons between the experimental results of the suspended particles PS (polystyrene microspheres) and P-PS (porous polystyrene microspheres) are shown in Fig. 4. Figures 4(a)-4(f) show the scattered polarization distributions and LDA distribution corresponding to H-, V-, P-, M-, R- and L- incident polarization states respectively. We can find that for each microsphere, the scattered [q u v] distributes in the certain ranges. For P-, M-, R- and L-incident polarizations, the distributions of the two types of particles are clearly different from each other. And for H-incident polarization, these two distributions overlap slightly but still can be separated. However, for V-incident polarization, the two distributions overlap pointedly and cannot be distinguished. From these figures, we can find that for these two particles, there should be an optimal incident polarization to separate the two distributions better.
Fig. 4 The scattered [q u v] (dots) and LDA distributions (lines in the dashed axes) of PS and P-PS. (a) H- incident, Lm = 2.10. (b) V- incident, Lm = 1.20. (c) P- incident, Lm = 9.30. (d) M- incident, Lm = 9.65. (e) R- incident, Lm = 8.04. (f) L- incident, Lm = 7.53.
To differentiate the two types of particles quantitatively, we use LDA to transform the distributions of the particles from 3-dimensional polarization space to one-dimensional distributions. The maximum separation Lm is calculated as Eq. (6). We can see the good performance of P-, M-, R- and L-incident more clearly. Comparing those LDA figures, we can find that the combination of the optimal incident polarization and LDA will help to differentiate the two types of particles by the scattered polarizations.
Beyond that, pairwise contrast experiments of other samples have also been done. The Lm has been calculated as a criterion for the differentiation capability (Table 1). We can see that R-incident, L-incident, M-incident and R-incident work better in the differentiation of the PS (polystyrene microspheres) and SD (solid silicon dioxide microspheres), CM (Chattonella marina algae) and AT (Alexandrium tamarense algae), PS and CM, SD and CM respectively. All these experiments demonstrate a good distinction between different suspended particles.
Table 1. Lm of different particulates with the different incident polarized light.
4. Analysis and simulations
From Table 1, we can find that the scattered polarization, [q, u, v], at 120° can differentiate particles with different physical properties, such as size, fine structure, shape and refractive index. In details, the PS and P-PS differ in the existence of the tiny holes on the surface. The PS, SD, and microalgae cells are different in refractive indices, and the two species of the microalgae cell have different shapes and internal structures. Finding the correlation between the scattered polarizations and the physical properties of the particles is important to interpret the experimental results.
It has been known that polarization states of scattered photons are sensitive to the fine structure of the scattering object [18]. We chose samples with distinctive microstructural features, and then tried simulations and calculations to examine how these features affect the polarizations of the scattered photons. For example, we use submicron microspheres to represent the effects of fine structures, such as the holes on the P-PS surface and tiny organelles in the cells, and use microspheres and microcylinders with different aspect ratios to simulate the effects of different shapes. The scattered polarizations of microspheres are calculated by the Mie theory [22], and those of microcylinders are calculated by the discrete dipole approximation (DDA) [32]. DDA is a flexible and powerful technique for computing scattering and absorption by targets of arbitrary geometry, which approximates the particle in terms of discrete dipoles [33].
4.1 Particles with different sizes and refractive indices
The 45° (P) linear polarization is used as the incident polarization. In Fig. 5(a), q, u and v vary with particle diameter between 0.1μm and 20.0μm, with the refractive index 1.46. We can see that q, u and v change as the diameters of the microspheres increase. When the diameter is less than 1μm, the polarization behaviors of the particles change sharply with the size of the particles, which explains why the scattered polarization exhibits more sensitivity to the submicron microspheres.
Fig. 5 The variation tendency of q, u and v scattered by different microspheres in the simulation. (a) The variation tendency of q, u and v scattered by microspheres with diameters varied from 0.1μm to 20.0μm, and the refractive index 1.46. (b) The variation tendency of q, u and v scattered by microspheres with refractive index varied from 1.07 to 1.66, and the diameter 10.0μm.
However, when the diameter is larger than 5μm the scattered polarization changes slightly and becomes constant with increasing diameter, which shows that the scattered polarizations are non-sensitive to these big particles.
