In order to assess the capability to optically identify small marine microbes, both simulations and experiments of angular resolved light scattering (ARLS) were performed. After calibration with 30-nm vesicles characterized by a nearly constant scattering distribution for vertically polarized light (azimuthal angle=90°), ARLS from suspensions of three types of marine picoplankton (two prokaryotes and one eukaryote) in seawater was measured with a scattering device that consisted of an elliptical mirror, a rotating aperture, and a PMT. Scattered light was recorded with adequate signal-to-noise in the 40–140°. Simulations modeled the cells as prolate spheroids with independently measured dimensions. For the prokaryotes, approximated as homogeneous spheroids, simulations were performed using the RM (Rayleigh-Mie) – I method, a hybrid of the Rayleigh-Debye approximation and the generalized Lorentz-Mie theory. For the picoeukaryote, an extended RM – I method was developed for a coated spheroid with different shell thickness distributions. The picoeukaryote was then modeled as a coated sphere with a spherical core. Good overall agreements were obtained between simulations and experiments. The distinctive scattering patterns of the different species hold promise for an identification system based on ARLS.
© 2006 Optical Society of America
The importance of marine microbes in both the evolution and maintenance of life on earth cannot be understated. Current theories about the origin of life and the evolution of an oxygen rich atmosphere strongly implicate marine microbes . Moreover, marine microbes are key players in the marine food web that fix over 70% of nitrogen and carbon in the oceans , are also responsible for over half of the global primary productivity , and over 50% of the atmosphere oxygen supply . Unfortunately, picoplankton with diameters smaller than 2 µm are least understood, nevertheless extremely important for oceanic productivity. Their small sizes ensure very high nutrient capture efficiency due to the large surface area to volume ratio.
Currently, the most popular optical technique for identifying marine microbes is flow cytometry, which detects the forward and side-scattered light and fluorescence signal from cells . Using this method, subgroups within the autotrophic picoplankton fraction (Synechococcus, Prochlorococcus, and picoeukaryotes) have been identified by differences in the fluorescence properties of their photosynthetic pigments . Absence of pigments in heterotrophic picobacteria has rendered them more difficult to classify. As one solution, bacterial DNA can be stained with fluorescence dye .
Additional methods that have recently gained popularity use PCR methods in order to identify DNA sequences, however, these are somewhat laborious, especially when compared with optical techniques.
It is generally acknowledged that a rapid, simple, non-invasive and inexpensive way to identify different groups of picoplankton would be of great use. One interesting option is to use angularly resolved light scattering (ARLS). When a cell is illuminated, it scatters light in all directions. The spatial distribution of the scattered light intensity depends on a cell’s size, shape, refraction index, density, and morphology. Studying the scattered light in specific angular ranges or the overall pattern can therefore potentially enable the determination of morphological information from the cell, which can be used for discriminating between different cell types.
Scattering measurement of either microbes or the bulk properties of seawater have been undertaken by other researchers. Such systems have typically relied on either a finite set of photomultiplier tube, i.e. PMTs  or the use of a pivoting or scanning transmitter and detector . A more recent design  uses a rotating periscopic prism which allows measurement of the Volume Scattering Function from 0.6° to 177.3° with an angular resolution of 0.3°. Here, a complete measurement takes at least 1.5 minutes. Besides free space optics, a wide-angle, semi-circular array of 175 optical fibers and a scanning disk were used for fast collection of multi-angle scattering to a single PMT , which allows scattering measurement from 3° to 177° with 1° angular resolution and 1 set of measurement in as short as 10 ms.
In this article, a rapid and high angular resolution measurement system based on free space optics is presented to provide label-free non-invasive elastic scattering measurements of lowconcentration cell suspensions. Here, via computer simulations and the rapid collection of data, it is demonstrated that the measurements have the potential to differentiate between subgroups in picoplankton with similar sizes and shapes.
The scattering phase functions of three representative species within the picoplankton are considered. A cyanobacteria, Synechococcus sp. strain CC9311, is among the smallest, but most abundant, photosynthetic microorganisms in the world’s oceans. A picoeukaroyote, Ostreococcus sp. strain CCE9901 is studied which is the smallest free living eukaryote on the earth. Lastly, scattering from Flavobacterium sp. strain ALC1 was measured. All three species have two dimensions not larger than 1 µm, one dimension of about 2–3 µm and resemble ellipsoids in shapes. Angularly resolved scattering was measured over the range of 0–180° with adequate signal-to-noise ratio in the 40–140° range. The scattering measurements from low-concentration suspensions show significant differences between species despite their similar sizes and shapes.
