Gulf of California Logbook
During the Spring of 2003 the R/V Western Flyer and ROV Tiburon traveled to the Gulf of California (left). During Leg 5 of this expedition (April 21st – May 11th), we had the opportunity to use DORISS to analyze natural gas seeping from the sea floor along a transform ridge just north of Guaymas Basin.
We were able to locate a single gas vent at 27°35.5’N, 111°28.5’W, 1582 m (below left). During four ROV dives to this site we attempted to collect in situ Raman spectra of natural gas bubbling from the sea floor, and solid gas hydrates found in the sediment. We also collected gas samples for GC shore-side analysis to verify the in situ data.
The method of obtaining gas spectra was through use of the immersion optic protruding into an open-bottomed cube (above right). Gas was trapped in the cube forming a gas space around the immersion optic (i.e., no seawater was present in the beam path). The one problem was the formation of a partially-opaque hydrate skin on the window of the immersion probe. This partially blocked both the exiting beam and the backscattered signal, thus greatly reducing the sensitivity. Once the sample was brought above the hydrate stability zone (~585 m depth, ~7.0°C) clean, high quality spectra could be obtained.
Gas samples for shore based analysis were collected in 150 ml stainless steel cylinders. Three evacuated cylinders were integrated into a sampling funnel with a heating element. During sample collection, gas hydrate formed in the sampling funnel which could clog the lines to the cylinders. The heating element was used to heat the sample and melt the hydrate. The gas could then be pulled into an evacuated cylinder by actuating a valve. Sixteen samples were collected in this manner. These samples include gas bubbling from the vent, gas bubbles captured above the vent, gas within the sediment, and melted hydrate from within the sediment.
Raman spectroscopy, which is capable of identifying covalently bonded gas species, can be used to determine the composition of natural gas. The primary constituent of the vent gas was assumed to be methane which has a strong Raman peak at ~2917 cm-1 (nu1 vibrational mode), and smaller peaks at 1535 (nu2), 3020 (nu3), and 3070 cm-1 (2nu2). Non-methane hydrocarbons (NMHC) were also assumed to be present in measurable quantities.
Data processing was performed with the GRAMS/AI spectral processing package from ThermoGalactic. The GRAMS peak fitting algorithm deconvolves the peaks, calculates and subtracts the baseline function, and determines peak positions (Raman shift), heights, widths and areas.
The highest quality spectra were collected with the immersion optic after the sample was raised above the hydrate stability zone (such that the hydrate skin formed on the optical window was dissipated). The spectrum above was collected at a depth of 500 m. The 1332 cm-1 diamond standard peak is clearly visible in the spectrum. The step at ~2200 cm-1 is where the two spectral windows are joined.
There are three main observations from this data:
1) Methane is the only component that has been identified in the spectra to date. While other higher hydrocarbons may be present, they are not in high enough concentrations to be detected by Raman analyses. GC analysis on shore verified that methane is present at ~97%.
2) Brunsgaard Hansen et al.  observed that the pressure (density) of a methane mixture can be inferred from the ratio of the 3020 (nu3) and 3070 cm-1 (2nu2) bands. The intensity ratios we observed are similar to those observed by Brunsgaard Hansen et al. : 0.67 @ 1582 m (~16.1 MPa); 1.75 @ 250 m (~2.6 MPa).
3) The intensity and resolution of the spectra reveal not only vibrational bands of methane (e.g., 2917, 3020, and 3070 cm-1), but also coupled rotational-vibrational bands from intramolecular Coriolis forces.
The hydrocarbons detected consist of methane to iso-pentane. Methane is the major component (96.7 - 98.8%). NMHC are trace components (less than 0.28%). CO2 is 2.7 - 3.2% except for sample T0578 (1.4%). The ratios of C1/(C2+C3) (less than 103) suggest the contribution of small amounts of thermogenic gases. In sample T0578, the material recovered contained both primary gas and hydrate fragments released by sediment disturbance. The hydrocarbon results are similar to the other samples. The presence of trace quantities of iso-C5, and undetectable n-C5 suggests some fractionation mechanism. We note that structure-H hydrate could provide this as iso-C5 can be included in this structure but n-C5 cannot. However, we have no direct evidence for the presence of this hydrate form. For T0563 and T0573, we see small changes in the concentration of C1 and CO2 gases near 5 m and 8-10 m above the vent. This may be controlled by bubble dissolution and rise rate, with the highly soluble CO2 component (solubility ~10.5 greater than methane) being rapidly lost.
GC analysis was performed by Noriko Nakayama and Ed Peltzer at MBARI.
Next: Hydrothermal VentsQuestions? Comments? Please contact Edward Peltzer.