We discuss a maritime surveillance and detection concept based on Raman scattering of water molecules. Using a range-gated scanning lidar that detects Raman scattered photons from water, the absence or change of signal indicates the presence of a non-water object. With sufficient spatial resolution, a two-dimensional outline of the object can be generated by the scanning lidar. Because Raman scattering is an inelastic process with a relatively large wavelength shift for water, this concept avoids the often problematic elastic scattering for objects at or very close to the water surface or from the bottom surface for shallow waters. The maximum detection depth for this concept is limited by the attenuation of the excitation and return Raman light in water. If excitation in the UV is used, fluorescence can be used for discrimination between organic and non-organic objects. In this paper, we present a lidar model for this concept and discuss results of proof-of-concept measurements. Using published cross section values, the model and measurements are in reasonable agreement and show that a sufficient number of Raman photons can be generated for modest lidar parameters to make this concept useful for near-surface detection.
© 2018 Optical Society of America
In discussions on physical security at maritime facilities, we became aware that automated detection of objects near the water surface presents a challenge for current technologies, especially for shallow littoral zones where the scattering environment is complex. Optical systems that utilize elastic scattering, such as the underwater laser-imaging system (UWLIS) , the airborne laser mine detection system (ALMDS) [2,3], and others, have been developed for underwater detection. Such systems rely on a contrast of the scattering signal from water and the objects of interest, which can be difficult to detect near the surface, especially when automated detection is required and the objects are dark without distinct reflectance features.
In this paper, we discuss a maritime detection concept based on Raman scattering by water molecules. It is proposed that an imaging lidar detects Raman scattering from water between the surface and a depth of a few attenuation lengths. The absence or change of Raman signal indicates the presence of a non-water object. With sufficient spatial resolution, a two-dimensional outline of the object can be generated. For objects below the surface, the detection signal can be enhanced by proper range gating. The detection depth is limited by overall attenuation of the incident beam and Raman scattered light.
For visible or ultraviolet (UV) light, the OH stretch band of liquid water has Raman shifts around . This provides ample spectral separation between excitation and Raman light and allows for elastic scattering to be practically eliminated by proper spectral filtering. Only passive backgrounds, such as solar or artificial lighting, would contribute to signal and noise for the proposed concept. A lidar system that measures the passive background can subtract it, and only the noise level is affected when the passive signal is comparable to the Raman signal. Thus, this concept is expected to work well close to the water surface and can be complementary to other technologies, such as ALMDS, which is capable of detection at significant depths.
Here we consider this concept and present a lidar model of the Raman scattering process for a body of water. Using published values of the cross section, it is shown that sufficiently large numbers of Raman photons are produced to make this concept attractive and, potentially, useful for many applications. Among potential applications are automated security in littoral regions, near-surface mine detection, discovery and mapping of floating or near-surface objects in day/night wide area searches, and other applications. Of particular interest is mapping of objects close to the surface where wave motion would make it difficult to interpret signals for elastic scattering-based technologies. If excitation in the UV range is used, the presence or absence of fluorescence from objects can add to discrimination capability. We also present results of proof-of-concept measurements using the Sandia Ares lidar at 355 nm excitation . Three water samples were measured: bottled drinking water, Pacific Ocean water, and San Francisco Bay water. For the bottled-water sample, the measurements and the lidar model results are in good agreement, assuming attenuation of incident and Raman light based on published values for pure water. These measurements indicate that the proposed detection concept is feasible and provides sufficient signal photons for moderate lidar parameters.
2. RAMAN CROSS SECTION
Raman scattering is an inelastic process in which a photon loses or gains energy by changing an energy level of the scattering molecule. For incident polarized light and scattered natural light, the band-integrated differential Raman cross section is denoted by , where is the scattering angle, and is the angle between the incident light polarization direction and the scattering plane (the plane formed by the incident and outgoing directions). Several measurements of this cross section are reported in the literature for the OH stretch vibrational mode of water molecules and for excitation at [5–11]. The published value in Ref.  for , for incident perpendicular polarized light, or , is . The angular dependence of this cross section is discussed in [10,11] and is given by11]. For , applicable to the lidar detection considered here (for co-located source and detector), Eq. (1) shows that the cross section is equal to that is reported in the literature.13]. Note that here we use frequency and wavenumber interchangeably, where the wavenumber is given in terms of wavelength by . At 488 nm incident light, , , and the result of Eq. (2) is , which is close to the above value. At 355 nm incident light, the wavelength at which lidar measurements discussed later in this paper are made, , , and the differential Raman scattering cross section is . In this paper, only liquid water is considered. Other applications, such as cloud and ice thickness measurements, use Raman cross sections and spectra for water vapor and ice [15–17].