In addition to the experimental studies, scattering simulations using different models were also performed. A homogeneous spheroid model based on the RM-I method developed by Latimer [11, 12] was used to calculate theoretical scattering curves from the prokaryotes. In addition, an extended RM-I method for a coated spheroid was developed to simulate the scattering from the eukaryote. The modeling results show reasonable agreement with the experimentally measured data, especially in the observed vs. predicted positions of the scattering maxima and minima.
2. Theoretical modeling
Generalized Lorentz-Mie Theory (GLMT) provides a rigorous solution for light scattering by a homogeneous, isotropic sphere of arbitrary size . However, most biological cells are not perfect spheres. To model cells more accurately, GLMT can be extended. For spherical eukaryote cells, a coated sphere model can be used , where boundary conditions at both core-shell interface and shell-medium interface need to be satisfied. On the other hand, nonspherical cells without significant internal organelles can be modeled as homogeneous spheroids . Among various approaches that can predict the scattering cross-section of a spheroid, one approximation that has high accuracy with low computation complexity is RM (Rayleigh-Mie) – I, a hybrid method of Rayleigh-Debye approximation and the exact GLMT .
For a spheroid defined by the equation x 2/(av)2+y 2/a 2+z 2/a 2=1 and orientation (γ, ψ), at each scattering angle θ, an equivalent sphere of radius a·g is defined according to the Rayleigh-Gans-Debye approximation. Here, g is a function of the spheroid’s axial ratio v, orientation (γ, ψ), and the scattering angle θ. The exact GLMT can then be used to determine the scattering of the equivalent sphere. The scattering phase function of the spheroid is obtained by scaling the corresponding scattering from the equivalent sphere by a factor of v 2/g 6. Figure 1 illustrates the coordinate system used in the following equations. Using standard trigonometry one can determine that
therefore g can be calculated as :
In order to simulate the scattering from a suspension, an incoherent superposition of uniformly sampled random angles was used. The angles γ and ψ were sampled so that the increment of the solid angle Ω is isotropic on the sphere. In spherical coordinates this implies that
Therefore, in order to sample dΩ uniformly, cos(ψ) was drawn from a uniform distribution on the interval [-1,1), while γ was equally spaced in [0°,180°).
Based on observation, an average non-dividing cell of Synechococcus sp. strain CC9311 has a spheroid shape with outer dimensions of approximately 0.8×0.8×2.0 µm; whereas a 1.0×1.0×3.5 µm spheroid can be used to approximate a cell of Flavobacterium sp. strain ALC1. Despite the internal structure of these prokaryote cells, they are all modeled as homogeneous spheroids with an index of refraction np =1.406 [14–16]. The refractive index of sea water used is nm =1.339.
In order to model the ARLS for the eukaryote, Ostreococcus sp. strain CCE9901, the spheroid model was first used. Unfortunately, this produced a poor match between the observed and experimentally measured data. This was presumably due to the effect of the nucleus. As one alternative, scattering from non-homogeneous spheroids has mainly been implemented with exact numerical approaches that require extensive computer time [17, 18]. In order to model the scattering from nonspherical eukaryote cells such as those of Ostreococcus sp. strain CCE9901 with higher accuracy and simpler computation, a new simulation method was developed, which combines the RM-I for the homogenous spheroid and the extended GLMT for the coated sphere. Similarly as in RM-I for the homogeneous spheroid, at each (γ, ψ, θ), an equivalent coated sphere with outer radius a·g is defined. The inner radius can be expressed as a·g-t(γ, ψ, θ) for arbitrary thickness distribution t(γ, ψ, θ). Specifically, if t(γ, ψ, θ) is constant, the coated spheroid has a uniform shell thickness, while for a·g-t(γ, ψ, θ)=r=constant, the core is a sphere with radius r. Table 1 illustrates the two cases and their mathematical definitions. The scattering phase function of the coated spheroid is then calculated by scaling the results obtained using the extended GLMT for the equivalent coated sphere at each (γ, ψ, θ). The scaling factor is the same with that from the homogeneous spheroid, which is v 2/g 6.