Figure 5(b) shows the variations of q, u and v as the refractive index of particles vary from 1.07 to 1.66, with the diameter 10.0μm. Relatively, we can see more changes in u, but smaller changes in q and v at the P-incident. Notably, q, u and v scattered by the particles with the refractive index less than 1.46 change more rapidly than those larger than 1.46. And in the latter case, the scattered u becomes constant with increasing refractive index but q and v still varies slightly.
4.2 Particles with different shapes
We use microcylinders with different aspect ratios (length/diameter) and microspheres to study the influence of the particle’s shape on the scattered polarizations [34]. It is known that the orientation of the microcylinders influences its scattered intensity and the polarization [22]. In the experiments, we try to reduce the orientation distribution of the suspended particles by stirring a flow field. In the simulations, we set the orientation of the microcylinders perpendicular to the scattering plane determined in Fig. 1. And then we compare the scattered polarizations of microcylinders with different aspect ratios. In the simulation, we keep the microcylinders with the same refractive index (1.46), and the same diameter (2μm). The incident light is right-circularly polarized.
Figure 6 shows the scattered polarizations, q, u and v, vary with the changing aspect ratio from 0.5 to 4, except for the first point which is a microsphere with the same refractive index and diameter. The scattered polarizations are quite different from those of the microspheres, but the main difference is the circular component v. However, the particles become more anisotropic with increasing aspect ratio, and the changes of v are the least. And after the aspect ratio is larger than 2, the scattered polarizations become steady. From Fig. 6, we can see that the shapes of the particles have a remarkable influence on the scattered polarizations till the aspect ratio is higher enough (larger than 2 in Fig. 6).
Fig. 6 The variation of q, u and v as the aspect ratio changes from 0.5 to 4. The first points marked circle are the calculated q, u and v of a microsphere with the diameter 2μm, and the rests marked diamond are that of microcylinders with the same diameter but different aspect ratios.
5. Discussions
5.1 Analysis of experimental results
Some samples in Table 1 are more complex than the microspheres and microcylinders in the simulation, which cannot be perfectly reproduced in the current calculations. For example, in the simplest sample, the PS, the diameter of individual particles may vary within 2% around its average value, 10μm; and each particle may cross the scattering volume following slightly different routes, all of which result in different polarization states in the scattered photons as shown in Fig. 4 by the scattered dots in the 3-dimensional polarization space. The P-PS particles have more complicated structure, so the scattered polarization states in Fig. 4 distribute more widely than those of the PS. By comparing the polarization behaviors of the two samples, we understand further on the dominant factors in the scattered polarization of the samples.
For spheres, the ideal Mueller matrix M can be represented as
For H-incident, [q u v] = [1 0 0]; and for V-incident, [q u v] = [-1 0 0]. So, for spheres, under H- and V-incident, the scattered [q u v] are the same and are not able to differentiate different spheres. So, we can find the polystyrene microspheres and the silicon dioxide microspheres are similar in the [q u v] and LDA distributions, as shown in Table 1. Moreover, H- and V-incident can be used to probe the different shapes of the suspended particles [22]. For the other 4 incident polarizations, the scattered [q u v] are related to m12, m33 and m34, which can characterize different microspheres.The main difference between the PS and P-PS is the existence of 0.1μm holes on the surface of the P-PS. Simulation results in Fig. 5(a) show that submicron holes (or particles with refractive index 1) will affect the scattered polarization more than the bigger particles. This evidence strengthens the correlation between the tiny holes and the different scattered polarizations.
Similarly, the main difference between the PS and SD is the refractive index, 1.59 versus 1.46. The simulation results in Fig. 5(b) show that these two refractive indices can introduce the differences in q and v, which helps to understand the experiments. However, we note that Lm of the PS and SD in Table 1 is much less than those of the PS and CM, and the SD and CM. This can be understood by Fig. 5(b) which indicates that for particles with their refractive indices less than 1.46 (CM), the polarization state of the scattered photons is much more sensitive to the changes in the refractive index.
The two microalgae cells, the CM and AT, as shown in Fig. 3, have different shapes and intracellular structures and organelles. The AT is almost spherical but the CM is spindle-like whose aspect ratio is around 3. As implied by Fig. 6, the aspect ratio of the CM should contribute to the difference of scattered polarization from the AT. In addition, the intracellular structures and organelle of these two cells with the submicron sizes will contribute to the difference. Meanwhile, these intracellular structures and organelles also contribute to the difference between the CM and PS or SD.
