Charles Mazel
Physical Sciences Inc.
20 New England Business Center
Andover, MA 01730
tel: (978) 689-0003
fax: (978) 689-3232
e-mail:
April 2000. Note - the information on this page is a bit outdated. The Benthic SpectroFluorometer (BSF) first went into the field in 1994 and served well for several years. It has now been superseded by the DiveSpec.
The Benthic SpectroFluorometer (BSF) is a diver-operated instrument for measuring the optical properties (fluorescence and reflectance) of benthic (bottom-dwelling) marine organisms. It was originally designed to measure the spectral emission characteristics of fluorescence from corals illuminated with ultraviolet or other wavelengths of light, but it is also capable of measuring downwelling light or the light reflected from an organism or surface of interest.
Reference:
Mazel, C. H., 1997. Diver-operated instrument for in situ measurement of spectral
fluorescence and reflectance of benthic marine organisms and substrates, Optical
Engineering, 36(9):2612-2617.
The heart of the unit is the Model S1000 Spectrometer from Ocean Optics, Inc. Light reaches the spectrometer via a 600 micron diameter fiber optic cable. A 100 micron entrance slit provides a nominal 10 nm spectral resolution. The spectrometer is fitted with a grating/CCD detector combination which measures at 1024 pixels covering the wavelength range from approximately 250 to 750 nm. Designed to operate connected to an IBM-compatible PC with an A/D board, for this project the unit is controlled by a Tattletale 7 data logger. A 120x64 element LCD text/graphics display with electroluminescent backlight shows the software's operational choices at each step and also provides a graphical data output for in situ quality control. The electronics and optics are contained in a Lexan housing, and the entire package measures approximately 23x12x32 cm. It's buoyancy is adjusted to be slightly negative in seawater.
Excitation energy for fluorescence measurements is provided by a 12V/50W halogen bulb. Emission from the bulb is passed through an interference filter and refocused onto the tip of a 3 mm diameter Liquid Light Guide. A filter wheel that can be operated while underwater permits any of four filters to be selected for a particular measurement.
The excitation and receiving fibers are one meter long. The probe head holds the excitation and receiving fibers at a 45 degree angle. A ring of black neoprene at the probe end provides a light-tight seal when pressed against a surface.
System operation is controlled by custom software. Software is written in C, debugged, and downloaded to the TattleTale EEPROM. The custom software developed for this application presents the operator with a menu-driven interface. The LCD shows the software's operational choices at each stage and also provides a graphical data output for in situ quality control. The operator activates the appropriate switch on a six-position keypad mounted on the hand-held probe to select the desired action. The keys are sealed piezoelectric switches (Tschudin and Heid, Inc.) that are not activated by ambient pressure. A small, waterproof speaker mounted outside the housing provides an audible indication of program functions, freeing the operator to concentrate on proper positioning of the measurement probe.
The keypad can also be used in conjunction with the software to enter alphanumeric information in the data file header. Species name or other desired information can be entered, eliminating the necessity to record information about each reading on a separate underwater notepad.
Data scans are saved to the onboard hard disk drive with a filename that is automatically created from the date and time of the reading. A pressure sensor on the housing is read when optical data is collected and the dpth of the sample is added to the file header. Data download is accomplished via an infrared optical link through the transparent housing.
 |
 |
 |
Above left: Charles Mazel with early prototype version of BSF. Above right: Mazel with BSF at 18 meters. (Photo by Phil Dustan.) Left: Making a measurement. (Photo by Phil Dustan.) |
 |
FLUORESCENCE This is a fluorescence emission spectrum from a specimen of Mussa angulosa (large polyp coral). The excitation wavelength for this scan was 440 nm. The small bump at 685 nm is fluorescence from chlorophyll in the zooxanthellae. The strong emission around 500 nm arises from substances in the coral host tissues. Depth: 7 meters. Dry Tortugas, Florida. |
 |
REFLECTED LIGHT This is a measurement of the spectrum of light reflected from the surface of a specimen of Montastrea cavernosa (large star coral). Depth: 4 meters. Lee Stocking Island, Bahamas. |
 |
REFLECTANCE This plot shows the ratio of the light reflected from a specimen of Montastrea cavernosa (the same one as in the plot above) to the downwelling ambient light reflected from a reference surface (Spectralon tm, Labsphere, Inc.). The prominent reflectance dip at 670 nm is due to absorption of light by the zooxanthellae. Depth: 4 meters. Lee Stocking Island, Bahamas. |
There are three post-processing steps that are routinely applied to the raw BSF data before distribution. These are: correction for dark current and electrical offset; correction for detector array spectral sensitivity; and smoothing.
The raw data is offset from the zero baseline due to electrical offsets and dark
current (buildup of charge in the CCD pixels due to thermal noise rather than incoming
photons.) The offset varies both with integration time and with temperature. An automated
routine for removing the offset has been developed and is automatically applied to the raw
data. Both the raw and offset-corrected data are saved to the disk. The offset correction
routine is described at length in the Optical Engineering paper cited above.
The detector response (electrical signal out for energy flux in) is not uniform for all
wavelengths. This is due to several factors, including: variation in spectral sensitivity
of the detector; blaze of the diffraction grating; and internal optics of the
spectrometer. The detector response peaks in the green and tapers off as you move to blue
or red wavelengths. The detector array spectral sensitivity can be determined by measuring
the output of a light source with known spectral distribution. The BSF is calibrated by
measuring the output of a lamp of standard irradiance (Optronics Laboratories). The ratio
of the known spectral distribution to the measured distribution provides the needed
correction factor at each wavelength. This is stored in a file on the data processing
computer (not the BSF itself) and applied in post-processing.
The data may be a bit noisy even after removal of the offset described above. A Savitzky-Golay smoothing routine,* commonly used for spectral data sets of this sort, may be applied to the data to remove some of the remaining noise.
* A. Savitzky and M. J. E. Golay, Anal. Chem., 36:1627-1639,
1964
J. Steinier et al., Anal. Chem., 44:1906-1909, 1972
The BSF was designed and built by Charles Mazel at the Massachusetts Institute of Technology, with assistance from graduate student Eran Fux. The original funding for the prototype unit was provided by the National Oceanic and Atmospheric Administration through the MIT Sea Grant Program. Continuing development has been funded by grants from the Environmental Optics Program of the Office of Naval Research.