Based on optical microscopy observation, the cell of Ostreococcus sp. strain CCE9901 is approximated by a spheroid of 1.0×1.0×2.0 µm with a concentric spherical core of radius 0.25 µm. The values used for refractive index were 1.37 for cytoplasm and 1.40 for nucleus . Table 2 lists the models used for scattering simulations for the three picoplankton studied.
3. Materials and method
In order to prepare Large Uni-lamellar Vesicles using the Extrusion Technique (LUVET) , 0.2 mL DOPC (1,2-Dioleoyl-sn-Glycero-3-Phosphocholine) chloroform solution (15mg/ml) was put into a small beaker, and dried with a nitrogen gun in a fume hood. After overnight evaporation in a vacuum desiccator, the lipid film was rehydrated with 1 ml PBS 1x solution, and agitated to insure full mixing. To increase the efficiency of entrapment of watersoluble compounds, the hydrated lipid suspension was poured into a round bottom tube after 30 minutes, and subjected to 3–5 freeze/thaw cycles by being alternately placed in a dry iceacetone bath and warm water bath. The sample was then extruded 15 times through a membrane. Because DOPC has a transition temperature of -20°C, it is not necessary to heat the extruder. For vesicles with 30 nm nominal diameters, the extrusion procedure was performed twice, first with a 100 nm membrane, then with a 30 nm membrane. The vesicle preparation was stored at 4°C for less than 3–4 days before experiments.
3.1. 2 Marine microbes
In order to perform the scattering measurements, three isolates, Synechococcus sp. strain CC9311, Ostreococcus sp. strain CCE9901, and Flavobacterium sp. strain ALC1 were used. Two isolates, Synechococcus sp. strain CC9311, and Ostreococcus sp. strain CCE9901 were grown in standard media (SN, F/4 media, respectively) at 20–22°C and constant light conditions (28 µmol quanta m-2s-1). Flavobacterium sp. strain ALC1 were collected at the location of the pier at the Scripps Institution of Oceaenography (32°53’N, 117°15’W), followed by plating in Zobell agar. The isolate was then grown in Zobell medium (seawater with 5 g peptone and 0.5 g yeast extract per liter, autoclaved) at 25° in a shaker for two days after inoculation.
Seawater for filtration (to be used for dilution) was collected from the top of the sea surface, gravity filtered through a 0.2-µm sterile acid-rinsed Gelman Supor filter capsule, and distributed to acid-cleaned 1-liter polycarbonate bottles.
3.2.1 Experimental setup
Figure 2 shows the experimental setup [21, 22]. Briefly, the system was designed to record light scattering over 360°. Specimens are contained in a cylindrical, polished glass cuvette (outer diameter of 10 mm, height 16 mm, Hellma Cells, Inc., Plainview, NY) held at the focal point of an ellipsoidal reflector. The orientation of the cuvette inside the elliptical mirror is arranged so that the light coming horizontally from the red laser (40 mW, 658 nm, Crystal Laser, Reno, NV) enters and exits the cuvette along the normal of the wall. The laser beam is scattered by particles and then reflected by the elliptical mirror toward its second focal point where the PMT (R3896, Hamamatsu, Bridgewater, NJ) is located. In order to insure that only light from the illuminated volume reaches the PMT, a rotating aperture is used whose height and radial position determines the azimuthal angle at which the scattered light is detected (0° in this case). The direction of scattered light to be detected is selected by the position of the rotating aperture. In practice, due to the symmetry in the scattered light for uniform randomly oriented particles only the range of 0–180° was recorded. Finite divergence of the excitation laser beam prevents detection of light scattered at very small angles and in the vicinity of 180°. Here, light scatter is recorded in the range of 0.5–179.5°. In order to insure that the large dynamic range of the scattered light fell within the detection capability of the PMT, a circularly graded intensity filter was placed after the rotating aperture. The filter attenuated the forward-scattered light the most with the back-scattered light being attenuated the least. SNR of the measurement was also increased by the insertion of an interference filter (646–666 nm) and a condensing lens.