3. LIDAR MODEL
The lidar model considered here is for a pencil beam and one detector pixel. At the water surface, given the incidence angle and beam polarization, the refracted beam angle and intensity are known from basic optics. As the refracted beam propagates in water, it will be attenuated and undergo Raman scattering. The lidar is assumed sufficiently far from the surface so that only photons scattered close to 180° are in the detector field of view. These photons will nearly follow the reverse path of the incident beam and will also be partly reflected at the water–air interface. An illustration of the lidar and beam propagation is shown in Fig. 1.
Along the beam path, at distance from the surface, the differential number of Raman photons produced in thickness per unit area of beam, per unit time, and in solid angle is3) over the beam cross section, the intensity is replaced with the local beam power.
Consider the time dependence of the detected Raman signal, let be measured from the start of the beam pulse at the lidar location, and let the lidar beam pulse shape be given by . For a differential segment of the beam pulse at time and width , Raman photons produced at path length in the water will arrive at the detector at time , where is the range from the lidar to the water surface, and and are the speeds of light in air and water, respectively. The factor of 2 accounts for the round-trip time. The differential energy measured by the detector for this beam segment from Raman photons produced in path length is4) over the interaction thickness of water and over the beam pulse width, resulting in
A. Numerical Estimate of Raman Photons
Consider a quick estimate of the expected number of Raman photons for a single pulse to get an idea of the expected signals and therefore the feasibility of the concept. Lidar parameters similar to those for the proof-of-concept measurements will be used. For this example, the transmissions and refracted fractions are set to unity (propagation normal to the water surface), thus , where is the overall lidar optical efficiency. Also consider the limiting case of zero gate delay (), a wide gate (), and neglect attenuation in water, that is, and . For these limits, Eq. (5) reduces to18].
B. Solar Background
For daytime applications, the solar background is relevant to the concept discussed here. For nighttime, other backgrounds such as lunar and passive lighting would need to be considered, but they are expected to be much smaller than the solar.
The spectral solar irradiance reaching a water surface as a function of wavelength can be estimated using the Moderate Resolution Atmospheric Transmission code MODTRAN . Spectral irradiance is the radiative power incident on a surface per unit area per unit wavenumber (spectral radiative flux) and is given in units of watts per per in MODTRAN. Figure 2 shows a MODTRAN estimate of the solar spectral irradiance. The integrated value of this spectrum (summing over all spectral bins and multiplying by the bin width of ) is . This is slightly less than the solar constant of 2.0 calories per per minute , or due to the effect of atmospheric attenuation.
An estimate of the solar energy detected by the lidar for each pulse in the Raman spectral range is obtained by assuming diffuse reflection of solar light at the water surface and is given by8) excludes the case of direct solar glint, which is orders of magnitude higher and needs to be avoided in application of the detection concept discussed here. The area , where is the field-of-view angle of the lidar system. Thus, is independent of the lidar range . Similarly, the solar background given by Eq. (8) is independent of range, except for the effect of atmospheric transmission. This is in contrast to the Raman signal, which has dependence, and implies that for any lidar there is a maximum detection range where noise from the solar background will dominate the Raman signal.
For a numerical estimate of the solar background, consider similar parameters to the previous section: , gate width , , and milliradians. Near , . Consider a spectral interval of width 10 nm around , which corresponds to . This interval is sufficient to contain the Raman signal for liquid water. For an upper limit estimate, we use and , resulting in per pulse. This value is much smaller than the Raman energy estimated in the previous section, and thus the solar background is expected to be negligible for the proof-of-concept measurements discussed here. However, because of the scaling with the different parameters and the possibility of spreading the beam over multiple pixels, the solar background has to be considered in modeling for any specific application, and the instrument needs to be designed with the capability to measure this background.
