First attempts at direct Raman detection of the oceanic carbonate system
Peter G Brewer1, Rachel, M. Dunk1, Sheri N White1, Edward T Peltzer III1, Bryan Bowie2, and Peter Walz1
1MBARI, 7700 Sandholdt Rd., Moss Landing, CA 95039
2Kaiser Optical Systems, Inc., 371 Parkland Plaza, Ann Arbor, MI 48103
MBARI’s Deep Ocean Raman In Situ Spectrometer (DORISS) has been deployed on a number of scientific missions in the ocean, with successful in situ acquisition of spectra from a wide range of targets including minerals, gases, and gas hydrates. We are now exploring the future use of this instrument for the direct detection of aqueous species in natural waters, with a particular focus on the oceanic carbonate system. The ability to acquire in situ data allowing a direct quantitative assessment of carbonate speciation would present a powerful new tool to the oceanographic community.
Ocean scientists have long relied on indirect methods for determination of the individual carbonate system species in seawater through linking observations with apparent thermodynamic constants. However, the direct Raman spectroscopic observation of HCO3- and CO32- in solution is possible, although acquisition of spectra in-situ is difficult due to low concentrations in natural waters. We are attempting to address this problem using the DORISS instrument. The first objective is to determine the Raman scattering efficiency of HCO3- and CO32-, and thereby define the instrument requirements to enable direct determination of carbonate speciation in a ~2mM TCO2 solution. The second step is to improve the sensitivity of the system through improvements of the optical path and advanced post-processing techniques.
To perform quantitative measurements using Raman spectroscopy a reference standard is required, where the relative scattering efficiency of the standard and target must also be known (the Raman cross section ratio, sstandard/starget). We have chosen to utilize SO42- as the reference standard for seawater species, as the seawater SO42- signal is readily detected in situ with short (sub 1 minute) spectral acquisition times. The scattering efficiencies of HCO3- and CO32- with respect to SO42- were determined in the laboratory (sSO4/sCO3 = ~2.5; sSO4/sHCO3 = ~4.2). As both HCO3- and CO32- are comparatively weak Raman scatterers, and the TCO2 of seawater is ca. 10 times lower than the SO42- concentration, it is necessary to increase instrument sensitivity by a factor of 10-100 to enable rapid (sub 1 minute) in-situ detection of the seawater carbonate system.
Current instrument sensitivity allows detection of ~15 mmolkg-1 HCO3-, or ca. 7 times the seawater concentration, using standard spectral processing software (GRAMS AI, ThermoGalactic) and long (~30 minute) spectral acquisition times. To increase sensitivity we have explored potential adaptations to DORISS and alternative methods of spectral processing. First, we have tested a new sampling optic, a 532 nm PhAT probe (Kaiser Optical Systems, Inc.). The current DORISS probe head utilizes a single optical fiber for signal collection, whereas the PhAT probe utilizes a bundle of 50 collection fibers, leading to a concomitant increase in instrument sensitivity. Secondly, the application of partial least squares spectral processing (Eigenvector Research, Inc.) notably improves the Raman peak detection limit. The incorporation of the PhAT probe, in addition to advanced spectral processing, could significantly increase the sensitivity of the DORISS instrument, enabling direct measurement of the seawater bicarbonate system.