In order to induce scatter, the excitation beam (linearly polarized) was focused into a spot of diameter predicted to be 116 µm. The intensity variation of the red laser was measured to be less than 0.5% (Crystal Laser, Reno, NV). Here, since the beam diameter is much larger than the particle diameter, one can assume a uniform beam profile. The data acquisition system enables PMT sampling frequencies of up to 200 kHz and the rotation velocity of the aperture can be varied from 60 RPM to 2000 RPM. The smallest inter-pulse time interval, defined as the time interval between the start of the collection of one set of 360° angular scattering and the start of the next measurement is 30 ms. The upper limit of the angular resolution is determined by the aperture size, while the actual angular step is determined by the sampling frequency of the PMT and the rotation velocity of the aperture. The azimuthal angle α, defined as the angle between the laser polarization and the detection plane is 0° for horizontal polarization (p-polarization), and 90° for vertical polarization (s-polarization). The diameter of the rotating aperture was 1 mm resulting in an angular resolution of 0.35°. To minimize the interference of extraneous light, all experiments were performed in the dark.
3.2.2 System calibration and validation
A necessary prerequisite prior to the interpretation of data collected from the scattering setup is a system calibration. In order to accomplish this, a preparation of LUVET, as described above in section 3.1.1, was used.
Since a small vesicle can be regarded as a single dipole in a uniform light field, the scattered field is constant for vertical polarization (α=90°), and is a cosine function of the scattering angle for horizontal polarization (α=0°). As a result, small vesicles yield a very flat scattering curve for vertical polarization that can readily be used for system calibration. However, one practical complication is that the sizes of the extruded vesicles are not all identical and, rather, obey a log normal distribution . In order to accommodate this distribution of sizes an extended GLMT for coated spheres  with polydispersity was used to simulate the angularly resolved scattering from suspension of vesicles with 30 nm nominal diameter in Phosphate Buffered Saline (PBS) 1x solution. Figure 3 shows a comparison of the results obtained via computer simulation (3a) with those obtained experimentally (3b). Experimental curves obtained from low-concentrations of suspensions, ~1 particle per scattering volume, agree well with the theoretical results, especially in the 40–140° range where the scattering curve for a=90° is considerably flat. Here, the scattering volume is calculated as V= ·Di , where ω0 =58 µm is the beam waist at the center of the cuvette, and Di =8 mm is the inner diameter of the cuvette. The ratio between the experimental α=90° data [the red curve in Fig. 3(b) and the theoretical α=90° data (the red curve in Fig. 3(a)] was used to obtain corrected experimental results for both microspheres and microbes.
After calibration, scattering measurements on polystyrene microspheres with diameters of 0.6 µm, 0.8 µm and 1.0 µm (PPS06, PPS08 and PPS10, Kisker, Germany) were performed in deionized water in order to validate the experimental apparatus using specimens with more structures in their scattering pattern. The experimental data [Fig. 4(b)] matches GLMT simulation [Fig. 4(a)] very well between 40° and 140°, providing confidence that the data calibrated by the 30 nm vesicles can be used to estimate light scattered from submicron particles.
Angularly dependent scattering diagrams were measured from low-concentration (~1 cell per scattering volume) suspensions of Synechococcus sp. strain CC9311, Ostreococcus sp. strain CCE9901 and Flavobacterium sp. strain ALC1. The data was collected with an aperture rotation rate of 1000 RPM which led to a total angular sampling rate of 1 scan (180°) every 30 ms. A total of 333 sets of scattering phase functions were recorded in 10 seconds. The data was then averaged and the systematic calibration correction was applied in order to estimate the true scattering distributions.
The results of both simulations and experiments are illustrated in Figs. 5, 6, and 7. For the prokaryote Synechococcus sp. strain CC9311 and Flavobacterium sp. strain ALC1, simulated scattering functions were computed using the RM-I method for homogeneous spheroids. Figures 5(a) and 6(a) show the simulations next to the experimentally observed values shown in Figs. 5(b) and 6(b). Reasonable agreement was found in the 40–140° range. In order to aid interpretation, significant features, obtained from visual inspection of the data, were identified and marked with the same color in the pair of scattering phase functions from the same organism.
Simulations and experimental measurements, performed under identical conditions as above, of the scattering curves for the eukaryote Ostreococcus sp. strain CCE9901 are illustrated in Fig. 7. The scattering simulations were performed using the newly developed extended RM-I method. As previously described, the microbe was modeled as a coated spheroid (a=0.5 µm, v=2) with a spherical core (r=0.25 µm) as the nucleus. Figure 7(a) shows the result of the computer simulations and Fig. 7(b) the experimentally observed data.