4. ATTENUATION IN WATER
Many measurements of light attenuation in a wide spectral range are reported in the literature. For pure water, the attenuation coefficient is a strong function of wavelength with a minimum of in the visible region near 480 nm [21–23]. Figure 3 shows this dependence along with the attenuation for the corresponding Raman wavelength. For natural waters, the attenuation strongly depends on particulate and organic matter content and can vary by several orders of magnitude at the same wavelength [24,25]. For clear ocean water,  it shows similar dependence of the attenuation on wavelength as Fig. 3, with a minimum slightly less than . This indicates that the effect of the salt content on attenuation is negligible in the wavelength range of interest.
For the proof-of-concept measurements presented in the next section, at 355 nm excitation and Raman at 404 nm, the attenuation coefficients of pure water are and , respectively, based on Fig. 3. Thus, in Eq. (5) is , corresponding to attenuation length of .
From Fig. 3, the optimum wavelength for pure water is in the range 425–450 nm for the proposed detection concept. For green light at 532 nm, the overall attenuation coefficient is estimated at , which is worse than for UV at 355. However, the effect of particulates and organic matter is expected to result in much worse attenuation for UV light. Since the Raman cross section is also a strong function of wavelength, the optimum excitation wavelength will depend on the particular water conditions for the application.
5. PROOF-OF-CONCEPT MEASUREMENTS
The Sandia Ares lidar  and a water column with a movable target were used to demonstrate the detection concept and validate the number of detected Raman photons estimated by the model. A schematic of the Ares lidar is shown in Fig. 4. The Ares laser (commercial Big Sky Laser CFR200) generates a nominal 30 mJ, wide pulse at 355 nm and is operated at 30 Hz. Elastic back-scattered, Raman, and laser-induced fluorescence (LIF) light is collected by an 18.8 cm telescope and directed to a long-pass dichroic beam splitter, which reflects light with wavelengths shorter than , effectively separating the elastically backscattered laser light from the LIF and the Raman scattered light at 404 nm for the band for water. LIF and Raman light is detected by a time-gated intensified charge-coupled device (CCD, Andor iStar ICCD) with a spectrally resolved channel of 512 spectral bins from 295 to 730 nm. This channel has no sensitivity below 360 nm due to the dichroic beam splitter. The elastic backscattered light is detected by a time-resolved photo-multiplier tube (PMT), which is used for ranging and setting the ICCD time gate. For each laser pulse, in addition to the spectrum, the passive background spectrum is also measured.
The measurement setup is shown in Fig. 5. The water column is a horizontal stainless steel pipe that is 152 cm in length and 14.0 cm. in diameter with several ports that allow insertion of an expandable target to serve as a beam stop at specific depths along the water column. Fused silica windows on both ends of the pipe were used and resulted in small beam losses estimated at . The lidar was located away from the pipe front window. At this range, the lidar overlap factor is about 22% (see below), and the beam spot size is less than 5 cm in diameter. The pulse energy was manually measured before and following the measurements. For all measurements the detector gain was adjusted to avoid saturation. Calibration factors at various gain settings are available from past measurements. The lidar overlap factor is a measure of the fraction of collected light that is focused on the detector and is estimated as a function of lidar range using atmospheric nitrogen () Raman measurements as shown in Fig. 6.
Three water samples were measured: (1) pure bottled water, (2) Pacific Ocean water, and (3) San Francisco Bay water. The first water sample was bottled water, and the Raman spectra are shown in Fig. 7 for several target locations. Figure 8 shows the summed spectra over the spectral range 390–410 nm, along with scaled model dependence using , where is the target location. This model results from Eq. (5) in the limit of a wide gate relative to the pulse width ( and ). For these measurements, the expandable target allowed a small fraction of the beam to escape ( as estimated using a green paper fluorescent target behind the rear window). The correction used in Fig. 8 and other details of the measurements are discussed in Ref. .
When attenuation in water is neglected, we earlier estimated Raman photons per pulse for , with , , , , and a water thickness . Scaling this to the Ares measurement for the first target location (using , , overlap at the lidar range based on Fig. 6), for the same pulse energy and optical efficiency, the scaled number of photons per pulse is photons. From Fig. 8, the corrected Raman signal is counts for 500 pulses or counts per pulse for the first target location. This corresponds to photons after application of the lidar absolute calibration factor at the detector gain setting (previous absolute calibrations, using a standard lamp, at a gain setting of 250, resulted in at 500 nm with accuracy to within a factor of two. The relative calibration factor between gain settings of 50 and 250 is 224 based on past relative calibration measurements). This number of photons is roughly within a factor of three of the model-obtained value, which is reasonable given the uncertainties in the parameters that were used (the optical efficiency is likely smaller than the 10% that was used—10% is based on measurements that were done several years ago). Therefore, these measurements give us confidence that the proposed detection concept has sound basis and the lidar model can be used for reasonable estimates of photon budgets.