5.2 Optimizing the incident polarization
From Fig. 4 and Table 1, we can see different recognition results with different incident polarization states, indicating there should be an optimal incident polarization for the best separation of the two samples. We take polystyrene microspheres and porous polystyrene microspheres for example. As shown in Fig. 4, the P-, M-, R- and L-incident polarization states result in better separations of the species. Since in many applications, linear polarizations are easier to obtain and control than circular polarization and are more commonly used [35–37], we first try to find out the optimal linear incident polarization. We remove the QW2 in the experimental setup and rotate the polarizer P to generate different incident polarizations. For each incident linear polarization, we measure the scattered polarizations [q, u, v] and evaluate the maximal value Lm of LDA to assess the differentiation effects quantitatively. Figure 7(a) shows how Lm changes as the linear polarization angle of the P rotates. Lm reaches its maximum value at 153°linear polarization for the incident light. An experiment is carried out on the suspension of these two kinds of microspheres by using the 153° linear polarization state as the incident light. Firstly, we measure the scattered polarizations of the two samples separately, which is shown as Fig. 7(b). Then, LDA function f in Eq. (5) can be calculated and inserted in Fig. 7(b) as same as those in Fig. 4. Finally, a mixed suspension of the two types of microspheres is measured. The scattered polarizations and LDA distributions are shown in Fig. 7(c).
Fig. 7 Optimizing the incident polarization. (a) The Lm changes as the linear polarization angle rotate. (b) The separately measured scattered [q u v] and LDA distributions of the PS and P-PS with the optimized incident polarized light, Lm = 9.82. (c) The scattered [q u v] and LDA distributions of the mixed sample with the optimized incident polarized light.
Comparing Figs. 7(b) and 7(c), we can find that the scattered polarizations of the mixed suspension well repeat those of the independent samples. And the two particles can be well differentiated by the scattered polarizations.
Combining Fig. 7 and Table 1, one can see that we can differentiate different types of particles suspended in water using their different polarization features. In real applications, we can vary the polarization states of the incident beam and take the Stokes of the scattered light. The experimental results will be analyzed and compared with a database which contains the polarization features of known particles such as different types of algae [38]. Such database can be expended continuously as the Mueller matrices of new types of particles are measured and their distinctive polarization features are identified.
6. Conclusions
In this paper, we present a polarized light scattering method for differentiating suspended particles of different physical and microstructural properties, which is important for monitoring microalgae, microplastics, and silt concentrations. An experimental setup is established to illuminate the suspended particles with polarized light and measures the polarization state of the scattered light at 120°. The polarization states of the scattered beam of some typical samples, such as polystyrene microspheres, porous polystyrene microsphere, silicon dioxide microsphere, and marine microalgae have been measured. Experimental results show that the scattered polarizations can differentiate any two kinds of these particles. Simulations based on the Mie theory and discrete dipole approximation confirm that the scattered polarizations are sensitive to the submicron microstructure, refractive index and shape of the particles, which help to explain how the microstructure and the physical properties of the particles are correlated to the scattered polarization states. The experimental results based on the laboratory system and the simulations prove the feasibility of a new technique and may serve as the proof-of-concept prototype for new instrumentations used on board ships or even with submersibles.
Funding
National Natural Science Foundation of China (NSFC) (41527901, 61527826); Shenzhen Municipal Government (JCYJ20160818143050110, JCYJ20150529164918736); Chinese Academy of Sciences (XDB06020203); Graduate School at Shenzhen, Tsinghua University (JC2015001).
Acknowledgments
The authors thank Dr. Jin Zhou and Dr. Yuelu Jiang in Division of Ocean Science and Technology, Graduate School at Shenzhen, Tsinghua University for providing the microalgae samples.