5. Discussion and conclusions
In this article, we have considered the use of angularly resolved light scattering (ARLS) in order to characterize three major types of picoplankton in the oceans and provide potential for differentiating among the various species. Towards this goal, a rapid and high angular resolution measurement system was employed for experimental measurements, and different models were selected to represent the microbes for simulation. The adequacies of the models are evaluated by comparing simulation results with the measurements.
The determination of the simulation models results from efforts to improve the theoryexperiment agreements. As most traditional studies, homogeneous spheres of different diameters were first used to represent the cells. Since the results of the GLMT simulations yielded a poor match to the experimentally measured scattering curves a second step was taken to model all three types of cells with homogeneous spheroids. Reasonable agreements between RM – I simulations and experiments were obtained for the two prokaryote cells, while the result for the eukaryote Ostreococcus sp. strain CCE9901 was still not satisfactory. As such, a coated sphere model was tried on Ostreococcus sp. strain CCE9901 but the simulated scattering pattern obtained from the extended GLMT showed even more deviation from the experimentally collected data than the RM – I simulation. As a result, the RM – I was extended so that scattering from the eukaryote Ostreococcus sp. strain CCE9901 could be simulated with a coated spheroid. Between the two models listed in Table 1, the coated spheroid with a concentric spherical core gives a better result comparing with the coated spheroid with uniform shell thickness.
Although overall theory-experiment agreements were obtained for the shape of the scattering curve, mismatches still present in some parts of the scattering patterns. The deviations at small (0–40°) and large angles (140–180°) between simulations and experiments are mostly caused by the extraneous light from the red laser beam reflected by the cylindrical cuvette, which limit the accuracy of the data especially for weak scattering from small particles. Apart from this, a main source of mismatch between simulation and experiment is the size and shape polydispersity in the real sample. All simulation used parameters that could represent an average non-dividing cell, however, as a matter of fact, cells in a suspension experience division at different times producing elongated shapes depending on their different stages. The inhomogeneity of the cells due to their internal structures also introduces heterogeneity in the samples that differs from the assumptions of the homogeneous spheroid model. In the case of the Ostreococcus sp. strain CCE9901, the assumption of a concentric and perfectly round nucleus in the simulation introduced additional errors in the results of the simulation.
Several improvements in the scattering setup could be introduced in the future and have the potential to yield superior results. For one, since the main source of mismatch between experiments and theory is from the strong reflection of the laser beam at the air-glass interface, index matching liquid could be an effective solution. As such, a new system is under development which has the laser beam inlet and outlet on the reflector covered with optical windows. The cuvette will then be immersed in the index matching liquid that will, hopefully, result in the elimination of the strong reflections at the incident and 180 degree scatter directions. This approach has been used successfully by others .
In summary, the angularly resolved scattering diagrams from three main types of marine picoplankton (<2µm) were measured in the angular range of 0.5–179.5° with a high angular resolution (<0.5°). Using a variety of models, good agreement was obtained between simulations and measurement, especially for the absolute positions of the scattering maxima and minima between 40° and 140°. Here, a homogeneous spheroid model was used to approximate prokaryote cells, and a RM-I method was used to simulate the phase functions. For improved accuracy, the picoeukaryote cell was modeled as a coated spheroid with a spherical core using a newly developed extended RM-I. Upon understanding scattering properties of different species and knowing the correct models for simulation, a data base for the finger prints of the picoplankton could be established for marine microbe characterization and identification.
J. S. Jaffe and B. Shao would like to thank the Office of Naval Research and the Seaver Institute for supporting this work. The authors would also like to thank Xavier Mayali and Professor Farooq Azam (Scripps Institution of Oceanography, University of California, San Diego) for providing the culture of Flavobacterium sp., thank Brian Palenik (Scripps Institution of Oceanography, University of California, San Diego) for providing the culture of Synechococcus sp. and Ostreococcus sp.
References and links
2. C. Osolin, “Genomes of tiny microbes yield clues to global climate change” (The office of science, US Department of Energy), August 13, 2003, http://www.er.doe.gov/News_Information/News_Room/2003/Genomes%20of%20Tiny%20Microbes.htm.
3. J. M. Garcia-Fernandez, N. T. de Marsac, and J. Diez, “Streamlined regulation and gene loss as
adaptive mechanisms in prochlorococcus for optimized nitrogen utilization in
Oligotrophic environments,” Microbiol. Mol. Biol.