Similar measurements were carried out for the ocean and SF Bay samples. The spectra are shown in Figs. 9 and 10 and show significant fluorescence, likely from organic matter. The organic content can be significantly different depending on location and water type [24,25]. To obtain the Raman counts requires partitioning of the Raman and fluorescence spectra. After partitioning, the maximum Raman counts were used to obtain an estimate of for the ocean sample and for the SF Bay sample, assuming the maximum signals reached asymptotic values. Even at of , a sufficient number of Raman photons is expected from the first 10 cm of water to allow for detection of objects just below the surface. It should be noted that the ocean and bay samples were obtained very close to the shoreline and therefore might not be accurate representations of these bodies of water. For all measurements reported here, the daytime passive background (ambient solar) is significantly smaller than the Raman signals for the same detector gate width .
The proof-of-concept measurements presented in this paper show that the proposed concept of using changes in water Raman scattering signals for detection of objects close to water surfaces is sound and could potentially complement other optical detection technologies in the region where elastic scattering is problematic. Lidar measurements of the Raman signal for the pure water sample at varying target positions are consistent with an overall attenuation coefficient of for an incident beam at 355 nm and Raman at 404 nm. Measurements of water samples from the Pacific Ocean and the San Francisco Bay show much larger attenuation coefficients estimated at 4.5 and for these two samples, respectively. The large attenuation is likely from organic matter content in the water. Longer wavelengths are expected to result in better penetration depths, especially for waters with high organic content. Even with the high attenuation and modest lidar parameters, sufficient Raman photons are generated to make near-surface detection of objects feasible using this concept.
In this paper, we discussed the general detection concept without a specific detection algorithm. It is anticipated that detection algorithms will need to be developed for specific applications and instruments; these algorithms will result in sensitivity and detection limits based on the lidar model. Because of the large variation of Raman signal for different waters, it is assumed for this detection concept that the water is reasonably uniform on the scale length of detection objects of interest. Under this condition, a detection algorithm can consider changes in Raman signal in the vicinity of the object, rather than the absolute value of the signals.
The measurements reported in this paper utilized the available Ares lidar with its gated ICCD to detect the Raman spectrum for demonstration of the detection concept. For a practical application, it is likely advantageous to use multiple PMTs, filtered for specific spectral bands, to detect the time-resolved Raman and fluorescence signals. The time-dependent Raman signals are expected to provide a clear target pattern without needing to set a time gate, making for easier and more precise target detection. The passive background is the PMT signal at long times after complete attenuation of the beam and decay of any fluorescence.
Sandia National Laboratories; U.S. Department of Energy’s National Nuclear Security Administration (DE-NA0003525).
The authors acknowledge Tom Kulp, Don Sheaffer, and Will Bolton for useful discussions and feedback.
This paper describes objective technical results and analysis. Any subjective views or opinions that might be expressed in the paper do not necessarily represent the views of the U.S. Department of Energy or the United States Government.
1. T. J. Kulp, D. Garvis, R. Kennedy, T. Salmon, and K. Cooper, “Development and testing of a synchronous-scanning underwater imaging system capable of rapid two-dimensional frame imaging,” Appl. Opt. 32, 3520–3530 (1993). [CrossRef]
2. M. E. Kushina, G. Heberle, M. Hope, D. Hall, M. Bethel, and L. K. Calmes, “ALMDS laser system,” Proc. SPIE 4968, 163–168 (2003). [CrossRef]
3. “AN/AES-1 Airborne laser mine detection system (ALMDS),” http://www.globalsecurity.org/military/systems/aircraft/systems/an-aes-1.htm.
4. R. L. Schmitt, K. L. Schroder, M. W. Smith, L. J. Krumel, P. J. Hargis Jr., I. R. Shokair, T. A. Ashlock, J. W. Daniels, J. R. Klarkowski, M. S. Johnson, C. M. Boney, P. J. Claassen, G. I. Magee, M. L. Pedroncelli, J. T. Spooner, and A. R. Lang, “Ares UV LIF standoff system development and testing,” in Proceedings of the 6th Joint Conference on Standoff Detection for Chemical and Biological Defense, Williamsburg, Virginia, October 2004.