References and links
1. T. Garrison, Essentials of Oceanography (Cengage Learning, 1999).
2. D. M. Anderson, “Turning back the harmful red tide,” Nature 388(6642), 513–514 (1997). [CrossRef]
3. S. L. Wright, R. C. Thompson, and T. S. Galloway, “The physical impacts of microplastics on marine organisms: a review,” Environ. Pollut. 178, 483–492 (2013). [CrossRef] [PubMed]
4. I. D. James, “Modelling pollution dispersion, the ecosystem and water quality in coastal waters: a review,” Environ. Model. Softw. 17(4), 363–385 (2002). [CrossRef]
5. R. J. Davies-Colley, C. W. Hickey, J. M. Quinn, and P. A. Ryan, “Effects of clay discharges on streams. 1. Optical properties and epilithon,” Hydrobiologia 248(3), 215–234 (1992). [CrossRef]
6. D. McKee, M. Chami, I. Brown, V. S. Calzado, D. Doxaran, and A. Cunningham, “Role of measurement uncertainties in observed variability in the spectral backscattering ratio: a case study in mineral-rich coastal waters,” Appl. Opt. 48(24), 4663–4675 (2009). [CrossRef] [PubMed]
7. X. Zhang, E. Boss, and D. J. Gray, “Significance of scattering by oceanic particles at angles around 120 degree,” Opt. Express 22(25), 31329–31336 (2014). [CrossRef] [PubMed]
8. E. Boss and W. S. Pegau, “Relationship of light scattering at an angle in the backward direction to the backscattering coefficient,” Appl. Opt. 40(30), 5503–5507 (2001). [CrossRef] [PubMed]
9. E. Boss, W. S. Pegau, M. Lee, M. Twardowski, E. Shybanov, G. Korotaev, and F. Baratange, “Particulate backscattering ratio at LEO 15 and its use to study particle composition and distribution,” J. Geophys. Res.-Oceans 109, C001514 (2004).
10. E. Boss, L. Taylor, S. Gilbert, K. Gundersen, N. Hawley, C. Janzen, T. Johengen, H. Purcell, C. Robertson, D. W. H. Schar, G. J. Smith, and M. N. Tamburri, “Comparison of inherent optical properties as a surrogate for particulate matter concentration in coastal waters,” Limnol. Oceanogr. Methods 7(11), 803–810 (2009). [CrossRef]
11. D. Stramski, “Corrigendum to “Relationships between the surface concentration of particulate organic carbon and optical properties in the eastern South Pacific and eastern Atlantic Oceans,” Biogeosciences 4, 93–99 (2008).
12. V. Martinez-Vicente, G. Dall’Olmo, G. Tarran, E. Boss, and S. Sathyendranath, “Optical backscattering is correlated with phytoplankton carbon across the Atlantic Ocean,” Geophys. Res. Lett. 40(6), 1154–1158 (2013). [CrossRef]
13. M. J. Behrenfeld, E. Boss, D. A. Siegel, and D. M. Shea, “Carbon‐based ocean productivity and phytoplankton physiology from space,” Glob. Biogeochem. Cycle 19, 2299 (2005).
14. R. A. Reynolds, D. Stramski, and G. Neukermans, “Optical backscattering by particles in Arctic seawater and relationships to particle mass concentration, size distribution, and bulk composition,” Limnol. Oceanogr. 61(5), 1869–1890 (2016). [CrossRef]
15. T. S. Moore, C. B. Mouw, J. M. Sullivan, M. S. Twardowski, A. M. Burtner, A. B. Ciochetto, M. N. McFarland, A. R. Nayak, D. Paladino, N. D. Stockley, T. H. Johengen, A. W. Yu, S. Ruberg, and A. Weidemann, “Bio-optical properties on Cyanobacteria blooms in Western Lake Erie,” Front. Mater. Sci. 4, 300 (2017). [CrossRef]
16. L. G. Sokoletsky and F. Shen, “Optical closure for remote-sensing reflectance based on accurate radiative transfer approximations: the case of the Changjiang (Yangtze) River Estuary and its adjacent coastal area, China,” Int. J. Remote Sens. 35(11–12), 4193–4224 (2014). [CrossRef]
17. A. Mignot, R. Ferrari, and H. Claustre, “Floats with bio-optical sensors reveal what processes trigger the North Atlantic bloom,” Nat. Commun. 9(1), 190 (2018). [CrossRef] [PubMed]
18. M. Sun, H. He, N. Zeng, E. Du, Y. Guo, S. Liu, J. Wu, Y. He, and H. Ma, “Characterizing the microstructures of biological tissues using Mueller matrix and transformed polarization parameters,” Biomed. Opt. Express 5(12), 4223–4234 (2014). [CrossRef] [PubMed]
19. M. Chami, “Importance of the polarization in the retrieval of oceanic constituents from the remote sensing reflectance,” J. Geophys. Res. Oceans 112, C05026 (2007).