Rev. 68, 630–638
4. N. Touchette, “How the other half lives: Marine Microbes capture the limelight,” (Genome News Network, 2003), September 5, 2003, http://www.genomenewsnetwork.org/articles/09_03/ocean.shtml.
5. R. J. Olson, R. R. Zettler, and O. K. Anderson, “Discrimination of Eukaryotic Phytoplankton cell
types from light scatter and autofluorescence properties measured by Flow
Cytometry,” Cytometry 10, 636–643
6. D. Marie, F. Partensky, S. Jacquet, and D. Vaulot, “Enumeration and cell cycle analysis of Natural
Populations of Marine Picoplankton by Flow Cytometry using the nucleic acid
stain SYBR Green I,” Appl. Environ.
Microbiol. 63, 186–193
7. P. J. Wyatt and C. Jackson, “Discrimination of Phytoplankton via
light-scattering properties,” Limnol.
Oceanogr. 34, 96–112
8. H. Volten, J. F. de Haan, J. W. Hovenier, R. Schreurs, W. Vassen, A. G. Dekker, H. J. Hoogenboom, F. Charlton, and R. Wouts, “Laboratory measurements of Angular
Distributions of light scattered by Phytoplankton and Silt,”
Limnol. Oceanogr. 43, 1180–1197
9. M. E. Lee and M. R. Lewis, “A New Method for the Measurement of the Optical
Volume Scattering Function in the Upper Ocean,” J.
Atmos Ocean. Tech. 20, 563–571
10. Z. Ulanowski, R. S. Greenaway, P. H. Kaye, and I. K. Ludlow, “Laser diffractometer for single-particle
scattering measurements,” Meas. Sci.
Technol 13, 292–296
11. P. Latimer, “Light scattering by
ellipsoids,” J. Colloid Interface Sci. 53, 102–109
12. P. Latimer and P. Barber, “Scattering by ellipsoids of revolution a
comparison of theoretical methods,” J. Colloid
Interface Sci. 63, 310–316
13. C. F. Bohren and D. R. Huffman, Absorption and Scattering of Light by Small Particles (John Wiley & Sons, New York, 1983).
14. D. Stramski, A. Bricaud, and A. Morel, “Modeling the inherent optical properties of the
ocean based on the detailed composition of the Planktonic
Community,” Appl. Opt. 40, 2929–2945
15. S. Ackleson, D. Robins, and J. Stephens, “Distributions in phytoplankton refractive index and size within the North Sea,” presented at the Ocean Optics IX, 1998.
16. Eyvind Aas, “Refractive index of phytoplankton derived from
its metabolite composition,” J. Plankton
Res. 18, 2223–2249
17. M. I. Mishchenko, L. D. Travis, and A. A. Lacis, Scattering, absorption, and emission of light by small particles, 1st edition, (Cambridge University Press, Cambridge, 2002).
18. A. Quirantes and S. Bernard, “Light scattering by marine algae: two-layer
spherical and nonspherical models,” J. Quant.
Spectrosc. Radiat. Transfer 89, 311–321
19. R. Drezek, A. Dunn, and R. Richards-Kortum, “A Pulsed Finite-Difference Time-Domain (FDTD)
method for calculating light scattering from biological cells over broad
wavelength ranges,” Opt. Express 6, 147–157
20. R. C. MacDonald, R. I. MacDonald, B. P. M. Menco, K. Takeshita, N. K. Subbarao, and L.-r. Hu, “Smallvolume extrusion apparatus for preparation
of large, unilamellar vesicles,” Biochim. Biophys.
Acta, Biomembr. 1061, 297–303
21. F. T. Gucker, J. Tuma, H. M. Lin, H. C. M., S. C. Ems, and T. R. Marshall, “Rapid measurement of lightscattering diagrams
from single particles in an aerosol stream and dtermination of latex
particle size,” Aerosol Sci. 4, 389–404
22. D. Watson, N. Hagen, J. Diver, P. Marchand, and M. Chachisvilis, “Elastic light scattering from single cells:
orientational dynamics in optical trap,” Biophys.
J. 87, 1298–1306
23. R. Drezek, A. Dunn, and R. Richards-Kortum, “Light scattering from cells: finite-difference
time-domain simulations and goniometric measurements,”
Appl. Opt. 38, 3651–3661