5. C. H. Chang and L. A. Young, “Seawater temperature measurements from Raman spectra,” N62269-72-C-0204 (Advanced Research Projects Agency, 1972).
6. N. P. Romanov and V. S. Shuklin, “Raman scattering cross section of liquid water,” Opt. Spectrosc. 38, 646–648 (1975).
7. R. B. Slusher and V. E. Derr, “Temperature dependence and cross sections of some Stokes and anti-Stokes Raman lines in ice Ih,” Appl. Opt. 14, 2116–2120 (1975). [CrossRef]
8. I. I. Kondilenko, P. A. Korotkov, V. A. Klimenko, and O. P. Demyanenko, “Transverse cross section of the Raman scattering of the u1 vibration of the water molecule in the liquid and gaseous states,” Opt. Spectrosc. 43, 384–386 (1978).
9. S. Sugihara, M. Kishino, and M. Okami, “Contribution of Raman scattering to upward irradiance in the sea,” J. Oceanogr. Soc. Japan 40, 397–404 (1984). [CrossRef]
10. B. R. Marshall and R. C. Smith, “Raman scattering and in-water ocean optical properties,” Appl. Opt. 29, 71–84 (1990). [CrossRef]
11. G. W. Kattawar and X. Xu, “Filling in of Fraunhofer lines in the ocean by Raman scattering,” Appl. Opt. 31, 6491–6500 (1992). [CrossRef]
12. W. K. Bischel and G. Black, “Wavelength dependence of Raman scattering cross sections from 200–600 nm,” AIP Conf. Proc. 100, 181–187 (1983). [CrossRef]
13. G. W. Faris and R. A. Copeland, “Wavelength dependence of the Raman cross section for liquid water,” Appl. Opt. 36, 2686–2688 (1997). [CrossRef]
14. J. S. Bartlett, K. J. Voss, S. Sathyendranath, and A. Vodacek, “Raman scattering by pure water and seawater,” Appl. Opt. 37, 3324–3332 (1998). [CrossRef]
15. V. Rizi, M. Iarlori, G. Rocci, and G. Visconti, “Raman lidar observations of cloud liquid water,” Appl. Opt. 43, 6440–6453 (2004). [CrossRef]
16. F. Liu and F. Yi, “Spectrally resolved Raman lidar measurements of gaseous and liquid water in the atmosphere,” Appl. Opt. 52, 6884–6895 (2013). [CrossRef]
17. S. M. Pershin, V. N. Lednev, V. K. Klinkov, R. N. Yulmetov, and A. F. Bunkin, “Ice thickness measurements by Raman scattering,” Opt. Lett. 39, 2573–2575 (2014). [CrossRef]
18. I. R. Shokair, M. S. Johnson, R. L. Schmitt, and S. Sickafoose, “Concept for maritime near-surface surveillance using water Raman scattering,” Report SAND2016-10361 (Sandia National Laboratories, 2016).
19. L. W. Abreu and G. P. Anderson, “The MODTRAN® 2/3 report and LOWTRAN 7 model,” Ontar Corporation for PL/GPOS (1996).
20. R. C. Weast, ed., Handbook of Chemistry and Physics, 66th ed. (CRC Press, 1985), section F-156.
21. G. M. Hale and M. R. Querry, “Optical constants of water in the 200-nm to 200 µm wavelength region,” Appl. Opt. 12, 555–563 (1973). [CrossRef]
22. J. D. Jackson, Classical Electrodynamics, 2nd ed. (Wiley, 1975).
23. R. C. Smith and K. S. Baker, “Optical properties of the clearest natural waters (200–800 nm),” Appl. Opt. 20, 177–184 (1981). [CrossRef]
24. C. D. Mobley, Radiative Transfer in the Ocean (Academic, 2001).
25. D. P. Morris, H. Zagarese, C. E. Williamson, E. G. Balseiro, B. R. Hargreaves, B. Modenutti, R. Moeller, and C. Queimalinos, “The attenuation of solar UV radiation in lakes and the role of dissolved organic carbon,” Limnol. Oceanogr. 40, 1381–1391 (1995). [CrossRef]