20. D. Li, N. Zeng, D. Zhan, Y. Chen, M. Zeng, and H. Ma, “Differentiation of soot particulates in air using polarized light scattering method,” Appl. Opt. 56(14), 4123–4129 (2017). [CrossRef] [PubMed]
21. N. Ghosh and I. A. Vitkin, “Tissue polarimetry: concepts, challenges, applications, and outlook,” J. Biomed. Opt. 16(11), 110801 (2011). [CrossRef] [PubMed]
22. C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (Wiley, 1983).
23. M. Chami, A. Thirouard, and T. Harmel, “POLVSM (Polarized Volume Scattering Meter) instrument: an innovative device to measure the directional and polarized scattering properties of hydrosols,” Opt. Express 22(21), 26403–26428 (2014). [CrossRef] [PubMed]
24. D. Li, N. Zeng, D. Zhan, Y. Chen, M. Zeng, and H. Ma, “Differentiation of soot particulates in air using polarized light scattering method,” Appl. Opt. 56(14), 4123–4129 (2017). [CrossRef] [PubMed]
25. R. Liao and H. Ma, “Study on errors of nonsimultaneous polarized-light scattering measurements of suspended rod-shaped particles,” Appl. Opt. 54(3), 418–424 (2015). [CrossRef]
26. J. Chang, N. Zeng, H. He, Y. He, and H. Ma, “Single-shot spatially modulated Stokes polarimeter based on a GRIN lens,” Opt. Lett. 39(9), 2656–2659 (2014). [CrossRef] [PubMed]
27. J. Chang, H. He, Y. Wang, Y. Huang, X. Li, C. He, R. Liao, N. Zeng, S. Liu, and H. Ma, “Division of focal plane polarimeter-based 3 × 4 Mueller matrix microscope: a potential tool for quick diagnosis of human carcinoma tissues,” J. Biomed. Opt. 21(5), 56002 (2016). [CrossRef] [PubMed]
28. G. R. Persich, D. M. Kulis, E. L. Lilly, D. M. Anderson, and V. M. T. Garcia, “Probable origin and toxin profile of Alexandrium tamarense (Lebour) Balech from southern Brazil,” Harmful Algae 5(1), 36–44 (2006). [CrossRef]
29. M. Shen, J. Xu, M. W. Chiang, and D. W. Au, “Unravelling the pathway of respiratory toxicity in goldlined seabream (Rhabdosargus sarba) induced by the harmful alga Chattonella marina,” Aquat. Toxicol. 104(3-4), 185–191 (2011). [CrossRef] [PubMed]
30. T. Sapatinas, Discriminant Analysis and Statistical Pattern Recognition (Wiley-Interscience, 2004).
31. R. O. Duda, P. E. Hart, and D. G. Stork, Pattern Classification, 2nd ed. (Wiley, 2001).
32. B. T. Draine, “Discrete-dipole approximation and its application to interstellar graphite grains,” Astrophys. J. 333, 848–872 (1988). [CrossRef]
33. B. T. Draine and P. J. Flatau, “Discrete-Dipole Approximation for Scattering Calculations,” J. Opt. Soc. Am. A 11(4), 1491–1499 (1994). [CrossRef]
34. H. R. Gordon, “Light scattering and absorption by randomly-oriented cylinders: dependence on aspect ratio for refractive indices applicable for marine particles,” Opt. Express 19(5), 4673–4691 (2011). [CrossRef] [PubMed]
35. J. S. Tyo, D. L. Goldstein, D. B. Chenault, and J. A. Shaw, “Review of passive imaging polarimetry for remote sensing applications,” Appl. Opt. 45(22), 5453–5469 (2006). [CrossRef] [PubMed]
36. M. K. Swami, S. Manhas, P. Buddhiwant, N. Ghosh, A. Uppal, and P. K. Gupta, “Polar decomposition of 3 x 3 Mueller matrix: a tool for quantitative tissue polarimetry,” Opt. Express 14(20), 9324–9337 (2006). [CrossRef] [PubMed]
37. H. He, N. Zeng, R. Liao, T. Yun, W. Li, Y. He, and H. Ma, “Application of sphere-cylinder scattering model to skeletal muscle,” Opt. Express 18(14), 15104–15112 (2010). [CrossRef] [PubMed]
38. X. Li, R. Liao, J. Zhou, P. T. Y. Leung, M. Yan, and H. Ma, “Classification of morphologically similar algae and cyanobacteria using Mueller matrix imaging and convolutional neural networks,” Appl. Opt. 56(23), 6520–6530 (2017). [CrossRef] [PubMed]
